Professional Documents
Culture Documents
AGRICULTURAL
ENGINEERING
EDITORIAL
Edward McKyes Inside front cover
Barrington, S. F., P. J. Jutras, and R. S. Broughton. 1987. The sealing of soils by manure. I. Preliminary inves-
tigations. Can. Agric. Eng. 29: 99-103.
Infiltration rates, groundwater contamination and changes in soil profile nutrient composition were observed for four
experimental field reservoirs filled with 10% total solids (TS) dairy manure. The results indicated poor correlation between
manure soil infiltration and soil saturated hydraulic conductivity with water. These findings were confirmed in the
laboratory using nine columns of four different soil textures exposed to 6% dairy manure.
facilities being expensive, government (1976) 0.4 - 0.8 10 2.5 Dairy Sand to Laboratory
clay
schemes to improve manure storage pro
Nova Scotia Canada
gressed at a slow pace. Priority was there Lo
fore given to the development of (1977) 1.16 - 0.06 3.5-7.75 2.4 Diluted Sand to Laboratory
construction guidelines for environmen dairy clay
tally acceptable low-cost manure storage tk refers to hydraulic conductivity value rather than an infiltration rate.
facilities, such as earthen reservoirs. Such
development would provide a more favor
able economic climate in which to acceler growth factor, Q'. Chemical mechanisms rates, as well as groundwater contam
ate environmental schemes improving require temperatures above 15C during ination, provided no bases for definite
surface water quality, by reducing pollu the soil reduction process (Mirtskhulara guidelines for the construction of earthen
tion from manure. et al. 1972). storage facilities as of 1980. At that time,
The authors summarized in Table 1 have Nordstedt et al. (1971), Collins et al. authorities based their bylaws on soil
suggested three possible sealing mecha (1975), Sewell (1978), Ciravolo et al. hydraulic conductivity values. In Ontario,
nisms: (1) physical mechanisms by which (1979) and Patni et al. (1981) investigated a value of 8.64 x 10~2 m/d (10"4 cm/s)
soil pores become clogged, (2) biological groundwater contamination from earthen was required for the soil on site. Quebec
mechanismsthrough bacterial activity, (3) reservoirs via seepage losses. They identi increased its 1975 value of 8.64 x 10-5
chemical mechanisms where soil clay par fied bacteria (total and fecal coliforms, m/d (10 "7 cm/s) to 8.64 x 10~4 m/d
ticles deflocculate and soil structure is fecal streptococci), ammonia, nitrates and (10-6 cm/s) in 1981. The U.S. state of
destroyed through reductive processes. chlorides as major contaminants. But they Pennsylvania required a permeability
The last two mechanisms were identi also observed that the concentration of value of 8.64 x 10'4 m/d (lO"6 cm/s)
fied as temperature dependent. Biological groundwater contaminants varied widely after sealing had taken place.
mechanisms rely on bacterial enzyme with time and location for any given reser A research project was initiated in 1981
activity which follow the general tem voir. to determine the extent of the sealing of
perature rate relationship expressed by the The various levels of manure infiltration soils by manure under Quebec conditions.
7 ) t /
7
/ *
111/11
-N '///>}
Ora/n
^200 H 450 cm ->lO0 H /Water 7ab/c?
Perforated Well
The results of this project wereto verify the water table maintained naturally some samples were tested for bacteria(total and
Quebec norm requiring a soil hydraulic 0.60 m below the reservoir floor. fecal coliforms and fecal streptococci),
conductivity under 8.64 x 10~4 m/d These four reservoirs were filled in ammonium, nitrate, phosphorous and
(10"6 cm/s). October 1981 to a depth of 2.4 m with 10% potassium. In November 1982, the reser
total solids (TS) liquid dairy manure. They voirs were emptied and their soil profiles
PROCEDURE were observed for infiltration rates and were analyzed for pH, phosphorous,
To investigate sealing efficiency under groundwater contamination until No potassium and total Kjeldahl nitrogen.
natural conditions, four small manure res vember 1982. Reservoir no. 4 had to be re Core samples were taken for this purpose
ervoirs were built on sites of various soil plenished in May 1982. at two places per site on the reservoir side
textures (Site no. 1 of clay, Site no. 2 of Infiltration rates were monitored reg walls, 100 cm above their floor. These
loam, Site no. 3 of coarse sand and Site no. ularly by measuring the levelof the surface cores were taken perpendicular to the res
4 of 1.2 m of coarse sand over a grey crust of each reservoir, using an engineer's ervoir surface, at 10-cm intervals to a depth
structureless clay). These reservoirs were level and a bench mark established within of 90 cm.
built of minimum dimension to reduce 5 m of each site. Evaporation rates and In September 1982, nine dairy manure
costs, thus a floor dimension of 1.0 m x rainfall data were obtained from weather columns (three of clay, three of loam, two
1.0 m, but of typical hydrauliccharge, thus stations located 2 km away (sites nos. 2, 3 of sand and one of gravelly silty clay) were
of 3.0 m depth (Fig. 1). Their side had a and 4) and 10 km away (site no. 1). This set up in an unheated laboratory (Fig. 2)
slope of 1.5:1.0 for stability, considering experimental method was based on pro This second project was conducted to
especiallythe sandy sites. At the firstthree cedures used by Meyer et al. (1972), Davis obtain more accurate readings on infiltra
sites, the groundwater table wascontrolled et al. (1973) and Robinson (1973). Infiltra tion rates and groundwater pollutioi haz
below the bottom of the reservoir, while at tion data were compared statistically using ards. Each column held a 10-cm bigh by
the fourth site, no such control was under the method of analysis of variance. 10-cm diameter undisturbed B Horizon
taken. This allowed the observation of Groundwater quality was monitored soil sample (sample dimension recom
groundwaterlevel effects. Twoof the con through the sampling of water from a well mended for soil hydraulic conductivity
trolled reservoirs required a peripheral 2.8 m deep within 3 m of each reservoir measurements), exposed to a head of 1.80
drain 0.30 m below the floor levels (sites and from a control unaffected by the reser m of 6% TS dairy slurry. The seepages
no. 1 and no. 2), emptying into a nearby voir seepages if any, thus located more (exfiltrates) recuperated under each soil
ditch. Site no. 3 demonstrated a ground than 200 m upgrade from each site. Water column were used to measure infiltrat/on
100
CANADIAN AGRICULTURAL ENGINEERING, VOL. 29, NO. 2, SUMMER 1987
rates. These exfiltrates were collected dur
ing the second month for nutrient and min
eral analyses. The columns were observed
for 3 mo, October through December
1982.
148 135 132 110 135 82 lected. Again, the sealing of soils by
50-60 275
105 97 145 212 67 manure was successful in reducing infiltra
90 116 130 201 114 93 71
11
major role; biologicaland chemical mech
30-40 22
50-60 22
5 11 8 5 Trace anisms were secondary, as their activity is
90
8 8 14 0 8 2 only significant at temperatures above
15C.
Laboratory column exfiltrate analysis
TABLE IV. LABORATORY COLUMNS INFILTRATION RATES
showed poor correlation between soilclay
content and sample contaminant con
Manure infiltration
centration. The color of the exfiltrate
rate
Hydraulic clearly indicated that the amount of
Column conductivity 48 h 840 d Exfiltrate
(10-6m/s) (10-8m/s) (10-9 m/s) turbidity t organic matter being carried through the
no. Soil
soil was a function of soil clay content.
1 Sand 2.35 6.7 5.0 Heavy
2.27 1.3 6.0 Light Furthermore, comparison of concen
2 Loam
3 Clay 0.06 0.01 2.2 Clear trations among all three clay soil assays
4 Sand 2.11 2.56 6.6 Heavy leads to the probable link of soil structures
5 Loam 3.58 1.50 7.0 Light or permeability to the quantities of exfil
6 Gravelly 1.74 6.6 Light
Barrington, S. F., P. J. Jutras, and R. S. Broughton. 1987. The sealing of soils by manure. II. Sealing mecha-
nisms. Can. Agric. Eng. 29: 105-108.
The sealing mechanisms for soils being infiltrated by dairy and swine slurries were investigated using laboratory
infiltration columns. Physical mechanisms were found to be predominant especially for dairy slurries. Biological
mechanismsinterveneto strengthen physical seals only where ambienttemperatures exceed 15C. Biological mechanisms
are especially important in the sealing of soils by swine slurries as their physical mechanisms are weaker than those of dairy
slurries. Chemical mechanisms were found to be insignificant.
INTRODUCTION nutrients accumulated within the reservoir TABLE I. PARTICLE SIZE ANALYSIS OF
COARSE SAND UTILIZED IN
A preliminary project (Barrington et al. profiles. Miller and Robinson (1981) also
INFILTRATION COLUMNS
1987) established that physical mecha monitored groundwater quality below a
Fraction
nisms play a primary role in the sealing of beef manure earthen reservoir to observe Particle size
were successful in sealing soils to infil several other authors having monitored 53-2 Trace
tration rates below 8.64 x 10-4 m/d groundwater quality around earthen 53-75 1
75-150 9
(10~6 cm/s), even on soils with initial per manure storages. Some of the authors 150-250 82
meability values exceeding the limits set observed seepage effects but not under a >250
by the Ontario and Quebec environmental systematic basis.
authorities. This project, following from that of
These findings were in agreement with Barrington et al. (1987), aims at investi by returning to the individual column, the
those of several previous research works. gating the specific mode of action as well liquids collected as seepages (exfiltrates).
DeTar (1977, 1979) demonstrated, using as the relative importance of the physi The second trial pertained to the sealing
manure seeping throughsiltysoils, thatTS cal, biological and chemical sealing effects of chemical mechanisms. Eight
content above 1.8% was far more signifi mechanisms. clay soil samples, 0.048 m in diameter by
cant in controlling final infiltration rates 0.010 m high, were compacted at the bot
than either soil saturated hydraulic con PROCEDURE tom of aluminum tubes (Fig. 2). After
ductivity to water(k) or manure hydraulic The investigation of the three sealing measuring the saturated hydraulic con
head. Lo (1977)found, using dairy lagoon mechanisms was carried out through four ductivity of each clay core, four columns
liquids over columns of various soil tex laboratoryinfiltration trials. Exceptfor the were soaked in tap water and four more
tures, that final infiltration rates were not chemical mechanism investigation, all columns were soaked in swine slurry. All
significantly different despite wide varia assays were performed using plexiglas eight columns were kept under tempera
tions in hydraulic heads. Rowsell (1980) tubing of 0.14 m insidediameter. A coarse tures of 0-5C for minimum bacterial
compared sealing effects of screened beef sand column, 0.15 m high, was placed at activity. Carewastaken as to soaktheclay
manure, formaldehyde-sterilized screened the lower extremity of the tubing and was samples from the bottom of the tubes to
beef manure and two salt solutions on vari subjected to various constant manure prevent any manure solids from infiltrating
ous soil textures. He concluded (1) that the hydraulic heads (Fig. 1). Both dairy and from the surface and thus creating a phys
sealing seemed to reside within theorganic ical seal. The saturated hydraulic con
swine slurries were used for these trials,
mat which accumulated at the soil surface, the former containing 9% total solids (TS) ductivity of each column was measured
(2) that biological sealing mechanisms and latter 6% TS. The particle size analy initially as well as after 1 wk and 2 wk of
were insignificant and (3) that chemical sis for the coarse sand is presented in soaking. These permeability measure
mechanismsresulted perhaps from the soil Table I. ments require the infiltration of a volume
destructuring effect of the manure's neu The first trial compared sealing effi of water, representing four times that of
tral pH. ciency between physical and biological the soil cores. Thus, the cores' per
Miller etal. (1976) sampled four earthen mechanisms. Using the setup described in meabilityto water was expectedto change
manure reservoirs, two in use for 2 yr and Fig. 1, formaldehyde-sterilized slurries between trials and soaking periods. Nev
two others in use for 8 yr. Their sampling (Rowsell 1980) and natural slurries were ertheless, chemical sealing effects should
indicated significant contaminant accumu tested for infiltration rate over coarse sand produce an additional change in per
lation as deep as 150 cm below the reser columns 0.14 m in diameter by 0.15 m in meability among core groups, if signifi
voir soil-manure interface. Sampling only height. The coarse sand was assumed to cant. Analysis of variance was used as
four reservoirs, Miller could not establish demonstrate minimal chemical activity. statistical method to compare the effects of
a correlation between soil clay content or Slurry hydraulic head as well as total the two treatments, water and swine slurry
reservoir age and the quantities of solids content were maintained constant soaking.
105
CANADIAN AGRICULTURAL ENGINEERING, VOL. 29, NO. 2, SUMMER 1987
TABLE III. HYDRAULIC CONDUCTIVITY
CHANGES RESULTING FROM SOAKING
WITH SWINE SLURRY
Figure 1. Experimental infiltration columns with piezometers to measure pressure drops. RESULTS AND DISCUSSIONS
The comparison of sealingperformance
TABLE II. RELATIVE PERFORMANCE OF PHYSICAL AND BIOLOGICAL SEALING between physical and biological mecha
MECHANISMS nismsis presented inTableII. Hogslurries
Infiltration ratef
were observed to ferment heavilyfrom the
(10~9 m/s"1) 12th to 48th h of eachtrial. Thisbiological
Temperature activity was, by far, greater than that of
Assay Dairy slurry Hog slurry (C)
dairy slurry. This superiority probably
Initial (water) 72 000 72 000 0-5 resulted from the better biodegradability
Sterilized slurry or lower fiber content of the hog manure
(physical mechanism) 8.0 400
Nature slurry
0-5
solids. The chemical sealing effects were
(physical and found to be insignificant, as no significant
biological mechanisms) 6.5 70 0-5 difference was found between the
7 20-30 hydraulic conductivities of the clay col
tlnfiltration rates are reported for an hydraulic head of 0.700m above thecoarse sand column. umns soaked in hog liquids and those
soaked in water (Table III).
The use of piezometers duringthe dairy
A fourth procedure was used to investi sand columns, with two replicates each for and hog slurry infiltration trials demon
gate various physical sealing mechanisms. dairy and hog manure. Piezometers were strated pressure gradient losses directly
Manure hydraulic heads of 0.125, 0.440 used inside the columns to locate the above and at the soil-manure interface but
and 0.700 m were established over coarse sealed layers (Fig. 1). Also, total solid not below. It was therefore concluded that
Infiltration ratef
Slurry type (10-9 m/s)
Heat-treated 1040 260
Natural 120 44
t Values corrected for 700 ppm and 3000ppm of NH4 for dairy and hogexfiltrates, respectively.
tExfiltrates collected after 50 d of infiltration.
REFERENCES
BARRINGTON, S. F. 1985. The sealing of
TABLE IX. SOIL SATURATION RATEt (10-7 cm/s). Assuming a groundwater soils by manure. Ph.D. Thesis. Agricultural
FOR EARTHEN RESERVOIR PROFILE
table some 60 cm below a hog manure Engineering Department of McGill Univer
Cation saturation rate earthen reservoir, contamination would sity, Montreal, Quebec.
Soil
(m/yr) become obvious within 3.3 and 15 yr for BARRINGTON, S. F., P. J. JUTRAS, and
texture CEC* Hog manure Dairy manure sandy and clay soils, respectively. This R. S. BROUGHTON. 1987. The sealing of
soils by manure. Part I. Preliminary inves
Sand
analysis indicates a need for further
5 0.18 0.04 tigations. Can. Agric. Eng. 29: 99-103.
Clay 25 0.04 0.01 groundwater protection for soils of low DeTAR, W. R. 1977. The concentration of
cation exchange capacity (CEC), even liquid manure effects its infiltration into
tThese rates assume 100% soil saturation.
$CEC, cation exchange capacity. though infiltrations were reduced to 8.64 soils. Annual Meeting, Am. Soc. Agric.
x 10-5 m/d (IO-7 cm/s). Eng., St. Joseph, Michigan. Paper no.
Column assays throughout the project, 77-2060.
Throughout this project, exfiltrate sam indicated a final infiltration rate tending DeTAR, W. R. 1979. Infiltration of liquid dairy
ples were analyzed for nutrient and min towards 8.64 x 10~5 m/d (10~7 cm/s). manure into soil. Trans. Am. Soc. of Agric.
eralcontent. Because most of the samples Perhaps this represents the maximum seal Eng. 22:520-528,531.
were collected from coarse sand columns, ing extent of soils in contact with manure. LO, K. V. 1977. A study of the infiltration
soils of low filtration capacity, the quality characteristics of dairy waste storage
of theseliquids wasassumed equivalent to lagoons. Can. Soc. Agric. Eng. Paper
CONCLUSIONS
that seepingtowardsthe groundwaterfrom 77-208. CSAE, Ottawa, Ont.
The investigation of individual sealing MILLER, M. H., J. B. ROBINSON, and
earthen reservoirs. Because of ammonia mechanisms demonstrated the superiority D. W. GALLAGHER. 1976. Accumulation
volatilization during collection, sample of thephysical processes whereby organic of nutrients in soil beneath hog manure
cation concentrations were corrected in solids are trapped within soil pores and at lagoons. J. Environ. Quality5(3): 279-282.
Table VIII to more probable ammonia lev the soil surface. The efficiency of this MILLER, M. H. and I. B. ROBINSON. 1981.
elsof 700 and 3000 ppm fordairy and hog physical process is a direct function of soil Natural sealing of earthen manure storage
exfiltrates, respectively. These new levels void geometry as well as dimension, and ponds. Workshopon swinemanure manage
of ammonia (based upon data from the manure solids particle size distribution. ment technologies, Agriculture Canada,
Quebec Ministry of Environment) gave Biological sealing mechanisms inter Ottawa, Ontario, Canada, pp. 72-90.
final liquid cation concentrations of 60 and OVERCASH, M. R., F. J. HUMENIK, and
vene as a function of manure solids bio-
280 meq/L. Table IX illustrates the rate of J.R. MINER. 1983. Livestock waste man
degradability and ambient temperature. agement. CRC Press, Boca Raton, FL.
cation saturation for the soil of earthen This mechanism binds manure solids to ROWSELL, J. G. 1980. Infiltration of liquid
reservoir profiles subjected to a regular soil particles thus strengthening the phys manure into soil materials. M.Sc. Thesis,
infiltration rate of 8.64 x 10~5 m/d ical sealing process. Guelph University, Guelph, Ont.
108
CANADIAN AGRICULTURAL ENGINEERING, VOL. 29, NO. 2, SUMMER 1987
SOIL EROSION LOSSES UNDER FREEZE/THAW AND WINTER
GROUND COVER USING ALABORATORY RAINFALL SIMULATOR
Linnell M. Edwards1 and J. R. Burney2
'Research Station, Agriculture Canada, Charlottetown, Prince Edward Island CIA 7M8 and
Department ofAgricultural Engineering, Technical University ofNova Scotia, Halifax, Nova Scotia
B3J 2X4
Edwards, Linnell M. and J. R. Burney. 1987. Soil erosion losses under freeze/thaw and winter ground cover using a
laboratory rainfall simulator. Can. Agric. Eng. 29: 109-115.
Alaboratory rainfall simulator was used to test three Prince Edward Island agricultural soils (varying in soil texture) for
runoff and splash volume and sediment loss under varying conditions of freeze/thaw, ground cover and erosivity. Wooden
soil boxes and ancillary collection frames (termed cassettes) were designed tofit four atatime under the rainfall simulator.
With bare soil, freeze/thaw significantly increased sediment loss by about 90%: additionally, sediment in runoff varied
significantly with soil type and, for a loam soil, was 15 and 31%, respectively, ofthe amounts for a fine sandy loam and a
sandy loam. Where the soil was seeded to a winter ryecover, sediment loss was reduced by 70-80% with no significant
effect of soil type or freeze/thaw. Where, however, theerosive force was increased by adding overland flow to simulated
rainfall, there was a significant increase insediment loss even with ground cover. Sediment splash was sampled foralltests,
and only ground cover indicated a significant effect.
INTRODUCTION or near saturation during periods of snow- erosion plots on a fine sandy loam soil
In the Atlantic Provinces of Canada, melt. The cool season of PE.I. is charac between 1973 and 1977 inclusive. Yearly
as in many other parts of the world, the terized by frequent showers and occur soil loss on a 12 and 7% slope averaged 411
Universal Soil Loss Equation (USLE) is rences of soil freeze/thaw cycles; and most ha-1 and 19 t ha-1, respectively, under
routinely used to predict soil loss on agri of the soil erosion in P.E.I, seems to occur fallow conditions; 20 t ha-1 and 10 t
cultural land. However, as with any statis during this period under these predispos ha-1, respectively, under potatoes culti
tically fitted regression model, the ing conditions. vated along the slope; and 7 t ha~ *and 41
application accuracy of the USLE is con The effects of frost action on soil sta ha-1, respectively, across the slope. Fur
fined to the range and nature of its data bility have been studied to some extent thermore, erosion during the snow-melt
base. In particular, the data base used for (Logsdail and Webber 1959; Sillanpan and period of March-April contributed sub
the USLE is predominantly continental Webber 1961; Hinman and Bisal 1968; stantially to total annual erosion load, thus
U.S.A. The erosivity factor (R) is a mea Benoit 1973; Richardson 1976), and results strengthening general observations in
sure of the summed product of the kinetic have varied from no effect to an effect in P.E.I, that erosion losses are highest dur
energy (E) and the maximum 30-min either direction, viz. improved aggrega ing the cool season. Stewart (1979) con
intensity (/) within each storm period over tion or degradation depending on soil tex tends, moreover, that the low erosivity
an average year. The data base therefore is ture, initial soil moisture, freezing rate and value (R) of 50-60 projected from rainfall
heavily biased towards high intensity sum temperature. Undercircumstances of slow data alone seems inadequate to account for
mer storm conditions as a causative factor freeze/thaw cycling with high initial annual soil losses in P.E.I, which com
and takes no account of erosion caused by moisture in the soil, as typify PE.I. natu monly exceed 10 t ha_i. It would there
non-precipitation events such as snow- rally and as prevailed in this experiment, fore appear that both increased erosivity
melt runoff or the stress of alternate freez Bryan (1971) found significant soil phys forces and factors which markedly reduce
ing and thawing. ical degradation and thus susceptibility to soil erodibility resistance substantially
The applicability of the USLE under erosion. increase erosion during the cool seasons.
the light rainfall conditions of the Pacific The first documented study on soil ero None of the aforementioned studies on
Northwest of the U.S.A. was examined by sion in PE.I. (Harza of Canada Ltd. 1968) soil erosion in P.E.I, provided definitive
McCool et al. (1982) and Onstad and advocated, among other things, that the information on the quantitative impact of
Young (1982). In general it was found that erosivity factor (R) and erodibility factor cool-season meteorologic processes on
the erosivity value (R) and the cropping (K) in the USLE be independently evalu soil loss or runoff. The present study was,
value (C) required substantial modifica ated in the Atlantic Provinces. Wall et al. therefore, instituted using soil boxes and
tion. In a related sense, Morgan (1979) (1976) computed rainfall erosivity (R) ancillary collection frames (termed cas
presented European data which showed values for Canada east of the Rocky Moun settes) in a laboratory rainfall simulator to
substantial soil loss for low intensity, long tains and Wall and Dickinson (1979) and investigate the effect of freezing/thawing
duration storms on saturated soil. Wall et al. (1983) developed seasonal rain on P.E.I, soils under the influence of a cool
On Prince Edward Island (P.E.I.) soil fall distribution curves. All of these are, season erosive force and ground cover.
erosion is a major concern particularly on however, limited in applicability under the
potato lands where the crop is often culti stochastic erosivity and erodibility condi
MATERIALS AND METHODS
vated up and down the slope, and the soil tions which prevail on PE.I.
usually remains bare during the cool-sea The only available soil loss data for General
son after potato harvesting. The cool sea P.E.I, were reported by Himelman and Samples of three agricultural soils of
son is moist and the soils could easily be at Stewart (1979) based on observations of P.E.I. (Table I), taken following a barley
109
CANADIAN AGRICULTURAL ENGINEERING, VOL. 29, NO. 2, SUMMER 1987
TABLE I. SOME PHYSICAL CHARACTERISTICS OF THREE PRINCE EDWARD ISLAND were designed to fit easily over any ofthe
soil boxes during rainfall simulator runs. A
Dunblane Charlottetown fine Dunstaffnage soil box complete with splash frame and
Characteristics loam sandy loam sandy loam runoff collector, termed a soil cassette, is
Organic matter (%) 2.7 3.1 3.7 shown on the left in Fig. 4. Four splash
frames and four runoff collectors were
Textural analysis
Clay (%) 18.0 11.7 5.6 constructed so that four soil cassettes could
37.1 34.0 23.7
Silt (%)
44.4 53.3
be run at once in the rainfall simulator.
Fine and very fine sand (%)t 37.4
Sand (%) 44.9 54.3 70.7 Each splash frame was constructed
Bulk density (Mg rrr3) 1.09 1.18 1.25 from 22-gauge (0.84-mm) galvanized iron
Hydraulic conductivity 26.2 26.2 28.2 and consisted of the following main parts:
(cmrrl) (1)A continuous splash collection
Aggregate stability {%)$ 97.0 86.1 80.1 trough which ran across the top and down
tSand fractions between 250 Ltm and 50 luti. each side; the edge above this trough lying
tKemper (1965). directly above the inside edge of the box,
and the section below the trough fitting
crop, were tested in a laboratory rainfall support along the bottom and side snuggly into the inner edge of the box.
simulator for runoff volume and soil loss. plywood butt joints. The frame was firmly (2) A cover at the lower end to prevent
Two replicates of a factorial arrangement fixed to the plywood by stainless steel raindrops from falling onto the flow con
were used to test the following factors: screws, and evenly spaced 20-mm-diame- centrating section.
three soils (Charlottetown fine sandy ter drain holes were drilled in the bottom of (3) A small perforated pipe across the
loam, Dunstaffnage sandy loam and Dun each box. A sheet of plastic flyscreen over upper end to spray water against the inside
blane loam); two temperature states the inside bottom of each box prevented back edge and from there onto the soil
(frozen and unfrozen); three ground cover soil loss through the drain holes. Two sta surface. Each pipe was designed to deliver
states (bare soil, winter rye early cover ples were welded to each end of the boxes water at an equivalent rate of 150 mm h ~'
seeded 30 Sept. and late cover seeded 31 and a short length of 20-mm rope was evenly across the top end of a soil box,
Oct.); two erosion agents (simulated rain fitted to facilitate handling. effectively extending the plot length.
and simulated rain plus overland flow). The soil boxes were designed to have a The runoff collector was fitted sepa
There were 72 experimental units. 45 bevel on the overflow, lowerend (Fig. rately over the sill of the soil box. For each
1). In order to provide for a sharp raindrop run, the edges between the splash frame
Equipment separation edge for erosion tests, a con and the runoff collector were sealed with a
Seventy-two wooden soil boxes (Fig. 1) centration of runoff for collection at the silicon sealer.
were constructed from 18-mm Douglas fir lower end, collection of splash, and the As shown in Fig. 5, the basic rainfall
select plywood which was pressure-treated provision of overland flow simulation at simulator unit consisted of a 3.05-m-long
with a preservative. A frame of 50 x 50 x the upper end, a specialsplash frame (Fig. x 0.94-m-wide and 0.25-m-high soil
3-mm rust-proofed steel angle provided 2) and a separate runoff collector (Fig. 3) trough suspended beneath a matrix of
SCALE:
18 mm PRESSURE
TREATED PLYWOOD
500
230
905
NOTE:
1. ALL VALUES IN mm
2. MATERIAL: 22 gge g.i.
PLAN
GZ mr
r. 130
l END
ELEVATION ELEVATION
Figure 2. Splash frame which fits over a soil box shown in Fig. 1.
Procedure
Soil boxes were filled in the field at
three sites of different soil types. Care was
taken to maintain the original profile orien
tation of the soil. The boxes were taken to a
common outdoor site, cover crop treat
ments seeded to Kodiak winter rye at 4 g
box-' (140 kg ha- *) and left under natu
ELEVATION SILL FLEVATION
ral meteorologic conditions until mid-
Figure 3. Runoff collector which fits overthe sill of a soil box shown in Fig. 1. November when they were moved to the
laboratory. All boxes were kept at near
field capacity through regular watering.
drop-former perspex boxes. The entire hollow perspex boxes based on the design Rainfall simulator runs were done in three
apparatus was enclosed in an insulated of Chow and Harbaugh (1965). Four time blocks: first, with bare soil treat
cooling chamber. nozzles in the upper surface of each box ments; second, with early seeded treat
The rainfall drop-formers consisted of a injected water at rates equivalent to 12.5, ments; and third, with late seeded
set of 430-mm-square x 30-mm-deep 25, 50 and 100 mm h ~' over the area of the treatments. For each time block, 12 boxes
-WATER
SUPPLY
VACUUM
LINE
REFRIGERATION
UNIT
2x7 MATRIX OF HOLLOW
0.43 m SQ. PERSPEX BOXES
-RIBBED
ALUMINUM
SHEET
(WALLS AND
CEILING) -TROUGH SUPPORT ROD x 4
SOIL TROUGH
HINGE-
SCISSOR
JACK
114
CANADIAN AGRICULTURAL ENGINEERING, VOL. 29, NO. 2, SUMMER 1987
shed experimentation. J. Geophys. Union Effect of frost action on structure of Haldi- drying cycles on soil aggregation. Can. J.
Res. 70: 6111-6119. mand clay. Can. J. Soil Sci. 39: 103-106. Soil. Sci. 41: 182-187.
HARZA OF CANADA, LTD. 1968. Soil con- McCOOL, D. K., W. H. WISCHMEIER, and STEWART, N. E. 1979. Some problems asso
servation program for the Atlantic L. C. JOHNSON. 1982. Adapting the uni ciated with implementing soil conservation
Provinces. Report for Dept. of Forestry and versal soil loss equation to the Pacific north programs. Paper, Soil Erosion Workshop,
Rural Development of Canada. 60 pp. west. Trans. ASAE (Am. Soc. Agric. Eng.) NSAC, Truro, N.S. 17-19 April. 14 pp.
HIMELMAN, D. and N. E. STEWART. 1979. 25: 928-934. WALL, G. J. and DICKINSON, W. T. 1979.
Soil erosion studies in Prince Edward MORGAN, R. P. C. 1979. Soil erosion. Aca Soil erosion prediction in Canada. Paper,
Island. Paper, Soil Erosion Workshop, demic Press, Don Mills, Ont. 113 pp. Soil Erosion Workshop, NSAC, Truro, N.S.
NSAC, Truro, N.S. 17-19 April. 9 pp. ONSTAD, C. A. and R. A. YOUNG. 1982. 17-19 April. 25 pp.
H1NMAN, W. C. and F. BISAL. 1968. Altera Erosion characteristics of three north WALL, G. J., L. J. P. VAN VLIET, and W. T.
tions of soil structure upon freezing and western soils. Trans. ASAE (Am. Soc. DICKINSON. 1976. The universal soil loss
thawing and subsequent drying. Can. J. Soil Agric. Eng.) 25: 367-371. equation a method for predicting soil loss
Sci. 48: 193-197. RICHARDSON, S. J. 1976. Effect of artificial in Canada. Abstract, Annual Meeting, Can.
KEMPER, W. D. 1965. Aggregate stability. weathering cycles on the structural stability Soc. Soil Sci., Halifax, N.S. 4-7 July.
Pages 511-519 in CA. Black ed. Methods of a dispersed silt soil. J. Soil Sci. 27: WALL, G. J., W. T. DICKINSON, and J.
of soil analysis. Part 1. Agronomy 9. Am. 287-294. GREVEL. 1983. Rainfall erosion indices
Soc. of Agron., Madison, Wis. SILLANPAN, M. and L. R. WEBBER. 1961. for Canada east of the Rocky Mountains.
LOGSDAIL, D. E. and L. R. WEBBER. 1959. The effect of freezing-thawing and wetting- Can. J. Soil Sci. 63: 271-280.
Bryant, G. J., R. W. Irwin andJ. A. Stone. 1987. Tiledrain discharge under different crops. Can. Agric. Eng. 29:
117-122.
Tiledrain discharge for a 14-yr period hasbeen examined to determine theeffect of cropcoveron the flow. Discharge
characteristics wereanalyzed from 12 drainage research plots having fertilized andnonfertilized cropsin a 4-yr rotationof
corn-oats-alfalfa-alfalfa, continuous bluegrass andcontinuous corn. Fertilized cropscontributeda greater volumeof drain
flow than unfertilized crops. The majorcontribution to draindischarge occurred in Marchand April. Continuous corn and
bluegrass contributed a largerdrainage volume thanrotation crops.Theeffectof cropcoveron drainagewas the leastfor a
2-yr return period. The analysis indicates that the crop grown and level of fertility have an importanteffect on tile drain
discharge in Brookston clay loam soils.
INTRODUCTION cantly influenced by cropping sequence. results. Field tests range from 0.05 m/day
Discharge from agricultural subsurface Daily flow rates for corn preceded by oats to 5.00 m/day; laboratory tests range from
drains is of concern to many people. Intu was higher than for corn preceded by sec 0.12 m/day to 1.60 m/day, with a median
itively, one feels that the installation of ond year meadow. He attributed this to the value of 0.85 m/day.
subsurface drains must increase the vol greater evapotranspiration of the meadow
which left the soil drier the next season.
Plot Description
ume and rate of flow of water from the land
Overall, there is a lack of knowledge Twelve plots, each 12.2 m by 76.2 m,
thus creating streamflow problems such as were established in 1956. A 100-mm tile
flooding, increased erosion and scour, regarding the influence of crop cover and
fertility level on subsurface drain dis drain, spaced 12.2 m apart, was installed
damage to ecological systems, to list but a in the center of each plot. The plot layout,
few. The downstream effects of subsur charge characteristics. The objective of
this study was to determine the effect of drain gradient (%), corner surface eleva
face drainage have been a concern for tions (m), and tile flow collection system
many years. crop cover and fertility level on tile drain
discharge from a Brookston clay loam soil is shown in Fig. 2. The average depth of
Analysis of Ontario streams by McCub- drains is 70 cm which is existing practice
bin (1938) and Archibald (1939) concluded in southwestern Ontario by analyzing
in the area.
there was little indication of increased long-term tile drainage records.
Each drain discharged into an access
flooding due to agricultural drainage well which contained a flow meter equipped
activity. More recently Eddie (1982) stud EXPERIMENTAL DATA
with a linear head-discharge weir plate.
ied the annual mean, annual maximum, Soil Characteristics
The access wells were heated for year-
and annual minimum streamflows of 10 Data were obtained from an experiment around operation.
rivers in western Ontario but failed to find at the Whelan Experimental Farm (ap
any trend that might be related to agri proximately 4213'N, 8244'W) of the Crop Cover
cultural drainage. Irwin and Whiteley Harrow Research Station, Agriculture Crops grown on the plots were ran
(1982) reviewed 48 papers which dealt Canada, (Fig. 1) on Brookston clay loam. domized duplicate plots of continuous
with the effect of drainage on flow regimes. The Brookston series is a naturally poorly corn (CC), continuous bluegrass sod (B)
The authors concluded that changes were drained lacustrine soil (orthic humic and a 4-yr rotation of corn (C), oats (O),
difficult to separate from other effects and gleysol). It is an important agricultural soil alfalfa (Al) and alfalfa (A2). Superim
in many cases a reduction in flow occurred in Ontario representing 1.1 million ha. The posed on the cropping treatments were two
rather than increased flow. typical surface slope on these soils is levels of fertility, unfertilized (U) and fer
The watershed of the Middle Thames 0.05-0.10%. Physical properties of the tilized (F). Plots 1to 6 received 16.8, 29.6,
river in southern Ontario has undergone a soil are presented in Table I. Crop and and 27.9 kg/ha of N, P and K, respec
substantial increase in drainage intensity. fertility treatment effects on soil structure tively. Corn plots received an additional
Serrano et al. (1985) examined a 30-yr for the experiment described in this study side-dressing of 112kg/ha/yr of N. Plots 7
record of this river and found no evidence are presented by McKeague et al. (1987). to 12 were unfertilized. Each crop was
of appreciable change in volume of runoff This soil often cracks extensively in late grown each year. Figure 3 shows the crop
for individual storms, annual streamflow summer. Hydraulic conductivity is there ping system for the plots during each
volume, or changes in seasonal distribu fore difficult to measure with consistent month of the year.
tion of streamflow.
Two major investigations which ana TABLE I. PHYSICAL PROPERTIES OF BROOKSTON SERIES
(1969) and the Sandusky drainage plots by Bgl 16-45 21 34 45 1.58 1.0
Bg2 45-116 21 34 45 1.58 0.4
Schwab et al. (1961). Hoover and Schwab
Ckg 116 + 21 39 40 1.52 0.6
(1969) found tile discharge was signifi
Hydrologic Data
PLOT ALFALFA 1 A continuous record of tile flow from
the 12 plots was analyzed for the years
1969 through 1982. Stage-time data from
PLOT ALFALFA PLOWED
the strip charts (Fig. 4) were digitized and
stored on a microcomputer hard disk in the
Figure 3. Time distribution of cropping and tillage operations. form of discharge. The discharge was in
millimeters of depth.
Friday The digitized data were used in a com
10 XII 2 4 puter program to develop time-to-peak
flow, total event flow duration, total event
volume, peak flow and maximum flow
volume expressed as an equivalent depth
for 12, 24, 36 and 48 h. These data were
then analyzed on a seasonal and annual
basis (Bryant 1986).
Climatological data were collected at
the site as part of the Environment Canada
meteorological observation network but
were not used in this analysis.
Surface runoff was measured from four
plots for the years 1980-1982. Runoff was
negligible except for one major storm in
this 3-yr period. Water table elevations
were not measured.
ANALYSIS OF DATA
Summary of the Data
There were 1927 tile flow events during
the 14-yr period, 1969-1982, of which
1641 were over 12 h in duration. The rec
Figure 4. Tile discharge hydrograph from plot showing time periods analyzed. ord was divided into three seasons. The
Planting season was the months of April,
TABLE II. YIELD OF GRAIN CORN FROM PLOTS May and June. The Growing season was
July, August, September and October. The
Rotation corn Continuous corn
balance of the year was termed a Dormant
Fertilized Unfertilized Fertilized Unfertilized season. Table III shows the distribution of
Plant Plant Plant Plant events by duration and season.
Yield height Yield height Yield height Yield height During the analysis it became apparent
Year (t/ha) (cm) (t/ha) (cm) (t/ha) (cm) (t/ha) (cm)
that maintenance and operation of the
1969 7.76 254 3.61 190 6.66 242 0.42 137 flow-measuring instruments was not con
1970 8.04 244 5.11 222 5.86 221 0.96 165 sistent over the period of record. The
1971 7.70 235 4.76 226 5.75 230 0.87 190 major problem was in defining the actual
1972 7.80 239 6.66 233 5.81 226 1.87 196
end of an event. Due to this uncertainty in
1973 7.03 258 4.93 224 5.50 244 0.42 158
1974 4.22 193 2.64 184 3.16 173 1.13 166
the data it was concluded that the max
1975 6.65 226 4.20 217 6.22 213 2.48 167 imum time period analyzed would be 48 h.
1976 5.82 235 4.01 212 3.71 189 0.81 131 Event volumes over 48 h and seasonal and
1977 8.75 NR 3.71 NRt 6.84 NR 1.41 NR
annual volumes are not reported here due
1978 6.20 NR 2.92 NR 3.86 NR 0.45 NR
to this uncertainty.
Mean 7.00 236 4.26 214 4.72 217 1.08 164
The 24-h discharge volumes and peak
tNR, no record. flows are often used in the design of field
Woodslee, Ontario
a
9
E
3
O
>
I
CM
E
3
E
X
o
2
o
3
C
C
<
Figure 5. Annual maximum 24-h volumes for each crop (F, fertilized; U, unfertilized; B, bluegrass; CC, continuous corn; C, rotation corn; O,oats;
Al, first-year alfalfa; A2, second-year alfalfa).
E
E
a
o
E
3
0
>
c
0
E
E
X
0
2
o
3
C
c
<
Figure 6. Annual maximum event volume for each crop (F, fertilized; U, unfertilized; B, bluegrass; CC, continuous corn; O, oats; Al, first-year
alfalfa; A2, second-year alfalfa).
TABLE V. COEFFICIENTS OF VARIATION FOR RETURN PERIODS The least variation among crops occurred
24-h volumes Event volumes
at the 2-yr return period. This was
Return
Period
expected since return periods greater than
X CT CV X CT CV
2-yr are associated with more extreme
1.05 12.5 5.8 0.467 33.3 10.7 0.323 storms. Extremity of storms may also
2 25.5 4.1 0.159 75.5 15.7 0.207 explain the greater differences in the tile
5 34.8 7.3 0.209 121.5 33.0 0.272
10
discharge among crops.
43.4 12.0 0.277 161.1 50.1 0.311
at the 2-yr return period for both the 24-h channel to the drain thus increasing the REFERENCES
volumes and event volumes. The coeffi volume of water discharged. ARCHIBALD, S. W. 1939. Ontario stream
cients of variation in Table V support this Continuous corn and bluegrass gener control. Can. Eng. Apr. 9: 4.
observation since they are lowest at the ally contributed a greater tile discharge BRYANT, G. J. 1986. The effect of crop cover
2-yr return period. volume than rotation crops. on subsurface tile drain flow. M.Sc. Thesis,
It should be stressed that this study is For the bluegrass plots this may be University of Guelph, Guelph, Ont.
EDDIE, J. D. 1982. Temporal variations of
site specific. Overland runoff at the site attributed to a more favorable soil struc
southwestern Ontario streamflows. Ontario
was negligible, and thus, was neglected. ture, therefore more water storage in the Ministry of Environment, Toronto, Ont.
In areas where overland runoff is signifi surface profile, which results in a long, Water Resources Paper 15.
cant, the combined effects of overland low flow. For the continuous corn plots it GOINS, T. A. 1956. Tile flow characteristics.
runoff and subsurface drainage must be may be associated with a compact soil Agr. Eng. 37: 30-32, 35.
considered. structure which results in longer wet peri HOOVER, J. R. and G. O. SCHWAB. 1969.
ods and ponding. Legumes in a rotation Effect of tile depth, spacing and cropping
CONCLUSIONS increase the number of biopores in the practice in drain discharge. Trans. ASAE
The results indicate fertilized crops surface soil and evapotranspiration from (Am. Soc. Agric. Eng.) 9(2): 150-152.
potentially contribute a greater volume of IRWIN, R. W. and H. R. WHITELEY. 1982.
the crop. Therefore, less water may be
tile discharge to a stream than unfertilized Effects of land drainage on streamflow.
available for drainage.
Conf. on Water-Based Recreation Issues,
crops. Fertilizer appears to promote root The largest contribution of tile dis Conflicts, Strategies, Minett, Ont.
proliferation and therefore more continu charge to a stream occurred in March and KITE, G. W. 1977. Frequency and risk analy
ous biopores and cracking through the sur April, due primarily to snowmelt and a sis in hydrology. Water Resources Publica
face profile which results in a more direct greater depth of rainfall in these months. tions. Littleton Co.
Natho-Jina, S., S. O. Prasher, C. A. Madramootoo, and R. S. Broughton. 1987. Measurements and analysis of
runoff from subsurface drained farmlands. Can. Agric. Eng. 29: 123-130.
The correlation between agricultural subsurface drainage systems and peak flow rates was investigated by measuring
bothsurfaceand subsurface componentsof runoffon an 11-ha field in Ormstown,Quebec. Depth of runoff, peak flow rate
and time parameters were estimated from measured hydrographs. Longer times to peak were observed on the subsurface
flow hydrographs compared to the surface runoff hydrographs.
INTRODUCTION research site is cropped with continuous Belfort F.W. 1 portable continuous water
Drainage improvements are essential in grain corn. The field is 11 ha with an aver level recorders with a weekly clock and
the humid regions of Canada. Approx age slope of 0.2% towards the Cha chart provided measurements of water
imately, two-thirds of the land in southern teauguay River, and has two open ditches level above the weirs. The recorders were
Ontario, Quebec and the Maritime for removing surface runoff, and subsur fixed to the top of a stand protected by a
Provinces that has soil and climate suitable face drains for subsurface water. wooden shelter. From the records of head
for food production needs drainage on the weir and the calibration curve,
improvement to achieve economic crop Soil Description hydrographs of outflow versus time were
production. It is estimated that as of March Particle size analyses were carried out drawn for selected runoff events.
1981, about 332 583 ha of a total of for a number of soil samples. The distribu
2 534 540 ha in Quebec have been drained tion was found to be 12% clay, 68% silt and Climatological Measurements
by subsurface drainage systems. In 20% sand and the soil was classified as a The hourly rainfall data were measured
1980-1981, 50 404 ha of land in Quebec silt loam. by a tipping bucket rain gauge. Due to
were drained by subsurface drainage sys gauge malfunction during some storm
tems, and large amounts of monies are Drainage System events, some rainfall data were used from
spent on this drainage work every year. Thirty-two laterals, complete with the closest weather station at Ste.
Some of the rivers which receive water envelope materials, were installed in Clothilde. In some cases, the total daily
from drainage ditches have flood and pol October 1983. Observation chambers were precipitation of the Ormstown Station was
lution problems. There is a need for mea installed at the lateral/mainline connection used.
surements and theoretical analysis to of 30 laterals. The observation chambers
determine the quantitative and water allowed measurements of flow rates from Soil Physical Properties
quality effects of various land drainage individual laterals which were spaced 15 m Hydraulic conductivity was measured
practices on the flood flows and the low apart. The average depth below the soil by the single auger hole method (van Beers
flows of the rivers downstream of the drai surface of all laterals ranged from 1.14 to 1976). The saturated hydraulic con
nage works (Hill 1976; Skaggs 1982; 1.62 m. ductivity values, Ksat, ranged from 0.17 to
Madramootoo 1984). 0.29 m/day, with a mean value of 0.23 m/
Irwin and Whiteley (1983) pointed to the FIELD MEASUREMENTS day.
fact that the magnitude of flood volumes Hydrometric Measurements Soil samples were taken at different
from drained and undrained land depended Three 90-degree V-notch weirs were sites at depths of 10, 35 and 65 cm from the
on several factors including antecedent placed in the field in order to measure the soil surface. The soil moisture charac
soil moisture, rainfall intensity and loca subsurface and surface flows. One weir teristic curve was obtained by both the
tion of drainage works. More field work is was installed to collect subsurface runoff sand suction table (Rashid-Noah 1981) and
needed to quantify the effect of subsurface at the outlet of the subsurface drainage the pressure plate apparatus. The soil
drainage on surface and subsurface runoff, system. An adjacent weir measured sur moisture content at saturation (zero suc
on soil properties, drainage intensity and face runoff from the northern half of the tion) and field capacity (1/3 bar) are shown
drainage conditions. field. The third weir was installed at the in Table I.
In this paper, field measurements of southeastern corner of the field to measure Bulk density was measured by drying
some actual runoff and peak flow rates are the surface runoff. The locations of these the soil core samples after the desorption
presented, and hydrographs due to surface weirs are shown in Fig. 1. The discharges
and subsurface runoff are also compared. from the weirs were measured with a vol TABLE I . SOIL MOISTURE CONTENTS
umetric cylinder and a stop watch follow
EXPERIMENTAL FIELD ing the cessation of a precipitation event. Depth Saturation Field
(cm) (ft) capacity (ft)
Site The corresponding heads of water above
The research was done on a farm located the V-notch weir were also measured. A 10 48 43
35 46 40
approximately 3 km west of Ormstown, calibration curve of head versus discharge
65 52 31
Quebec, near the Chateauguay River. The was then plotted.
m
73
1 1 measurements were completed. The mean
bulk density was found to be 1.70 g/cm3.
8.0 2.0
SSF=SUBSURFACE FLOW
rr
4.0
0.0 12 24 12 24 12 24 12 24 12 24 12 24 HOUR
APRIL 15 APRIL 16 APRIL 17 APRIL 18 APRIL 19 APRIL 20 DATE
1.0
3.0
4.0 5.0
.3.0 6.0
Id
O
rr
<
o 20 7.0
CO
i.o-
1 i i i i i 1
0.0
12 24 12 24 12 24 12 24 12 24 HOUR
MAY 7 MAY 8 MAY 9 MAY 10 MAY II DATE
10
2.0
PERIOD^ MAY 29-31, 1984
(STE-CLOTHILDE RAINFALL DATA)
3.0 ?
<
rr
5.0 5.0
uj 40
CD
rr
<
x
co 30
ou
2.0
00 L- 12 24 12 24 12 24 12 24 12 24 HOUR
MAY 28 MAY 29 MAY 30 MAY 31 JUNE I DATE
TOTAL PRECIPITATION
70 2.0
4.5mm
18.2 mm
5.0mm
6.0 3.0
5.0
UJ
2 4.0
<
X
O
CO
Q 3.0
2.0
10
0.0,
4 12 24 12 24 12 24 12 24 12 24 12 24 HOUR
MARCH 27 MARCH 28 MARCH 29 MARCH 30 MARCH 31 APRIL DATE
Time Parameters
Time to peak, 7p, represents the time
interval from the beginning of the runoff
until the peak discharge rate, Qp, occurs.
Time to start, Ts, is the time interval from
the beginning of rainfall to the beginning
of the runoff. Lag time, 7,, is defined as the
time interval from the maximum rainfall
until the peak discharge rate, Qp, occurs.
Time to peak, Tp, occurs later during the
early spring storm runoff than during other
periods. Thegreater Tp values observed in
spring may be due to the water flow altered
by the high soil moisture condition or to
either the gradual release of water from the
snowpack or the remaining snow and
12 12 24 12 12 24 12 24 HOUR
rough plowed ground retarding runoff.
APR. 3 APR. 4 APR. 5 APR. 6 APR. 7 DATE Time to peak, Tp, was also longer for
Figure 7. Hydrograph for the 2-7 Apr. 1985 event
1.0
PERIOD- APRIL 18-25, 1985
(ORMSTOWN RAINFALL DATA) 2.0 |
SSF* SUBSURFACE FLOW
SR * SURFACE RUNOFF 2
z
4.0 <
UI2 24 12 24 12 24 12 24 12 24 12 24 12 24 12 24 12 24 HOUR
APR. 17 APR. 19 APR. 21 APR. 23 APR. 25 DATE
1.0
3.0
<
a:
5.0
3.0" 6.0
,2.0 h 7.0
e>
108mm
rr
<
5 1.0 8.0
CO
0.0 l/V
24 12 24 12 24 12 24 12 24 12 24 HOUR
JUNE 17 JUNE 18 JUNE 19 JUNE 20 JUNE 21 DATE
1.0
E
E
2.0
_T
<
PERIOD' JULY 15-19, 1985
(ORMSTOWN RAINFALL DATA) 30 2
5.0
3.0
r 2.0
9.8mm
1.0
0.0
24 12 24 12 24 12 24 12 24 12 24 HOUR
JULY 15 JULY 16 JULY 17 JULY 18 JULY 19 DATE
Sommerfeldt, T. G., C. Chang, T. Entz, and B. J. Lamond. 1987. Soil characteristics around mole drains installed
at different depths. Can. Agric. Eng. 29: 131-135.
A study was carried out to determine the bulk density distribution of soil around mole drains and also the smear and
conditions at the surface of the mole. Drains were installed at 50, 70 and 90 cm depths in a clay loam soil. Three water
treatments were usedon the soil: not flooded (D), flooded 1wk prior(Fl) to installation, and flooded 1day prior (F2) to
installation of the drains. The water treatments affected the conditions of the moles, the fracturing of the soil above the
moles, and the bulk density of the soilaround the moles. Thegreatest amount of smearing around the perimeter of moles
was observed in the F2 treatment and the least was in the D treatment. Visiblefracturing of the soil above the drains was
affected by depth of installation as well as by watertreatment. In 50-cm-deep drains, fracturing started at the side of the
mole and fanned out upward to the surface. In 90-cm drains there was minimal fracturing, starting at about the 50-cm
depth. Most fracturing occurred under the D water treatment, whereasfracturing was least in the F2 treatment. The mean
bulk density of the soil above the moles was smaller than that below the moles, where it was similar to that of the
undisturbed soil. Compaction of the soil around the mole, though significant, shouldnot impededrainage. Drag between
the bullet and the soil should remove segments of the smeared skin, leaving sufficient areas of unsmeared soil through
whichwatercan easilyenter the drains. Thus, waterentry intothe moles shouldnot be restricted by compactionor smear,
regardless of depth of installation or water content of the soil during installation.
INTRODUCTION drain, which affect water entry into the In previous studies (Sommerfeldt
Soil bulk density and pore size distribu drain. 1984), the moles were installed for recla
tion are determining factors for the perco If the mole is installed shallow enough, mation of a waterlogged saline soil with
lation of water through the soil (Wingate- the upward thrust should cause soil frac out regard for the critical depth and
Hill 1978). When mole drains are installed, turing to the surface. The depth at which conditions of the mole. The major concern
soil material is displaced by the bullet, the upward thrust does not fracture the soil was whether the drains functioned or not.
which exerts compressive forces in all to the surface is known as the critical Mole drainage has good potential for
directions. These compressive forces depth. To produce well-formed, stable draining fine-textured soils in southern
should increase the soil bulk density and mole channels, the moles must be installed Alberta (Sommerfeldt 1984; Sommerfeldt
reduce pore size distribution around the deeper than the critical depth (Godwin et and Chang 1986), either by itself or in
drain, which should restrict water entry al. 1981), but installation below the critical combination with subsurface tube drai
into the drain. However, because of the depth may cause soil compaction around nage. If soil compaction around the mole
shape and dip of the bullet, which is the moles, which limits water entry into is critical to water entry (Naarding 1977),
tapered from top to bottom on the leading the drains (Naarding 1977). This compac then drains installed below the critical
end and set on sufficient pitch to cause tion can persist for years. depth may not function well. As indicated
suction, the major thrust is upward (Childs
1942). With a tillage tool, such as a plow TABLE I. WATER CONTENT OF THE SOIL AT EACH DRAIN, TO DRAIN DEPTH, AT TIME
share or cultivator blade, there can be a OF INSTALLING THE MOLE DRAINS
1.65
1 Radial 2,6
Radial 3,4,5
eral, the cause is attributed to the mole
plow, which could have become slightly
misaligned over the years. At 70-cm
Radial 7
depth, fracturing began at the top of the
1.60 moles and extended upward to the suface,
fanning out with distance. The angle of the
1.55 fan was smaller than that of the 50-cm-
deep moles. Where lateral fracturing
occurred above the moles at 90-cm depth,
1.50 it started at a depth similar to that of the 70-
cm moles, but the angle of the fan was
1.45 - small or nonexistent, depending on the
water treatment.
The width of fracturing was also indi
1.40-
cated by the mound of soil deposited at
\ ground level while the drains were being
1.35 installed. The greatest amount of mound
50 cm depth ing was observed when the moles were
installed at 50-cm depth. The mounds,
1.30
TI 1 I I I 1 I I I I I I I 10-15 cm high, extended laterally 45-60
CO
cm from the shank. After the plow had
c passed through, the soil settled to leave
0
only a small mound (see surface Fig. 2,
O
50-cm depth). Mounds for the 70-cm deep
moles were similar to but smaller than
DO those for 50-cm-deep moles. For 90-cm
deep moles, the mound was limited to an
area near the shank. With the F2 treatment
there was essentially no mound.
At the 50- and 70-cm depths the soil
bulk densities in radial positions 1,2,6 and
7, at and above the center line of the
moles, were smaller than those in radial
I I I I I I I I positions 3, 4 and 5 (Fig. 3). The soil in
radial positions 1 and 7 would be in that
1.65 -. location of disturbed soil, above the crit
ical depth, while drains were installed.
Radials 3,4 and 5 were below the center of
1.60 -
the moles and below the soil disturbed
during drain installation, and the densities
1.55 - were not significantly different from the
undistrubed soil. For the 90-cm-deep
1.50 - moles the greatest bulk densities were lat
eral at the depths of the mole, along radials
1.45 - 2 and 6, while the smallest was along
90 cm depth radial position 4, at the bottom of the
1.40 mole. The mean bulk densities of the soil
"1 1 I I I I I I I I I I I I in radial positions 1 and 7 were signifi
10 60 110 160 210 260 310
cantly different ( P < 0.05) from each
Distance (mm) other, while those in positions 2 and
6 were not significantly different
Figure 4. Mean bulk density ofthe soil along radial arms, asaffected by mole depth and radial (P < 0.05) and neither were those in posi
position around the drain. tions 3, 4 and 5. The soil density in radial
positions 1 and 7, zone of visible fractur
treatment (Fig. 1). Within each water Soil disturbance caused by the moles at ing (Fig. 2), was significantly less
regime, sections of the smeared perimeter, 50-cm depth started at the sides of the (P < 0.01) than that in positions 2 and 6,
5 to 10 mm deep, had been pulled away by mole and extended upward to the ground and that in positions 2 and 6 was signifi
the bullet, leaving areas of exposed surface, fanning out with distance (Fig. 2). cantly less (P < 0.01) than that in posi
unsmeared soil. The width of the fan was greatest for the D tions 3, 4 and 5.
Visible fracturing of the soil above the water treatment and least for the F2 treat The bulk density varied along the radi
drains, more evident in the field than in the ment. The angle of the fan, from the ver als, depending on the radial position and
photographs, was affected by depth of tical, was larger on the left side, number 7 the depth of the drain. According to the
installation as well as by water treatment. position, than on the right, number 1posi bulk densities at 335 mm, the bulk density
Nemon, N. A., B. Von Hoyningen Huene, J. Gailichand, and R. S. Broughton. 1987. Subsurface irrigation and
drainage on sandy soil in southern Quebec. Can. Agric. Eng. 29: 137-142.
An experiment was conducted in 1982 and 1983 on flat sandy soil in southern Quebec to study the feasibility of
subsurface irrigation inthis region. The field experiment was designed on 10 hawith two treatments and eight replicates.
These treatments consisted of irrigated andnonirrigated maize plots. Itwas found that thesubsurface irrigation raised the
water table satisfactorily toa predefined water table position. This resulted inadequate water supply to thecroproot zone.
Subsurface irrigation treatment produced grain maize yields twice as high as the nonirrigated treatment.
INTRODUCTION MATERIALS AND METHODS ment was designed to use an existing sub
Drainage involves removal of excess A field experiment was conducted in the surface drainage system, which had laterals
surface and subsurface water by means of 1982 and 1983 growing seasons on the 29 m apart and 1.05 m deep. This system
ditches and subsurface drain pipes to farm of Mr. Leandre Charbonneau, St. was modified to give a randomized com
provide sufficient aeration in the root zone Louis Parish, Richelieu county, Quebec. plete block design with two treatments and
in a humid climate and control of salinity The soil is a St. Samuel sand. The experi eight replicates. The treatments consisted
in soils in irrigated semi-arid areas.
In some humid areas where water sur
pluses and deficits occur, drainage alone \ 1 \
> 29m
may not be able to provide suitable moisture TYP.
conditions for optimum crop growth. Sup
plementary irrigation is needed to provide
water during periods of drought to achieve
A-1
<D A-2
<S> B-2
ing region is suitable for subsurface irriga O
1958)and the United States (Renfro 1955).
The design procedure for drainage is more
established than for subsurface irrigation.
t > I -
ROAD
In recent years subsurface irrigation has
O MUNICIPAL WATER COURSE
begun to be used on field scale in North
and South Carolina and Florida in the LEGEND^ fy NUMBER OF THE REPLICATION
U.S.A. (Kriz and Skaggs 1973; Skaggs et DIVISION BETWEEN REPLICATIONS ^ _ _ ^ _ ^
al. 1972; Doty et al. 1975;Doty and Chris- -e- CONTROL CHAMBER EEEE3ZZZEE 3
tenbury 1979; Doering et al. 1982). D DRAIN OUTLETS 0 25 50 100m
A PLOTS SUBSURFACE IRRIGATION
This paper describes the first series of B PLOTS SUBSURFACE DRAINAGE ONLY
experiments to determine the feasibility of
subsurface irrigation in southern Quebec. Figure 1. Drainage system after modification and location of the plots.
=3 ROAD
MUNICIPAL WATER COURSE
expected when water was added to the
water level control chambers. It was found
LEGEND--0-CONTROL CHAMBER that some of the drain pipes were blocked
WATER TABLE PIPE
I 1 I- with sand. They had been installed in 1972
TENSIOMETERS 0 25 50 100m without being wrapped with an envelope
SUBSURFACE DRAIN material. New subsurface drain laterals,
enrobed with a knitted polyester envelope,
Figure 2. Location of water table pipes and tensiometers.
were installed in October 1982 on those six
field plots.
During irrigation in 1983, all plots
J L-
-PT-
TT accepted water adequately from the sub
120 . surface irrigation system approximately as
10 might be expected.
100 Water table observation pipes were
y\
s installed in each of the irrigated and non-
20
E
<
irrigated plots. Four tensiometers were
E 80 installed at 15-, 30-, 45-, 60-cm depths in
SOIL AVAILABLE
DEPTH WATER
30 each of the 16 plots (Fig. 2).
60 cm While irrigation was underway, mea
I IRRIGATED
surements were made of the depth to the
UJ
30 water table. The soil moisture content was
NON-IRRIGATED
-j 40 measured gravimetrically for samples
CD
<
IRRIGATED
taken at 7-, 22-, 37-, 52-, 67-cm depths to
| 20 45
NON-IRRIGATED 10 E represent the depth ranges of 0-15, 15-
E 30, 30-45, 45-60, and 60-75 cm,
respectively. The bulk density was mea
1 cr
sured at 5-, 20-, 35-, 50- and 65-cm depths
to convert moisture content by weight
basis to a volume basis. The soil moisture
o tension was measured with tensiometers.
r"-1"i 1 1 1 1 1 The amount of water added to the cham
10 20 30 10 20 30 10 20 30 bers was measured with water meters. The
JULY AUGUST SEPTEMBER
yield of maize was measured in Octoberof
Figure 3. Average remaining available water in 30- and 40-cm soil depths in irrigated and each year following the methodof Raghavan
nonirrigated plots, 1982. and McKyes (1977).
increased the soil moisture content in the Field capacity is equal to 25.7% by volumebased on a fieldcondition with approximately 600-mm suction in
30- to 45-cm soil layer, thus increasing the the pore water. Permanent wilting point is 3.1% by volume at 15 bars suction.
Thooyamani, K. P. and D. I. Norum. 1987. Performance of low-pressure center-pivot sprinkler irrigation systems in
Saskatchewan. Can. Agric. Eng. 29: 143-148.
Uniformity of water distributions under six low-pressure center-pivot sprinkler irrigation systems was measured to
evaluate the systems. For a uniform overlapped spray pattern, uniformity of water distribution as per design was estimated
from sprinkler charts for each system. This design uniformity was compared to uniformities measured under field
conditions to find the suitability of design and/or management of the systems for field conditions. It was concluded that the
low-pressure center-pivot systems were properly designed for sprinkler discharge rates, but the resulting uniformity of
overlapped spray pattern was not properly considered. Thus, there is room for improvement in the design of low-pressure
center-pivot systems.
INTRODUCTION ate towers. Thooyamani (1982) conducted application and energy consumption. The
Center-pivot irrigation systems are clas 16 field trials to measure the circular uni performance was compared to that of an
sified as low-pressure or high-pressure formity, and reported that the circular uni automated surface irrigation system. They
systems based on the operating pressure at formities were above the acceptable level reported that energy requirements of high-
the pivot. Although there is no definite of 0.8. Thus, radial uniformity was pressure and low-pressure center pivots
boundary between low and high pressure, selected to evaluate the systems. were approximately 100 and 6.50 per
in general, a system is classified as low The objectives of this research were: ha-mm of water, respectively, based on
pressure if the pivot pressure is less than (a) To evaluate the performance of 1980electrical power rates (approximately
350 kPa. center-pivot sprinkler irrigation systems 4.20 kWh). It was concluded that an effi
In spite of high capital investment, under field conditions, in terms of water cient automated surface irrigation system
sprinkler irrigation systems are replacing application uniformity. can be designed to distribute water as
surface irrigation systems because of auto (b) To evaluate the design of the center- uniformly as low-pressure center-pivot
mation and flexibility of management. Out pivot systems in terms of the uniformity. systems.
of 76 890 ha of total irrigated area in Sas (c) To compare the design uniformity Norum et al. (1983) studied the perfor
katchewan, 37% was irrigated by sprinkler to the observed uniformity to find the suit mance of low-pressure center-pivot sprin
systems in 1978. It was 3% in 1950 and 8% ability of design and/or management for kler irrigation systems in the Outlook
in 1960 and 1970 as shown in Fig. 1 field conditions. Irrigation District, Saskatchewan. Unifor
(Prairie Farm Rehabilitation Administra mity of application in both the circular and
tion (PFRA), 1982). In 1985, 70% of total radial directions was evaluated under eight
irrigated area was serviced by sprinkler REVIEW OF LITERATURE different systems. It was concluded that all
irrigation systems (Hill 1985). Murray et al. (1982) studied the perfor the systems were distributing water above
To avoid runoff and erosion a center- mance of low-pressure and high-pressure the generally accepted level of uniformity
pivot system is usually designed and oper center pivots in terms of uniformity of (0.8).
ated to apply water at an application rate
equal to or less than the infiltration rate of 110
o
gation systems is the nonuniform applica
tion of water. Thus, uniformity of applica 60 .
/ 70%
tion was selected to evaluate the systems. 50
&/
Uniformity of application in circular as
well as radial directions should be consid 40
V &/
ered for a complete evaluation of a sprin V
30 -
This paper was presentedat the annual meetingof the 1950 I960 1970 1980 1990
Canadian Society of Agricultural Engineers, 6-10 YEAR
July 1986, Saskatoon, Saskatchewan. Paper no
86-303.
Figure 1. Growth of sprinkler irrigation in Saskatchewan.
Figure 2. (a) An example of a poor distribution of water in radial direction. (System A, average RESULTS AND DISCUSSION
application = 10.2mm, Ucc = 0.54); (b) Infiltration profile under System A.
Radial Uniformity for Field Conditions
The average depths of application in the
PROCEDURE operating under normal field conditions, 27 trials ranged from 7.5 to 24.2 mm, with
except that the test periods were chosen so an overall average of 15.0 mm. Figures 2
Estimation of Uniformities under Field that the wind was relatively light. One- and 3 show examples of two radial dis
Conditions liter cans of 100 mm diameter were posi tributions.
All field tests were carried out on far tioned on stands approximately 0.90 m From Fig. 2(a), (b), it can be seen that
mer-operated center-pivot machines in the above ground level, and were spaced System A performed poorly. Only 43% of
vicinity of the Outlook Irrigation District mostly at 5.0 m, but in some instances at the area received an irrigation of average
in Saskatchewan (Murray et al. 1982; 10.0 or 15.0 m. depth or greater. Another parameter to
Norum et al. 1983). Water distributions For each trial, a weighted Christiansen evaluate the water distribution is the aver
were measured while each machine was uniformity coefficient was estimated from age depth of application in the least irri-
144
CANADIAN AGRICULTURAL ENGINEERING, VOL. 29, NO. 2, SUMMER 1987
Figure 3(a), (b) shows the performance
20. 0 of a well-designed and operated center-
Q pivot system. Even though only 45% of
UJ
the land received a greater-than-average
5 15.
Q.
depth of irrigation, the low quarter average
< was about 90% of the average depth. Max
cc imum depth of application was only one
10.0 and one-half times the average depth.
<
Thus, for System F, most of the land was
u. adequately and efficiently irrigated.
o
5. 0 Radial uniformity of application, as
X
h-
observed under field conditions for the 27
Q_
UJ
trials under the six systems is given in
_L _L _J_
Q
0
_i i i_
Fig. 4. Taking 0.8 as the acceptable uni
0. 0 100.0 200.0 300.0 400.0 formity level, the uniformities for 17trials
were below the acceptable level. Thus, out
RADIAL DISTANCE FROM PIVOT, m
of 27 random samples, only 37% of the
PERCENTAGE OF IRRIGATION AREA time did the low-pressure systems perform
0.0 25. 0 50. 0 75. 0 100.0
adequately under field conditions.
0. 0| 1 r -j 1 1 1 n 1 1 1 1 1 1 r- -r-ii r-
UJ
o
Radial Design Uniformities
<
Radial design uniformity of a system is
UJ
> the uniformity that can be obtained under
< 0.5
ideal field conditions such as no wind,
level ground surface, etc., provided the
Q
UJ
LOW
overlapped spray pattern is uniform. Thus,
~ 1.0-
_l
Q_
the design uniformity is the theoretical
AVERAGE QUARTER
Q_
<
maximum possible uniformity of applica
CC
AVERAGE tion that can be expected under a system.
1.5
UJ
Examples of radial distributions that could
be expected according to the designs for
two systems, namely System A and Sys
2. 0L tem F, are given in Figs. 5 and 6.
Figure 3. (a) An example of good distribution of water in radial direction. (System F, average Figures 5(a), (b) and 6(a), (b) show that
application = 12.0mm, Ucc = 0.87); (b) Infiltration profile under System F. efficient and adequately uniform distribu
tion of water can be achieved under the
system if a uniform overlapped spray pat
1. 0 tern can be produced. Under System A
more than 50% of the area can be expected
z 0. 9 to receive an irrigation greater than the
UJ
-
s average depth, while the maximum depth
(J 4
0. 8
_ _ _ ^_ _ of irrigation is only 1.25 times the average,
u.
u_ m and the low quarter average is 90% of the
UJ
o
0. 7
average. Under System F, more than 55%
u
of the area can be irrigated with at least the
average depth, while the maximum and low
0. 6 "*
2:
quarter average depths are only 1.25 and
CC
o
0.9 times the average depth, respectively.
0. 5
Ll Design uniformities for the six systems,
z
1 1 1
as operated on the Outlook Irrigation Dis
n a 1 1 1
trict area, are given in Fig. 7. The design
A B C D uniformity for the systems ranged from
SYSTEM IDENTIFICATION
0.90 to 0.95. Apparently all six systems
had the capability to distribute water well
Figure 4. Radial uniformities under low-pressure sprinklers.
above the acceptable level of uniformity
(0.8). Thus, a low-pressure center-pivot
gated 25% of area, known as the low the average. Thus, some areas had a sprinkler irrigation system can be designed
quarter average (Hart and Heermann severe moisture deficiency while other to distribute adequate water uniformity, if
1976). Essentially, the low quarter average areas had heavy deep percolation losses. a uniform overlapped spray pattern can be
indicates the severity of moisture defi Crop yields may be reduced in both areas produced under the system.
ciency in the least irrigated area. The low resulting in a double loss to the farmer. A
quarter average for System A was less than well-designed and operated surface irriga Comparison of Uniformities
one-half of the average. Maximum depth tion system can distribute water better than Design uniformity and uniformities
of application was two and one-half times the infiltration profile for System A. under field conditions were compared by
Seasonal Uniformity
2. O1- Crop yield is not determined by a one
time distribution of water under a pivot
Figure 6. (a) Expected radial distribution under System F. (average application = 14.8 mm,
Ucc = 0.96); (b) Expected infiltration profile under System F.
system. Seasonal uniformity of applica
tion and overall spatial distribution of
water should be considered. Such exten
1. 0 |- sive data were not collected due to time
and cost constraints. Nevertheless, one
z 0. 9 time distribution of water does influence
UJ
yield potential. For example, for System
CJ
0. 8
A, an area may be repetitively under-irri
gated or another area may be under-irri
UJ
o 0. 7
gated at different times. In any case, crops
u
are subjected to moisture stress, and yield
0. 6
may be reduced. Thus, repetitive good
z water distributions help to realize full
cr
a
potential yield.
0. 5
u_
z
0. 4 J_ _L _L _L SUMMARY AND CONCLUSIONS
A B O D E Distributions of water in radial direc
tions under six low-pressure center-pivot
SYSTEM IDENTIFICATION
sprinkler irrigation systems were mea
Figure 7. Radial design uniformities for six low-pressure systems. sured to evaluate the systems. Twenty
seven trials were conducted under the sys
tems as operated by the farmers. The
(Thooyamani 1982). If the sprinkler spac- with the high peak application rate, farm expected distributions of water under the
ings were designed based on only average ers operated the systems at high rotational systems were estimated from sprinkler
application rate, and not spray pattern speeds, resulting in shallow depths of specification charts. To evaluate the sys
shape, the resulting overlapped spray pat application. This could explain the reason tems for their suitability, the observed and
tern may be quite nonuniform. The ulti for the average depth of application to be the design distributions of water were
mate result is nonuniform application of below 10 mm in six field tests and the compared.
water with very high peak application overall average depth of application to be It was found that the low-pressure sys
rates. To avoid surface runoff associated at 15 mm. Thus, the design of sprinkler tems distributed water at or above the gen-
z
o for economic development. Proceedings
1. 0
CO
X Fourth Annual West Prov. Conf., Ration
UJ
a Q * * A mm alization of Water and Soil Research and
\
Management. Saskatchewan Water Corp.,
0. 8 Moose Jaw, Sask. pp. 233-245.
u
L>
D
+vtfft MURRAY, J. M., D. I. NORUM, and G. E.
f 'i
DYCK. 1982. Development and evaluation
a
UJ
of automated surface irrigation system.
0. 6
> Final Report on DSS Contract No. OSZ79-
cr
UJ 00061 to Canada Department of Agriculture.
CO
CD 61pp.
o 1 1 1 1 1 1 1 1 1 i i i i i i i i i i i i i i i i i i
0. 4 NORUM, D. I., J. M. MURRAY, and
1 4 7 10 13 16 19 22 25 28 S. DUBETZ. 1983. Evaluation of, and rec
ommendations for, operation of low pres
TRIAL NUMBERS sure center pivot sprinklers. Prairie Regional
HQ, Agriculture Canada, Ottawa, Ont.
Figure 8. Performance of six low-pressure systems. Publ. no. 1-4791. 136 pp.
PRAIRIE FARM REHABILITATION AD
erally acceptable level of uniformity co ciated with high peak application rates MINISTRATION. 1982. History of irriga
efficient (0.8) only 37% of the time. It was which, in turn, were due to the noncon- tion in Western Canada. PFRA, Regina,
also found that the six systems distributed sideration of spray pattern shapes during Sask.
water consistently below their design design. This is an additional factor causing THOOYAMANI, K. P. 1982. Efficiency of
(expected) uniformity level. poor uniformity. low-pressure sprinkler irrigation systems.
During the field tests, wind speeds were The potential detrimental factors caus M.Sc. Thesis, Department of Agricultural
ing the poor uniformity of the systems can Engineering, University of Saskatchewan,
in the range of 0-30 km/h. Installation of
Saskatoon, Sask.
flow regulators and excessive reduction of be controlled by design. Thus, there is
nozzle pressure (from line pressure) can room for improvement in the design of
make the systems more vulnerable to center-pivot sprinkler irrigation machines. LIST OF SYMBOLS
wind, resulting in nonuniform overlapped A = total area of irrigation (m2).
spray patterns and low uniformity of dx = depth of applicationat/th station, or at
application. Thus, wind and excessive REFERENCES base of /th sprinkler (mm).
reduction of nozzle pressure were identi HART, W. E. 1961. Overhead irrigation d = average depth of application (mm).
fied as potential factors causing the poor pattern parameters. Agric. Eng. 42(7): Q = discharge rate from /th sprinkler
354-355. (L/sec).
performance.
HART, W. E. 1972. Subsurface distribution /?j = radial location of /th station (m).
For the 27 field tests, overall average of nonuniformly applied surface waters. 5j = radial incrementrepresented by the/th
depth of application was 15 mm. The low Trans. ASAE (Am Soc. Agric. Eng.) 15(4): station (m).
average depths of application were the 656-661,666. / = duration of application per revolution
result of high rotational speed of the sys HART, W. E. and D. F. HEERMANN. 1976. (sec).
tems. The high rotational speed was due to Evaluating water distribution of sprinkler Ucc = Christiansen uniformity coefficient.
the attempt to avoid surface runoff asso irrigation systems. Technical Bulletin 128. UR = uniformity ratio.
'Departmentof Agricultural Engineering, University of Saskatchewan, Saskatoon, Sask. S7N 0W0, and
2Soil Science Section, Agriculture Canada, Research Station, Lethbridge, Alberta TIJ4BI
Received 19 November 1986, accepted 18 February 1987
Thooyamani, K. P., D. I. Norum, and S. Dubetz. 1987. Application rates and uniformity under center-pivot
sprinkler irrigation systems using spray nozzles. Can. Agric. Eng. 29: 149-154.
Uniformity of water application in circular and radial directions under various center-pivot irrigation systems using
spray nozzles was recorded inthefield onfarmer-operated systems. Forthe eight systems studied, theaverage application
was 15.2 mm and the Christiansen uniformity coefficient values were all above 80%. Acomputer model was developed to
simulate field water application distributions using single sprinkler distributions that had been determined in the
laboratory. These distributions were used to predict potential runoff rates under various operating conditions. Simulation
showed that it is possible to obtain a high uniformity coefficient with any spray pattern provided the nozzles are spaced
properly; however, the result may be an application rate that is great enough to produce runoff.
tory tests. Q_
UJ
(c) To use simulated distributions to cal Q
culateuniformityof applicationand poten
tial runoff. 0
IOO 200 300 400
Potential Runoff
Potential runoff is defined as the per
centage of total depth of water applied,
that is, applied when the application rate is
greater than the soil intake rate. For sim
ulation purposes, it was assumed that the
ponded water infiltration rate for the soil
could be described by the Kostiakov equa
0 100 200 300 400 500 600 tion
Figure 4. Predicted application distribution with nozzles operating at 3.7 m height (28.8 h per where / = infiltration rate under ponded
revolution at 100% speed). conditions, (mm/h), t = time since water
was first applied (h), and K and n are
coefficients dependent on the soil type. As
for much of the time (preferably all the
TRIANGULAR time) the application rate from the sprin
V PATTERN kler system is less than the infiltration rate,
Eq. 3 must be modified to take that into
account. The method used was that of
n ELLIPTICAL
Cook (1946) and Kincaid et al. (1969) in
PATTERN which for nonponded conditions
ACTUAL (4)
/, - /
PATTERN D,
LIST OF SYMBOLS
= depth of water actually infiltrated
0 100 200 300 400 500 600 (mm).
Dn = depth of water infiltrated under
DISTANCE ALONG A RADIAL LINE f m ponded conditions (mm).
= infiltration rate under ponded condi
Figure 9. Potential runoff along radial lines for four sprinklertypes with nozzlesat 3.7 m height tions (mm/h).
and 30% speed of rotation (28.8 h per revolution at 100% speed). = modified infiltration rate (mm/h).
K = infiltration coefficient dependent on
the soil type (mm/hn).
UCC values were above 80%. to produce runoff. The application rate can L = length of the lateral (m).
A computer program was developed to be reduced by increasing the height of the n = infiltration coefficient dependent on
simulate the distribution from a center- nozzles but this may result in increased the soil type.
pivot system using single-nozzle distribu losses due to wind drift. Results from the = pressure at point s (kPa).
tions that had been determined in the labo simulations showed that the application = pressure at the pivot (kPa).
ratory. By assuming a particular infiltra rate was essentially independent of the = pressure at the end of the lateral
tion function for the soil, it was possible to rotational speed of the system. (kPa).
Q = flow rate at point s (L/sec).
use the program to determine potential
ACKNOWLEDGMENT = flow rate at the pivot (L/sec).
runoff for various operating conditions.
Financial aid by the Government of India to S = distance along the lateral from the
Simulation showed that it is possible to the first author is hereby acknowledged. This pivot (m).
obtain a high uniformity coefficient with work was supported by funds provided by the t = time since water was first applied (h).
any spray pattern provided the nozzles are Department of Supply and Services, Govern X = s/L.
spaced properly; however, the result may ment of Canada under contract OSU81-00005 At = time offset to make / equal to /m (h).
be an application rate that is great enough to the University of Saskatchewan.
Tossell, R. W., W. T. Dickinson, R. P. Rudra, andG. J. Wall. 1987. A portable rainfall simulator. Can. Agric. Eng.
29: 155-162.
Anew portable rainfall simulator has been developed for both field and laboratory soil erosion research attheUniversity
of Guelph. Employing several low- to medium-flow-rate, full-jet nozzles in a continuous-spray design, the simulator is
capable of reproducing storm intensities ranging from 17.5 mm/h to inexcess of 200mm/h. Simulated rainfall intensity
varies with nozzle size, water pressure atthenozzle, and height ofthe nozzle above theplot surface. Uniformity of rainfall
intensity overa 1-m by 1-m plot averages 88.33% overallnozzles andsettings, with many nozzles producing averages over
90%. The simulator and support equipment are portable to the extent that they can be loaded onto a half-ton truck for
transport to the field.
INTRODUCTION downward flow spray simulators (Lyles number of new features to provide an
Rainfall simulation has become a very etal. 1969). improvement over the rotadisk system.
effective technique for assessing soil ero A number of rainfall and operational The GRS 11 is completely portable and can
sion, particle detachment, overland flow, characteristics, identified to be important be assembled at field locations with a mini
and chemical runoff (Meyer and Harmon for rainfall simulation for soil detachment mum of site disturbance. The continuous
1979; Bubenzer and Jones 1971; Turner and transport studies, have been docu spray generated by the small nozzle design
1965; Beasley et al. 1985). The benefits of mented by Bubenzer (1979), Meyer (1965, is also more physically realistic than the
using a rainfall simulation approach to soil 1979), and Mutchler and McGregor intermittent spray produced by the rotadisk
erosion research are well documented by (1979). These include: and some other simulator systems.
Meyer (1965), Neff (1979), and Molden- 1. Drop-size distribution similar to natu As noted briefly above, each rainfall
hauer (1979), with versatility being the ral rainfall given comparable rainfall simulator type meets only a subset of the
foremost advantage. Rainfall simulation intensities. desired criteria. Drop formers have a nar
provides control of precipitation charac 2. Drop impact velocity approximating row range of drop sizes, and a nonrandom,
teristics such as intensity, and spatial and terminal velocity of natural raindrops. biased, drop-size distribution. F-type
temporal frequency and duration in both 3. Rainfall intensity representing the designs are usually nonportable and are
field and laboratory studies (Pall et al. geographical region where studies are to susceptible to wind-sorting effects. Porta
1983). be conducted. bility is also a problem with many rotating
Several types and designs of rainfall 4. Uniform rainfall over the study area. and oscillating simulators. A noncon-
simulators have been proposed to meet the 5. Energy characteristics correspond tinuous spray, combined with a narrow
objectives of researchers. These objectives ing to natural rainfall for comparable range of drop sizes at all intensities are the
in turn have often dictated the type or intensities. major drawbacks of the rotating disk
design of simulator appropriate to the 6. Rainfall intensity continuous over the design. On the other hand, the continuous-
research studies. For example drop-former storm event. spray, single-nozzle simulators have not
simulators (Ellison and Pomerene 1944; 7. Storm pattern reproduction. been very useful for reproducing low
Adams et al. 1957; Chow and Harbaugh 8. Sufficient area of coverage. intensity storms, and do not readily allow
1965; Langford 1970; Black 1972) produce 9. Drop impact angle near vertical. the changing of intensity over the duration
a relatively uniform, narrow, drop-size 10. Site to site portability. of a simulated storm. However, a wide
distribution. These simulators are there ftill et al. (1983) have added that a sim range of nozzles are now available to re
fore suitable for studies of rainsplash or ulator should be efficient and simple to produce a variety of storm intensities.
particle detachment, where known homo operate to facilitate reliability. The purpose of this paper is to provide a
geneous drop-size characteristics are The Guelph Rainfall Simulator I (GRS I) summary of the design and development of
desired. Recent design improvements in was developed on the basis ofthe above list the GRS II rainfall simulator, including
sprinkler nozzles have greatly improved of criteria for both laboratory and field preliminary calibration results regarding
the drop characteristics produced by erosion research (Pall et al. 1983). It em simulated intensities and spatial unifor
nozzle-type simulators. A wide range of ploys a large-capacity wide-angle spray, mity. A thorough analysis of drop-size,
randomly-distributed drops are produced rotadisk system. The GRS I has proven to drop velocity, and energy and momentum
by such simulators, rendering the simu be very useful for laboratory studies; how characteristics will be presented in a sub
lated rainfall more comparable to natural ever, the system has not been sufficiently sequent paper.
rainfall. Devices employing nozzles portable to be readily used at various field
include: F-type simulators producing con locations. DEVELOPMENT AND
tinuous spray upward (Hart 1984; Meeu- A second rainfall simulator, desig DESCRIPTION OF THE GUELPH
wig 1969), rotating-boom and oscillating nated as the Guelph Rainfall Simulator II RAINFALL SIMULATOR II
simulators (Luk and Morgan 1981; Meyer (GRS II), has been developed. This sim In light of the merits and shortcomings
1979), rotadisk simulators (Morin et al. ulator incorporates many of the good of the various simulator types, a new con
1967; Pall et al. 1983), and continuous qualities of the original simulator, with a tinuous-spray simulator design has been
selected and developed for both field and TABLE I. SPECIFICATIONS FOR NOZZLES EMPLOYED IN GRS II DESIGN
laboratory soil erosion studies at the Uni
Flowrate capacity at
versity of Guelph. Special attention has
specific pressure (kPa) Spray angle
<
study area from outside the plot. At the according to its position within the plot
base of the plot, a galvanized steel trough area (recorded by gauge number). Average where UC is the uniformity coefficient as a
(with cover) channels runoff water and rainfall intensity over the plot was deter percentage; jc, is the deviation from the
sediment through a sieve system into a mined by the following equation: mean value m; and n is the number of
collection apparatus (Fig. 3). observations or gauges employed. Two
The simulator, water supply system, / =10 [(2 VjAf)/n) x60/rj (I) replications were conducted for each inten
generator, and all related facilities are suf sity/uniformity trial, and a third replicate
ficiently portable to be loaded onto a half- where /p is the plot-average intensity was added when observations from the first
ton truck and transported from the labora (mm/h); Vt is the volume in the ith gauge two replicates were dissimilar. Simulated
tory to field sites. (cm3); Ag is the gauge collection area rainfall intensity and uniformity measure
(cm2); / is the time of each run in (min); ments were determined for simulation runs
AREAL UNIFORMITY AND and n is the number of gauges; the coeffi of 20-min duration.
INTENSITY CALIBRATION cient 10 converts measurements from cm/h
Performance tests regarding rainfall to mm/h. RESULTS AND DISCUSSION
intensity and areal rainfall uniformity over Collection containers of the same size Rainfall intensity and uniformity were
the plot area were conducted as described were used (i.e., 108-mm-diameterby 178- observed to be dependent upon nozzle
by Pall et al. (1983). The procedure mm-high cylindrical cans); however, the size, water pressure at the nozzle, and the
involved placing collection apparatus or number of cans was reduced to 17 from 49, height of the nozzle above the plot surface.
gauges throughout the target area of and they were placed in strategic positions Each nozzle was tested using operating
application, and taking a representative within the calibration area, as shown in pressures of 48.3, 69.0 and 96.5 kPa;
sample of simulated rainfall. In the case of Fig. 4. The reduced number of containers while the distance from the nozzle to the
the GRS II, the 1-m by 1-m plot was the helped to decrease the time needed for plot was varied from 0.8 m to 1.7 m in
target area. The collection gauges were calibration by a factor of 6 for each run, increments of 0.1 m or 0.2 m. The 1/8
TOP VIEW
ft""
3-
2
Plot boundary
U 1\ Plot surface
Surface-trough
interface or 'lip'
4 Collection trough
5 Collection trough cover
6 Collection pit
7 Collection system
8 Sieves
SIDE VIEW 9 Sieve support
ring
Collection bucket
2.8-W nozzle was not tested at the 48.3 The larger nozzles (i.e. the 3/8-, 1/2-inch With the exception of a few anomalies
kPa pressure because the specified mini nozzles) provided the largest range of (i.e. the 1/8 2.8-W nozzle and 1/8 4.3-W
mum operating pressure was 69.0 kPa. The intensity with changes in height. This nozzle at 48.3 kPa), an increase in water
1/2 30-W nozzle was not tested at 96.5 kPa effect was more evident at the lower pressure at the nozzle increased rainfall
because rainfall intensities were consid heights with the rate of change in intensity intensity. The smaller 1/8-inch nozzles,
ered to be unrealistically high for Ontario decreasing with increasing height. Inten however, showed a small change in inten
conditions. Not all heights were tested for sity tended to level off with nozzle heights sity with pressure. In fact, the 1/8 2.8-W
every nozzle type because of several over above 1.7 m for the 1/8 2.8-W and 1/8 4.3- nozzle at 96.5 kft* and the 1/8 4.3-W
lapping rainfall intensities. In such cases a W nozzles, while the intensities generated nozzle at 69.0 kPa produced lower inten
coarser increment of nozzle height (i.e. by the larger nozzles leveled out at greater sities than the corresponding lower pres
0.2 m) was used. Results ofthe intensity/ heights. Results were not presented for sures. Originally it was hoped that the 1/8
uniformity trials are shown in Table II. nozzle settings above 1.7 m in height 2.8-W nozzle would produce intensities
With the exception of the 1/8 2.8-W because of the leveling out of intensities down to 10.0 mm/h, but the lowest inten
nozzle, the rainfall intensity increased with some nozzles. Furthermore, opera sity attained was 20.7 mm/h. The 1/8
with increasing nozzle size. The smaller tion of the simulator above the 2.0-m 4.3-W nozzle produced the lowest inten
flow-rate nozzles (i.e. the 1/8-inch height becomes functionally more diffi sity of 17.5 mm/h.
nozzles) showed the smallest range in cult, and the spray distribution becomes The various combinations of nozzle
intensity as a function of nozzle height. more susceptible to wind-sorting effects. size, nozzle height, and water pressure
120
100
o
**-
c 80
o
a:
60
40
20 4
0 T T~ I I I 1 1 1
0.80 0.90 1.00 1.10 1.20 1.30 1.40 1.50 1.60 1.70
Height (m)
Figure 5. Range of rainfall intensities produced by the GRS II.
90 ~
89 -
c 88 -
0
87 -
0
o
86 -
85 - Nozzle \\
c Water \\
Pressure \\
84 -
D - 48.3 kPa \\
83 - -F - 69.0 kPa Ml
0 - 96.5 kPa ^
Ave V
82 -
81 - I 1 I I I i '
1/8 2,8W 1/8 4,3W 1/4 10W 1/4 14W 3/8 20W 1/2 30W
Nozzle Size
Figure 6. Average uniformity coefficients for all nozzles used with the GRS II.
larger nozzles (i.e. the 1/2 30-W and 3/8 research needs. Toprovidea wide range of distribution which is too narrow to be con
20-W)exhibited a slight increase in unifor rainfall intensities, three nozzles will sidered as a representation of natural rain
mity with nozzle height, whereas the re cover many research requirements. The fall for the same given intensity based on
maining nozzles showed no distinct rela 1/8 4.3-W nozzle will provide relatively preliminary investigations into drop size
tionship between rainfall uniformity and low rainfall intensities, while giving rea measurements. Even with an increase in
height. sonably consistent uniformity coeffi dropvelocities associated with an increase
The 1/4 14-W nozzle had the highest cients. Intermediate intensities can be in nozzle water pressure, the increased
average uniformity coefficient of 90.78%, covered with nozzles such as the 1/4 14-W velocity effect is offset by a reduction in
and the 1/2 30-W nozzle produced the or 3/8 20-W. Higher rainfall rates, such as theproportion of larger drops which make
lowest at 81.49% (averaged over all thunderstorms, can be produced with the up a large percentage of the rainfalls ero
heights andpressures). Otheraverage uni 1/2 30-W nozzle. If there is an option, sive characteristics.
formity coefficients were 89.93, 89.47, choosing between one of two nozzles, the
CONCLUSIONS
89.34 and 85.47% for the 1/8 4.3-W, 1/4 larger nozzle is recommended because of
10-W, 3/8 20-W, and the 1/8 2.8-W the widerdrop sizedistributions produced. The Guelph Rainfall Simulator II em
nozzles, respectively. The most inconsis The simulators nozzle height above the ploys a continuous, downward-flow wide-
tent nozzle proved to be the 1/8 2.8-W at study area, should be kept within func angle full-jet nozzle spray system which is
69.0 kPa, producing uniformities which tional limits, such as 1.0-2.0 m. It is suitable for both field and laboratory
ranged from 80.45 to 92.80% (Table II). suggested, however, that if drop velocities research. A selection of low- to medium-
All other nozzles had uniformity coeffi are of importance, the greater heights are flow-rate nozzles provides a goodrange of
cientranges of less than 7%. Average uni recommended. The increase in fall height, rainfall intensities and uniformity charac
formity for all nozzles and settings was will increase individual drop velocities. teristics. The simulator is completely por
88.33%. Operational nozzle water pressure table for field experimentation, and can be
In terms of uniformity of rainfall should also be kept within certain limits. transported in a half-ton truck. Thecontin
application, the GRS II outperformed its Pressures ranging from 45 to 100 kP& are uous-spray, random-drop distribution pat
predecessor, the GRS I, by approximately recommended. The smaller nozzles (1/8 tern is superior to the intermittent spray
4% on average. The uniformities of ap 2.8-W and 1/8 4.3-W) will not provide a pattern providing a closer representation of
proximately 90%are among the highest of full 120 spray angle and may yield low the natural rainfall phenomenon.
any presented in the literature. uniformity coefficients if nozzle pressures The simulator design provides a flex
The GRS II rainfall simulator has the are below 45 kPa. Conversely, pressures in ibility that renders it suitable for a wide
potential to accommodate a wide range of excess of 100 kPa will produce a drop size variety of research applications. To sup-
161
CANADIAN AGRICULTURAL ENGINEERING, VOL. 29, NO. 2, SUMMER 1987
plement the great range of nozzle sizes, 309-329. in northern Utah. J. Range Manage. 23:
combinations of nozzle heights above the BUBENZER, G. D. 1979. Inventory of rainfall 185-188.
study surface and nozzle water pressures simulators. Rainfall Simulator Workshop MEYER, L. D. 1965. Simulation of rainfall for
Tucson, Arizona., Agric. Review Manual soil erosion research. Trans. ASAE (Am.
will provide other desirable rainfall
ARM-W-10. pp. 120-130. Soc. Agric. Eng.) 8: 63-65.
characteristics. However, the combination
BUBENZER, G. D. and B. A. JONES, Jr. MEYER, L. D. and W. C. HARMON. 1979.
of such variables will ultimately depend on 1971. Drop size and impact velocity effects Multiple intensity rainfall simulation for ero
the objectives of individual research on the detachment of soils under rainfall sim sion research on row sideslopes. Trans.
projects. ulation. Trans. ASAE (Am. Soc. Agric. ASAE (Am. Soc. Agric. Eng.) 22:100-103.
Eng.) 14: 625-628. MEYER, L. D. 1979. Methods of attaining
ACKNOWLEDGMENTS CHOW, V. R. and T. E. HARBAUGH. 1965. desired characteristics in rainfall simulation.
The authors would like to thank D. Beals for Raindrop production for laboratory water Rainfall Simulator Workshop Tucson, Ari
his constructive suggestions with the initial shed experimentation. J. Geophys. Res. 70: zona. Agric. Review Manual ARM-W-10.
simulator design. Special appreciation is 6111-6119. pp. 35-44.
extended to B. K. Hohner, I. J. Shelton, CHRISTIANSEN, J. E. 1942. Irrigation by MOLDENHAUER, W. C. 1979. Rainfall simu
E. E. Vaughan and B. G. Johnson for their sprinkling. University of California Agric. lation as a research tool. Rainfall Simulator
helpful suggestions regarding the development Exp. Sta. Bull. 670. Workshop Tucson, Arizona. Agric. Review
of the simulator and the calibration procedure. ELLISON, W. D. and W. H. POMERENE Manual ARM-W-10. pp. 90-95.
Development funds were provided by the 1944. A rainfall applicator. Agric. Eng. 25: MORIN, J., D. GOl DBERG, and I. SEGINER.
Ontario Ministry of Agriculture and Food, the 220. 1967. A rainfall simulator with a rotating
Natural Sciences and Engineering Research HART, G. E. 1984. Erosion from simulated disk. Trans. ASAE (Am. Soc. Agric. Eng.)
Council and Agriculture Canada (Great Lakes rainfall on mountain rangeland in Utah. J. 10: 74-79.
Working Group). Soil Water Conserv. 39: 330-334. MUTCHLER, C. K. and K. C. McGREGOR.
LANGFORD, K. J. 1970. A review of designs 1979. Geographical differences in rainfall.
REFERENCES for simulators. University of Melbourne. Rainfall Simulator Workshop Tucson, Ari
ADAMS, J. E., D. KIRKHAM, and P. P. Agric. Eng. Rep. 17. zona Agric. ReviewManual ARM-W-10. pp
NEILSEN. 1957. A portable infiltrometer LUK, S. andC. MORGAN. 1981. Spatial vari 8-16.
and physical assessment of soil in place. Soil ations of rainwash and runoff within appar NEFF, E. L. 1979. Why rainfall simulation?
Sci. Soc. Am. Proc. 21: 473-477. ently homogeneous areas. Catena 8: Rainfall Simulator Workshop Tucson, Ari
BEASLEY, D. B., E. J. MONKE, E. R. MIL 383-402. zona.,- Agric. Review Manual ARM-W-10.
LER, and L. F. HUGGINS. 1985. Using LYLES, L., A. DISRUD, and W. P. WOOD pp. 3-7.
simulation to assess the impacts of con RUFF. 1969. Effectsof soil physical proper PALL, R., W. T. DICKINSON, D. BEALS,
servation tillage on movement of sediment ties, rainfall characteristics and wind and R. McGIRR. 1983. Development and
and phosphorus into Lake Erie. J. Soil Water velocity on clod disintegration by rainfall calibration of a rainfall simulator. Can.
Conserv. 40: 233-237. simulation. Soil Sci. Am. Proc. 33: Agric. Eng. 25: 181-187.
BLACK, P. E. 1972. Hydrograph responses to 195-205. TURNER, A. K. 1965. The simulation of rain
geomorphic model watershed characteristics MEEUWIG, R. O. 1969. Infiltration and soil fallsfor studiesin overland flow. J. Inst. Eng.
and precipitation variables. J. Hydrol. 17: erosion as influenced by vegetation and soil Austr. 37:9-15.
Rousseau, A., W. T. Dickinson, and R. P. Rudra. 1987. Evaluation of best management practices to control
phosphorus nonpoint sourcepollution. Can. Agric. Eng. 29: 163-168.
The Guelph model for evaluating the effects of Agricultural Management Systems on Erosion, Sedimentation and
Phosphorus yields (GAMESP) has been used toevaluate potential reductions in downstream sediment and phosphorus
loadings from agricultural land when best management practices are applied on problem areas of a small Ontario
agricultural watershed. This paper introduces a brief overview of different means of controlling phosphorus nonpoint
source pollution, and reveals that best management practices using a crop residue cover such as no-till and spring
moldboard plowing can potentially reduce downstream phosphorus loadings, attributed tosheet and rill erosion during the
hydrologically active spring period byas much as 83 and 74%, respectively, when compared to fall moldboard plowing
without a crop residue cover.
INTRODUCTION pollution is to select and use a deter load by the amount of soil-derived particu
In 1970, the International Joint Com ministic watershed model. GAMESP, the late phosphorus in overland runoff, using
mission reported to the Governments of Guelph model for evaluating the effects of the average total phosphorus con
Canada and the United States that 72% of agricultural management systems on ero centration ofthe surface soil (10 mm), the
the phosphorus inputs to Lakes Erie and sion, sedimentation and phosphorus yields potential soil loss, the sediment delivery
Ontario was from municipal sources, 7% was developed by Rousseau et al. (1985) ratio, and the phosphorus enrichment
from industrial sources, and 21% from and has been used for the delineation of ratio. This concept is based on the hypoth
agricultural sources. In the past 15 yr, seasonal source areas of sediment and esis that most, if not all, ofthe phosphorus
since phosphorus inputs from municipal phosphorus attributed to sheet and rill ero carried in overland runoff is derived from
and industrial sources have been reduced sion in small agricultural watersheds. The soil particles detached from the land sur
by advanced waste water treatments, model is easy to use and makes use of face by sheet and rill erosion. The result
phosphorus from agricultural sources limited amounts of readily available data ing seasonal phosphorus load at the outlet
(nonpoint sources) have received a great bases and of easily quantifiable physical ofthe watershed is calculated by summing
deal of attention. Phosphorus is recog and hydrological watershed charac up all predicted cell-to-stream phosphorus
nized to be the nutrient limiting the pro teristics. The model is based on the dis loads.
duction of plant biomass in surface waters cretization of a watershed into land and This paper introduces a brief overview
(Lee 1973; Lee et al. 1978); hence, non- stream cells of various shapes. Each land of different means of controlling phos
point sources of phosphorus have been cell is a field-sized cell with homogeneous phorus nonpoint source pollution, and
identified to be the main cause of surface characteristics of land use, soil type and presents an application of GAMESP to
water pollution (Myers 1986). Inputs of class of slope. This modelling approach evaluate the potential reductions in down
phosphorus from agricultural sources have offers an excellent means to define and stream sediment and phosphorus loadings
been observed to be mainly associated evaluate the implementation of different from cropland areas when selected best
with sediments (Miller et al. 1982; Nelson environmental management strategies on management practices (i.e. a combination
and Logan 1983) carried by cropland specific sites of an agricultural watershed. of land management practices) are applied
runoff during the spring period (Coote et The GAMESP model consists of two on problem areas of a small Ontario agri
al. 1982), consequently, it is expected that major components: the potential soil loss/ cultural watershed.
land management practices prescribed to sedimentation component and the phos
reduce sediment loadings to surface waters phorus component. The potential soil loss/ BEST MANAGEMENT PRACTICES
would reduce phosphorus nonpoint source sedimentation component estimates for Land management practices that can be
pollution (U.S. Army Corps of Engineers each land cell the seasonal potential soil used to control phosphorus nonpoint
1982). However, since sediment loadings loss, using a seasonal application of the source pollution and that are socially and
are spatially and temporarily variable, it universal soil loss equation, and the subse economically acceptable are termed "best
becomes a major task to identify which quent seasonal amount of sediment deliv management practices" (BMP). The types
erodible land in an agricultural watershed ered to the adjacent downstream land cell of best management practices that are
requires alternative land management and to the downstream stream cell, using a technically feasible and socially and eco
practices from the perspective of down seasonal sediment delivery ratio function. nomically acceptable are limited. Best
stream water quality impacts (Ogg and Based on the assumption that no instream management practices have been grouped
Pionke 1986). sedimentation processes occur, the sea by Baker and Johnson (1983) into those
One cost-effective approach to identify sonal sediment load at the outlet of the which reduce: (i) the volume ofthe carrier
seasonal source areas of sediment-associ watershed is estimated by summing up all (amount of sediments or runoff volume),
ated phosphorus and to investigate the predicted cell-to-stream sediment loads. (ii) the phosphorus concentration in the
impact of alternative land management The phosphorus component predicts for carrier, or (iii) the delivery from the field
practices on phosphorus nonpoint source each land cell the seasonal phosphorus to surface waters.
/<')
Water Survey (
of Canada Downie Township \
Stratford/Avon
Station 02GD01& '' . Demonstration
South Easthope TownshipL_-Watershecj
0 1 2 3 4 5 km.
Figure 1. Location of the Stratford/Avon Demonstration Watershed in the Avon River Basin.
(After Stratford Avon River Environmental Management Project 1983.)
METHODOLOGY
To explore the potential impact of best
management practices during the hydro-
logically active spring period in South
western Ontario, GAMESP was first used
to identify the source areas of sediment INTERVALS C t ]
and phosphorus delivery to surface waters
000-0.49
of the Stratford/Avon Demonstration
Watershed for the 1983 spring season 0.50-1.24
(Rousseau et al. 1985). The application of 1.25-2.49
GAMESP involved: (i) the discretization 2.50 -4.99
of the study watershed into land and >5.00
stream cells, (ii) the determination of the
drainage network of each cell to the water
shed outlet, (iii) the quantification for each
cell ofthe input parameters required by the
model, and (iv) the estimation of sediment
and phosphorus loads at the watershed out
let (attributed to sheet and rill erosion from
cropland for the season under considera
tion). From water discharge, sediment and Figure 2. Spatial variability of cell-to-stream sedimentyield in the Stratford/Avon Demonstra
phosphorus data collected at the watershed tion Watershed for the 1983 spring conditions.
agement systemson erosion andsedimenta trol of diffuse phosphorus loads to Lake ROUSSEAU, A. 1985. Development of a
tion, user's manual. School of Engineering, Erie. LEWMS Technical Report, U.S. phosphorus transport model for small agri
Army Corps of Engineers, Buffalo, N.Y. cultural watersheds. M.Sc. Thesis. Univer
University of Guelph, Guelph, Ont.
COOTE, D. R., E. M. MacDONALD, W. T. McDOWELL, L. L. and K. C. McGREGOR. sity of Guelph, Guelph, Ont.
DICKINSON, R. C. OSTRY, and R. 1984. Plant nutrient losses in runoff from ROUSSEAU, A., W. T. DICKINSON, and
FRANK. 1982. Agriculture and water conservation tillage corn. Soil Tillage Res. R. P. RUDRA. 1985. A phosphorus trans
quality in the Canadian Great Lakes Basin. 4:79-91. port model for small agricultural watersheds.
I. representative agricultural watersheds. MILLER, M. H., J. B. ROBINSON, D. R. ASAE Paper No. 85-2526.
J. Environ. Qual. 11:473-481. COOTE, A. C. SPIRES, and D. W. STRATFORD AVON RIVER ENVIRON
DICKINSON, W. T., R. PALL, and G. J. DRAPER. 1982. Agriculture and water MENTAL MANAGEMENT PROJECT.
WALL. 1983. Identification of soil erosion quality in the Canadian Great Lakes Basin. 1983. Physical characteristics and land uses
and fluvial sediment problem. ASAE Paper III. phosphorus. J. Environ. Qual. 11: of the Avon River drainage basin. Technical
NAR 83-204. 487-493. Report R-9. Upper Thames River Con
DREW, M. C. and L. R. SAKER. 1978. MUELLER, D. H., B. J. ANDRASKI, T. C. servation Authority, London, Ont.
Effects of direct drilling and ploughing on DANIEL, and B. LOWERY. 1983. Effect of SHEAR, G. M. and W. W. MOSCHLER.
root distribution in spring barley, and on the conservation tillage on runoff water quality: 1969. Continuous corn by the no-tillage and
concentrations of extractable phosphorus total, dissolved and algae-available phos conventional tillage methods: a six year
and potassium in the upper horizons of aclay phorus losses. ASAE Paper No. 83-2535. comparison. Agron. J. 61: 524-526.
soil. J. Sci. Food Agric. 29: 201-206. MYERS, P. C. 1986. Nonpoint source pollu TRIPLETT, G. B. Jr. and D. M. VAN
DREW, M. C. and L. R. SAKER. 1980. Direct tion control: the USDA position. J. Soil and DORAN Jr. 1969. Nitrogen, phosphorus and
drilling and ploughing: their effects on the Water Cons. 41(3): 156-158. potassium fertilization of non-tilled maize.
distribution of extractable phosphorus and NELSON. D. W. and T. J. LOGAN. 1983. Agron. J. 61: 637-639.
potassium, and of roots, in the upper hori Chemical processes and phosphorus trans U.S. ARMY CORPS OF ENGINEERS, BUF
zons of two clay soils under winter wheat port. Chapter 5 in F. W. Shaller and G. W. FALO DISTRICT. 1982. Lake Erie Man
and spring barley. J. Agric. Sci. 94: Bailey, eds. Agricultural management and agement Study, Final Report, Buffalo, N.Y.
411-423. water quality. Iowa State University Press, VAN VLIET, L. J. P., G. J. WALL, and W. T.
KETCHESON, J. W. 1980. Effect of tillage on Ames, Iowa. DICKINSON. 1978. Erosion losses and sed
fertilizer requirements for corn on a silt loam OGG, C. W. and H. B. PIONKE. 1986. Water iment delivery ratios for agricultural water
soil. Agron. J. 72: 540-542. quality and the new farm policy initiatives. sheds. I.J.C. Tech. Rep. re: Great Lakes
LEE, G. F. 1973. Role of phosphorus in J. Soil Water Cons. 41(2): 85-88. Pollution from Land Use Activities, Wind
eutrophication and diffuse source control. O'HALLORAN, I. P., M. H. MILLER, and sor, Ont.
Water Res. 7: 111-128. G. ARNOLD. 1986. Absorption of Pby corn WISCHMEIER, W. H. and D. D. SMITH.
LEE G. F., W. ROST, and R. A. JONES. (Zea mays L.) as influenced by soil distur 1978. Predicting rainfall erosion lossesA
1978. Eutrophication of water bodies: bance. Can. J. Soil Sci. 66: 287-302. guide to conservation planning. Agricultural
Insight for an age-old problem. Environ. OLOYA, T. O. andT. J. LOGAN. 1980. Phos Handbook No. 537, USDA, Washington,
Sci. Techcl. 12: 900-908. phate desorption from soils and sediments D.C.
Research Station, Agriculture Canada, P.O. Box 20280, Fredericton, New Brunswick E3B 4Z7
Received 28 August, 1986, accepted 16 March, 1987
M^on^QG;^'and C*D'McLeod- mi- Evaluation ofpotato-stone-clod separation mechanisms. Can. Agric. Eng.
^v: lo" l/I.
Methods of separating potatoes from stones and soil clods in storage were evaluated and factors affecting performance
were determined. The three methods ofseparation included a rotating brush separator, an electronic separator and an
impact roller separator. All separators effectively sorted the stones from the potatoes but the brush separator failed to
differentiate between the soil clods and potatoes. Parameters affecting the performance ofthe brush separator were the
feed rate and the brush gap while the performance of the electronic and impact roller separators was affected by the feed
rate and the stone/potato ratio, respectively.
INTRODUCTION ejected from the flow by means of high The objective of the study described in
Thedevelopment of mechanisms to sep speed pneumatic hammers. The basic pro this paper was to quantify the performance
arate potatoes from stones and clods has cess is similar to the roller with the advan of the three types of in-storage separators
been directed towards mechanical devices tage that the impact differences are in terms of their capability to separate
located on the harvesters. These devices amplified electronically as opposed to potatoes from stones and soil clods.
have been designed to separate potatoes on creating larger differences by changing the
the basis of physical differences between contact angle on the roller. Larger dif TEST PROCEDURES
the potato and the associated material. The ferences increase separating sensitivity. The laboratory arrangement included a
differences include rolling resistance, One disadvantage is having to align the feeding conveyor on which the products
elasticity, specific gravity, terminal product in order to optimize the perfor were placed and fed at controlled rates
velocity, resistance to X-ray penetration, mance of the electronic and impact ham onto each of the three separators. The
optical reflectance, and vibrational fre mer systems. A third method of separation three separators are shown schematically
quency response. Sides and Smith (1970) which is less expensive is presently at the in Figs. 1, 2 and 3. Russet Burbank
describe the mechanisms available for sep experimental stage. This method utilizes a potatoes were distributed evenly at a
arating potatoes and stones. Methods transaxial rotary brush separator in which known weight on the conveyor and then
which have been used or investigated over the material flow is across the brush axis. either stones or clods were uniformly
the past several years have also been Separation is achieved by a differential mixed with the potatoes. The conveyor
described by Story and Raghavan (1976). deflection of the bristles caused by the speed was constant during the tests while
More recently several mechanisms have difference in specific gravities of the the weight of material per unit length on
been developed to operate in storages dur potatoes, stones and soil clods as the conveyor was varied to deliver various
ing postharvest potato handling. The described by Eaton and Hansen (1969). loading rates to the separators. Prelimi-
advantage of doing a portion ofthe separa
tion at storage rather than on the harvester
is to reduce the number of people working
on the harvester under unfavorable condi
tions. Often in Eastern Canada from two to
four people are required to sort the ma
terial on the harvester. Two approaches
for in-storage use appear promising and
have been developed into commercial ma
chines. A commercial separator developed
by Feller et al. (1984) is based on the Figure 1. Cross-sectional view of brush separator.
difference between the coefficient of res
titution of the products being separated.
The device involves bouncing the objects DISTRIBUTION
Brunswick conditions. The field stone/ were conducted at four feed rates of 9.0, Pp + Pr sr +
potato ratio depends on the soil type as 18.0, 27.0 and 36.0 t/h over a 1.3 m width pr
well as to the extent of the stone removal with three stone/potato ratios of 0.01, (l -z)d
Pp + Pr sr + SD (1)
operation. The clod/potato ratio has a 0.055 and 0.1. In order to measure the
higher variability depending on soil type effect that moisture content of soil clods where Pp = potatoes in the product exit;
and moisture as well as tillage practices. going into storage in early fall and leaving = potatoes rejected with the stones;
The feed rate selected during the evalua during winter has on the performance of = stones in the reject exit; and
tion of the separator represents sufficient the electronic separator, soil clods with s = stones remaining with the product.
capacity to handle the output from two
harvesters under New Brunswick condi
tions. A large percentage of growers oper Brush Gap
Brush Separator L
O
"' ^^^><^_
The critical parameter affecting ma -H
O
. 9
"^^^ -
chine performance is the gap between CD
li
.8
adjacent pairs of brushes. Eaton and en
"*^-
10.0, 13.0 and 16.0 mm. Due to the small Feed Rate t/h
171
CANADIAN AGRICULTURAL ENGINEERING, VOL. 29, NO. 2, SUMMER 1987
SIMULATION OF ADSORPTION DRYING OF CORN, WHEAT,
BARLEY AND OATS USING SODIUM BENTONITE
Watts, K. C. and W. K. Bilanski. 1987. Simulationof adsorptiondrying of corn, wheat, barley and oats using sodium
bentonite. Can. Agric. Eng. 29: 173-178.
Bentonite is an abundant clay mineral that is suitable for the drying of various grains by adsorption. The absorbency
characteristics are a function of the composition of the clay and the adsorbed ions. This paper presents a mathematical
model for adsorption drying of various grains using existing equations for drying individual kernels at low temperatures.
The equilibrium moisture content ofthe grain was derived from accepted data in the literature. It was necessary to obtain
the adsorption characteristics of the specific sodium bentonite being utilized using clay /grain mixtures. Comparison of
simulation results with experimental data showed satisfactory agreement when a mass ratio of grain to clay, which was
derived by imposing a moisture balance on the experimental data, was used in the simulation.
INTRODUCTION mixed. This mixture can be analyzed as a bentonite would rapidly adsorb any
One major source of energy consump three-stage process: moisture that was on the surface of the
tion on the farm is the drying of grains by (1) The moisture movement out of the grain. Cereal grains are normally consid
conventional systems which force natural grain into the air which is assumed to ered to dry during the falling-rate period.
or heated air through the bed of material to surround the grain. This is due to the reasonably low moisture
be dried. In the search for alternate meth (2) The moisture adsorption by the clay content at which grains are harvested for
ods of drying, three requirements were from the air. optimal operation of machinery and mini
determined: (1) low conventional energy (3) The resultant relative humidity ofthe mal seed damage. Hence, all moisture
usage; (2) the material should not be sub air due to the resultant moisture content of relationships noted assume that only fall
jected to high temperatures; (3) the dryer the clay. ing-rate drying is occurring where
must be available at all times, independent When the clay is dry, the resultant rela moisture diffusion in the grain is the limit
of weather or season. tive humidity will be low causing a greater ing moisture transfer mechanism.
The only method of drying that meets potential for extracting moisture from the The diffusion process has been model
all three requirements is that of desiccant grain. As the clay adsorbs more moisture, led by various workers (Ficks second law
drying in which the moisture which is the relative humidity increases and it is of diffusion; Graham et al. 1983; Keey
removed from the grains is temporarily less able to extract moisture from the 1972) and may be written, on simplifica
adsorbed onto the surface ofthe desiccant. grain. Since the relative humidity sur tion for constant diffusion coefficient, D,
Low-grade forms of energy, e.g. solar, rounding the grain particles changes with as
could be used to dry the desiccant when time, it is not possible to directly apply
= DV2M (1)
they are available so that the dryer is avail conventional drying equations that assume dt
able at any time throughout the year. a drying medium which has a constant
where / is time and M is the decimal
The desiccant that was identified by relative humidity on entering. On the con
trary, it is necessary to analyze this process moisture content (db) ofthe individual par
Sturton et al. (1981) and Graham et al.
(1983) as having the most potential is ben with time as outlined below.
ticle. Equation 1can be solved for different
tonite, a widely distributed clay mineral shapes (Crank 1975) or as an infinite series
that is relatively inexpensive and yet has a which is claimed to be independent of the
Grain Moisture Relationships
relatively high adsorption rate and capac size and shape of the drying object:
Conventional drying of grain is gener
ity. A comparison of drying corn using ally accomplished using deep-bed drying MR = 2 Ate1^ (2)
sodium and calcium bentonite was re which is somewhat different than if the n = I
ported by Watts et al. (1986). kernels were dried separately. In this paper
The objectives of this study are to pre or for spherical coordinates
each kernel is envisaged as being isolated
sent the mathematical models used in cal from all others and surrounded by the dry b V 1
culating the simultaneous drying curve of ing medium. MR 2, e
tt" ., = . n~
(3)
the grain and the moisture adsorption A general drying process is considered
curve ofthe clay and to compare the results to comprise a constant-rate drying period and
to experimental data (from the literature) and a falling-rate drying period. The con MR = (M - Mcq)/(A/0 - Mcq) (4)
obtained when using sodium bentonite to stant-rate drying period occurs when the
dry four different types of grain. internal resistance to moisture transport is where D = diffusion coefficient (m2 s~'
much less than the external resistance to or m2 h ~ l),M = volume average decimal
MATHEMATICAL MODEL water vapor removal from the grain sur moisture content (db), t = drying time
Throughout this paper, it is assumed face. This does not occur in the drying (hours or seconds), M0 = original
that the clay and grain are intimately process under consideration here since the moisture content (db), Meq = equilibrium
corn has been experimentally determined moisture and the equilibrium moisture to Meqoa(s = 0.0022 + 0.5548 RH -
(Brooker et al. 1974) as 1.0911 RH2 + 0.8481 RH3 (19)
the difference between the initial moisture
Kam =0.54 exp (- 5023\
(7)
and the equilibrium moisture (Eq. 3). It might be argued that, due to the tran
Thus, the value of the equilibrium sient process, a varying equilibrium
where 0 is in R and K has units of s~ K moisture content must be evaluated to use moisture content is unrealistic. It should
Values of the drying constant k for wheat the above equation. However, the equi be noted, however, that at the start of the
and barley are (O'Callaghan et al. 1971) librium moisture content is itself a func process the final equilibrium moisture con
tion of the relative humidity of the air tent of either the grain, clay or air is
surrounding the grain. Since the relative unknown. Also, because the kernels are
*whea, =2000 exp (-^2) (8) humidity ofthe air changes throughout the separated from one another by the clay,
drying period as the clay adsorbs moisture, equilibrium moisture contents are estab
. lloo the equilibrium moisture and drying lished quickly.
^barley = 139.2 CXp /\~ ~Z
7976\
) (9)
.14
These equations, as well as ones noted
below, will be used in predicting data for rs
the grains studied. The assumption inher
ent in using these equations, as well as
8
w
12
VISUM. CURVE FIT
ones noted below, is that the predominant
mechanism is diffusion within the kernel, 2SP . 1
UJ
independent of what mechanism is being l-
z
used to dry the air surrounding the grain Q
(e.g. moving air or desiccant). .08 * *
Because the above equations have UJ
C
/
sometimes been found to be unsatisfactory 3
.06
for conventional drying, workers have
CO :/ 0ATA FROM STURTQN t BILANSKI (I860)
proposed empirical drying equations for -*
o
cereal grains. For example, Becker (1959) .04
proposed the following drying rate equa
tion for wheat:
or MR = 0.509 exp (-629.68Dt) 0 SO 100 ISO 200 250 300 350 400
for (D/),/2 > 0.034 (ID
TINE (HOURS)
u D
where r = 76.8
n^o exp II H080A1. (12) Figure 1. Wetting curve for sodium bentonite derived by placing the clay above a standard salt
V 6abs ' solution.
/
Q) V /
N /
-
/
#
, <*. / /
about 180 h, well beyond the time when
equilibria of the grains and clay were
A0/ /
/WB
established. It was thought that the
/ hygroscopicity characteristics of the
;
/ wV M
grains were better established than those of
the clay because of the variation of the
/ /
.06 - chemical make-up and adsorbed ions of
the clay. With this assumption, it is possi
PURITY = .78-. 11 RH +. 28 RH2 ble to determine the equilbrium relative
humidity of the air surrounding both the
.04 1 i 1 i 1 . 1 i, 1 i
clay and the grain by substituting the final
.4 5 .6 7 8 9 experimental moisture content ofthe grain
RELATIVE HUMIDITY into the appropriate equation (Eqs. 16-19).
Knowing the equilibrium moisture content
Figure 3. Moisture isothermdatafor sodium bentonite determinedfrom grain drying experiments
of Sturton and Bilanski (1980) and Graham et al. (1983).
of the clay, it is then possible to derive a
moisture adsorption isotherm for the clay
(Fig. 3). The individual data points are
noted for the different grains by their first
letter. A fairly consistent moisture iso
DRYING CORN WITH SODIUM BENTONITE
therm is obtained, although some scatter is
J.25 evident. Also plotted on the graph are lines
of constant "purity", where purity is
2.2 defined as the ratio ofthe relative humidity
predicted from the data of Keren and
O # * # * Shainberg (1979) and a given relative
".IS humidity at the same moisture content. It
UJ can be seen that the grains that equilibrated
x "
| at higher relative humidity (>0.7) follow a
o> . 1 X 0.85 purity line, and those that equili
brated at lower relative humidities (<0.7)
EXPERIMENTAL CORN DRYING DATA
follow a 0.78 purity line. However, since
.05 x EXPERIMENTAL CLAY WETTING DATA
only one clay was being used, one continu
- THEORETICAL PREDICTIONS
ous curve as noted on the graph was deter
i 1 i 1 i_ 1 , 1 . 1 . 1 . 1 . mined for use in computing the drying
0
10 20 30 40 SO 60 70 80 curves for all grains. The equation for that
curve was
TINE (HOURS)
Purity = 0.78 - 0.11 RH + 0.28 RH2 (21)
Figure 4. Comparison of experimental and predicted results of drying corn using sodium ben
tonite. Data from Sturton and Bilanski (1980). Run SCORNI. It is evident that this equation is valid only
for the sodium bentonite that was used in
SIMULATION AND PREDICTION reported in Sturton et al. (1981) and the work of Sturton and Graham. Further
Drying trials with freshly harvested Graham et al. (1983) in which the authors work at the University of Guelph using
wheat, corn, oats and barley were carried fitted their experimental data with expo nominally the same clay of a different
out at the University of Guelph. The nential equations but did not actually mesh size found a slightly lower purity.
experimental method and results are model the drying process. The experimen Thus, it is evident that to use Black Hills
Prediction Results
Twenty-one experimental drying
u .1 curves were predicted with the initial
UJ
moisture contents of grain and clay and
OC their mass ratios as the only inputs. The
3
r- data represented a variation of initial
U)
S . OS
moisture contents of grain and clay and
EXPERIMENTAL BARLEY DRYING DATA mass ratios of grain to clay. As space did
x EXPERIMENTAL CLAY WETTING DATA not premit inclusion of all, one example of
- THEORETICAL PREDICTIONS the prediction curves of wheat, corn, oats
I I and barley are shown in Figs. 47, respec
10 20 30 40 SO 60 70 80 tively. The drying curves for the wheat
were virtually identical whether Eqs. 6 and
TINE (HOURS) 8 or Eqs. 10-12 were used. The predic
tions for barley used Eqs. 6 and 9. The
Figure 5. Comparison of experimental and predicted results of drying wheat using sodium corn drying curves were obtained using
bentonite. Data from Sturton and Bilanski (1980). SWHEAT1.
Eqs. 5 and 7. The equations used by Flood
et al. (1972) were empirically derived for
.2 air drying, hence, apparently do not apply
DRYING WHEAT WITH SODIUM BENTONITE to adsorption drying, as they predicted a
curve which did not reach equilibrium
moisture content even after 180 h. The
IS prediction of those runs which had a low
clay content relative to the grain became
UJ unstable, probably due to the inaccuracy of
the clay moisture isotherms at high
.1 moisture contents.
One interesting and fairly consistent
observation that can be seen in the experi
mental data (e.g. oats data) and that could
r: .05 not be predicted, is the interim plateau that
EXPERIMENTAL WHEAT DRYING OATA
is reached in the clay adsorption curve
x EXPERIMENTAL CLAY VETTING DATA
between 4 and 30 h. The wetting process of
- THEORETICAL PREDICTIONS
bentonites is quite complex. The initial
1,1,1,1,1,
i. moisture adsorption for single-layer wet
ting of the clay particles follows the BET
10 20 30 40 50 60 70 80 equation. As more water molecules are
TIME (HOURS) added, there is some reorientation of the
clay particles, a process which takes time
Figure 6. Comparison of experimental and predicted results of drying barley using sodium (Grim 1968). This reorganizing ofthe clay
bentonite. Data from Sturton and Bilanski (1980). Run SBARL2.
particles is thought to be responsible for
the hysteresis in the adsorption-desorption
Wyoming bentonite clay will not neces curves. It is postulated that this is what is
workers. This may indicate that there is
sarily yield consistent moisture adsorption more illite (curve 9) or kaolinite (curve 10) happening during this plateau. For the pre
isotherms for the clay. dictions of the grains it is assumed that the
present in the clay used by Sturton and
Instead of using only the moisture iso moisture adsorption is instantaneous;
Bilanski (1980) and Graham et al. (1983)
therms found experimentally, we thought hence, the plateau is not predicted.
than in those of other workers.
it wise to relate all moisture isotherms to a Also noted on Fig. 2 are two points
standard. Since Keren and Shainberg's which were experimentally derived CONCLUSIONS
data satisfies the BET theory at low relative moisture contents at known relative (1) Using the conventional single-ker
humidities, it was felt that these data would humidities for the clay used in the experi nel drying equations and the given
make a reasonable standard. ments. Both differ somewhat from the equilibrium moisture isotherms noted in
The experimentally derived moisture curve derived above. This discrepancy this paper, drying curves for wheat, corn,
adsorption isotherm is plotted as curve 8 may be due to the fact that clay mixed with barley and oats were predicted for one type
on Fig. 2. It is noted that this isotherm is grain cannot expand readily. Even though of sodium bentonite.
somewhat below most curves for sodium expansion is minimal at low moisture con (2) The interim plateau reached in the
montmorillonite determined by other tents, some reorganization ofthe clay par clay adsorption curve between 4 and 30 h
Brown, R. B., L. Otten, and J. E. Brubaker. 1987. Automatic timer control for a batch-in-bin dryer. Can. Agric
Eng. 29: 179-182.
Acomputer-based control system for abatch-in-bin grain dryer was developed and tested with afarm-scale grain dryer
in 1985. Nine drying tests were conducted with wet corn harvested at about 30% initial moisture content (wet basis). The
control system utilized an inexpensive personal computer (Radio Shack TRS-80 Model 100) and an A/D interface unit
designed tocommunicate with the computer through the RS232 port. The control algorithm was based upon simulation of
the drying process, with known initial conditions and parameters. Control to within 0.3 percentage points ofthe target final
moisture content of the corn dried was realized in three of the tests.
INTRODUCTION resulting from uneven stirring or pockets units. Since batch-in-bin drying uses air
Batch-in-bin dryers are commonly used of fines make it necessaryto obtain a large temperatures above 60C, equilibrium is
to dry corn on farmswherethe annualcrop number of samples from the grain bulk. attained for only a small fraction of the
does not exceed about 500 t. This system These samples must be blended and cooled bed, and the method is inappropriate.
requires a relatively low initial investment, before a moisture tester can be used to A set of four partial differential equa
is flexible and with careful management measure average moisture content. The tions can be derived to define the changes
can produce high-quality grain. In addi procedure must be repeated until the batch in grain moisture and temperature and air
tion, energyefficiency is potentially excel is finished. In practice, dryer operators do specific humidity and temperature for a
lent especially with exhaust heat recovery not test the grain but rely upon experience bed of grain (Meiering et al. 1977; Bakker-
(Otten 1985). to time the batch. This approach is satisfac Arkema et al. 1974). These may be solved
A typical drying bin is about 7000 mm tory when variations in grain moisture, numerically, but very small increments of
in diameter and has a fully perforated dry depth and ambient conditions are slight. time and depth are necessary for solution
ing floor. A fan and burner unit delivers More typically, the operator acts in such a stability. For a deep bed and a high initial
heated drying air, usually at 60-100C, to way that the corn is always overdried by moisture content simulation time can
an underfloor plenum. This drying air 2-3 percentage points to be on the safe become excessive.
picks up moisture as it passes through a side. A similar approach is to perform heat
1000- to 2000-mm depth of wet corn in the Apart from a reduction in grain quality, and mass balances around a thin section of
bin, and is exhaustedthrough roof vents. A deliberate overdrying costs at least $5.00 the bed, and then treat the bed as a stack of
level fill of wet grain is dried to an accept per tonne of grain marketed at present adjacent layers (Thompson et al. 1968).
able moisture content and then cooled in prices. The cost is due to additional fuel The thin-layer drying rate is calculated
the bin and unloaded to storage. The time costs for drying and to weight loss from from an appropriate empirical equation.
required for one cycle varies from about 12 unnecessary moisture removal. The resulting system of equations can be
to 20 h depending upon initial moisture A control system to regulate final aver solved quickly if several simplifying
content, grain depth and drying tem age moisture content for a batch-in-bin assumptions are made. For example, ker
perature. dryer is desirable. A closed-loopfeedback nel temperature is assumed to be in equi
The deep bed of grain in the dryer cre control system for moisture content is not librium with the drying air. It was also
ates two problems with the system. First, a possible because average moisture content assumed that although condensation
steep gradient of moisture content exists is not readily measured. An alternative occurs in the upper grain layers, rewetting
from the bottom to the top of the grain bed approach is to use a predictor, or a calcu does not occur due to the relatively short
at the end of drying. Although grain is lated estimate of the moisture content, to drying time. This assumption greatly
mixed as the dryer is unloading, the dif control the dryer. The objective of this reduced the computation time for the first
ference in moisture content among indi study was to develop and field-test such a half of the drying period whengraincondi
vidual kernels may be as great as 20 control system. tionsabove the drying front areessentially
percentage points. Various schemes to stir constant.
or mix the grain as it dries have been SIMULATION MODELS The model used wasdevelopedfrom the
introduced to eliminate the gradient, and Three approaches to the mathematical concepts presented by Thompson (1970).
dryer manufacturers supply either auger modelling of deep-bed corn drying were A time increment of 300 s (5 min) and a
stirring machines or grain recirculating investigated. The logarithmic model (Sab- layer thickness of 50 mm (2 inches) was
devices. bah et al. 1979; Barre et al. 1971) is the used for simulation. Psychrometric rela
The second operational problem is one simplest and fastest to use. It is suited to tionships given by Wilhelm (1976) and
of control. It is very difficult to monitor the near-equilibrium drying since it neglects published by the American Society of
moisture content of grain in the dryer in sensible heating of grain by the drying air. Agricultural Engineers (1983, D271.2)
order to determine when to shut off the Moisture and temperature gradients in the were used to calculatemoist air properties.
burner. The axial moisture gradients men grain bed are assumed to be logarithmic The Zeroin search technique reported by
tioned earlier and also radial fluctuations functions of timeanddimensionless depth Bakker-Arkema et al. (1974) was used
180
CANADIAN AGRICULTURAL ENGINEERING, VOL. 29, NO. 2, SUMMER 1987
TABLE 1. SUMMARY OF CONDITIONS FOR 1985 CONTROL SYSTEM TESTS 2.45 MJ/kg of water removed for the dryer
with the heat exchanger. Without the heat
Ambient conditions
Initial Wet Bed Drying Grain Duration Final exchanger, specific energy consumption
moisture weight depth Temp. RH temperature temperature of cooling moisture
(% WB)
was 2.98 MJ/kg water removed, an
Test (tonnes) (m) (C) (%) (C) (C) (h) (% WB)
increase of almost 22%.
1 29.9 28.1 1.0 6 80 90 10 2.0 11.9
2 31.4 29.0 0.9 16 85 no 22 2.5 12.2 DISCUSSION
3 27.3 37.5 1.3 11 85 108 11 2.5 12.6 Once the amount of drying that occurred
4 28.0 37.2 1.3 16 50 103 12 3.5 15.2
5 30.4 35.8 1.2 10
with cooling was known, the control sys
45 109 4 3.5 15.8
6 31.5 33.2 1.2 8 55 88 5 3.5 11.9
tem worked very well. A target moisture
7 28.9 32.8 1.6 6 60 107 5 3.5 16.4 set 3 percentage points above the desired
8 27.5 43.5 1.5 8 80 108 6 3.5 15.3 final value (i.e., 18.5% for 15.5% final
9 27.0 40.0 1.3 5 85 92 4 3.0 16.5 average) gave good results. For other dry
ers this parameter may vary, but it would
be determined as part of the tuning pro
endar, battery back-up power pack in case first three tests was severely overdried cedure. Final moisture content of cool
of power fluctuations, small size, built in (Table I). It was impossible to determine
grain was controlled to within 0.3 per
display for data input prompts, and low the average moisture content at the time
centage points for three of the test runs. If
cost. the burner was shut off because there were
the initial moisture content of tests 7 and 9
A controller and interface unit was built radially distributed zones of wet and dry
had been more accurately determined, it is
to set the gas solenoid valve and dryer fan corn created by the stirring auger paths.
probable that the final moisture contents
relay. Two solid state relays were ad Therefore the third test was unstirred, and
would have been within the same range. It
dressed by setting the two bottom bits of a the average moisture content was 17.3%
is doubtful that greater accuracy could be
control byte and writing the result to the wet basis compared to a simulated value of
maintained consistently because commer
controller via the RS-232-C port. 17.0%, a difference of less than 2%. After
cial moisture testers have an error range of
cooling the average moisture content was
the same magnitude.
FIELD TESTING 12.6% indicating that more than 4 percent
The dryer used for the field tests was
Level filling and uniform mixing by the
age points of moisture were removed with
stirring device was assumed in the model
located near Ayr, Ontario. It was 7300 mm cooling.
in diameter and 6100 mm to the eave with a
and should be realized in the actual sys
The control program was altered assum
full perforated drying floor. A 660-mm tem. An implicit assumption in the control
ing a conservative value of 3 percentage scheme is that the ambient absolute
diameter, 7.5-kW axial fan and a natural points moisture removal with cooling. The
gas burner provided heated drying air. humidity is constant throughout drying.
subsequent two tests yielded a final
Grain was distributed by a centrifugal Although ambient temperature and relative
moisture content within 0.3 percentage
spreader as it was loaded into the bin, and a humidity can vary significantly in any 10-
points of the target level of 15.5% wet
six-auger stirring machine traversed the to 20-h period, it is unlikely that the
basis. Test 6 indicated that the moisture
bin once every 4 h during drying. humidity ratio will deviate greatly over the
removal during cooling was the same with
An exhaust air heat recovery system was same period. This drying method is quite
or without the heat exchanger. The seventh
installed on the dryer utilizing a crossflow insensitive to ambient conditions in any
and ninth tests both produced corn wetter
flat plate heat exchanger which was tested event due to the large temperature increase
than desired. In both cases there was a
with the burner.
the previous year (Otten 1985). discrepancy of about 1 percentage point
The first dryer test commenced 16Oct., The drying temperatures used in the
between the moisture tester readings at the
tests were higher than normal for deep-bed
and nine tests were conducted between that farm and the oven-dry moisture deter
date and 9 Nov. A poor field drydown rate batch drying because the dryer was oper
mination in the laboratory for initial grain
coupled with persistent rainy weather ated for maximum throughput. For a tem
conditions. The tester results were used to
made for sporadic testing. The conditions perature range of 60-70C moisture
control the dryer, and these had been in
for the tests are summarized in Table I.
gradients within the bed would be reduced
good agreement with the oven-dry method
The grain stirrer drive reduction ratio and a more uniform product moisture
in the earlier tests. Test 8 produced good
could be maintained.
was changed prior to test 4 since the oper results, and the moisture tester determina
ator felt that better mixing and more uni tion of initial moisture content was again CONCLUSIONS
form drying would result from a 2.5-h close to the oven-dry method. The main conclusions drawn from the
stirring cycle. The stirrer was not started The grain stirring augers did not do an study are as follows:
until 6 h of drying time had elapsed, in adequatejob of mixingthe grain, although 1. The simulation model allowed an
order to establish a uniform drying zone there was some reduction of moisture gra accurate prediction of the time required to
before stirring activity disrupted the air dient in the dried grain. The delayed stir dry corn to a specific average moisture
flow pattern in the bed. Three tests (3, 6 ring scheme (Test 4 and subsequent tests) content.
and 9) were conducted without stirring. improved the situation, but radial zones of 2. With accurate determination of grain
The heat exchanger was disconnected for wet grain were encountered between adja and system operating conditions, the cal
two tests to evaluate its effect on dryer cent auger paths. The solution to this prob culated drying time can be used to control
performance. lem would be more augers, or variable- a batch-in-bin dryer and regulate the final
path stirring augers. The overall energy grain moisture.
RESULTS efficiencyof drying and the accuracyofthe 3. Moisture removal during the cooling
Moisture removal during cooling was simulation model did not appear to be stage averaged 3 percentage points for the
greater than the 1.5 percentage points affected by stirring strategy. particular combinations of temperature
anticipated, consequently the corn of the Specific energy consumption averaged and airflow rates in this study.
Jayas, D. S., S. Sokhansanj, E. B. Moysey,and E. M. Barber. 1987. Distribution of foreign material in canola bins
filled using a spreaderor spout. Can. Agric. Eng. 29: 183-188.
The distribution offoreign material (chaff and fines) was studied in a4.6-m-diameter bin filled with Tobin canola. The
bin was filled twice using a central spout and once using a conical spreader. The distribution ofchaff and fines can be
described using a quadratic equation with radial coordinates. The use ofa conical spreader to fill the bin with canola
increased the uniformity of chaff and fines distribution only slightly. Contrary to general belief, fines were not highly
concentrated in the center of the bin for either ofthe fill methods. The variation ofchaff and fines with height was
insignificant.
INTRODUCTION 3.2 million tonnes of canola worth 1.2 tors such as moisture content ofthe plants,
The drying of farm crops is an integral billion dollars were produced in Canada combine setting, presence of weeds and
part of agricultural production systems. (Anonymous 1984). Depending on the atmospheric conditions at the time of com
The current trend is to dry farm crops, weather conditions during harvest, drying bining. The amount of foreign material in
including canola, using near-ambient tem of canola may be a necessity. High- tem the harvested canola was difficult to con
perature air. Aeration is the most com perature air drying is undesirable for can trol. Therefore, instead of keeping con
monly used method for cooling stored ola because of the fear of contamination stant foreign material at harvest, it was
grains and oilseeds. In both the near- from burnt kernels, and nonuniform air decided to sample each truck load for for
ambient drying and the aeration systems, flow distribution in near-ambient drying eign material. Trucks were sampledat the
air in varying quantities, 10-30 L/(s m3) systems can result in localized regions of time of unloading. The dropping streamof
for near-ambient drying and 1-2 L/(s m3) wet seeds causing mouldy or heated ker canola from the truck was collected every
for aeration systems (Friesen et al. 1984), nels. Both the burnt and heated kernels 120 sec. The canola stream was cut by
is forced through a perforated area and is complicate refining ofthe canola oil. As a moving a container in one direction per
then distributed throughout the grain first step in an attempt to predict the uni pendicularto the falling stream.The speed
mass. The success ofthe systems is highly formity of airflow within stored canola, of the container was controlled manually
dependent on uniformity of the airflow this study was undertaken to investigate and was assumed to be nearly constant.
within the stored mass. The uniformity of the distribution patterns of foreign mate The time to empty the truck ranged from
airflow is affected by the configuration of rial (chaff and fines) in canola bins as 20 to 30 min depending on the setting of
the perforated area used for introduction of affected by the method used to fill the bin. the unloading gate. The samples collected
the air; bin geometry; resistance to airflow Based on average particle diameter of for an individualtruck were mixed to give
of the product; and the amount, size and Tobin canola, chaff was defined as over a representative sample for that truck.
distribution of foreign material in the flow of No. 12 mesh US Standard sieve The bin was filled on two occasions
stored bulk. If a clean product at known and fines as underflow of No. 16 mesh US using a central spout with a drop heightof
moisture content is used to fill the storage Standard sieve. 7.0 m (measured from the perforated floor
bins, then theeffectofthe configuration of to the outlet of the unloading screw con
the perforated area and the bin geometry MATERIALS AND METHODS
veyor). Additional canolawasaddedusing
can be predicted reasonably well using the Sampling for Foreign Material a conical spreader to the partially filled
mathematical models of several inves Distribution bin. The dropping height for the conical
tigators (Brooker 1969; Marchant 1976; The amount, size and distribution of spreader filled bin was 3.0 m (measured
Segerlind 1982; Singh(Jayas)and Sokhan foreign material was studied in a farm bin, from apex of the inverted bottom cone to
sanj 1985; Miketinac et al. 1986). 4.6 m in diameter and 6.89 m in overall the leveled canola surface). Since the bin
The effect of the amount, size and dis height. The bin was erected on a farm near was being monitored for another experi
tribution offoreign material onuniformity Saskatoon. The perforated floor was 0.48 ment, the canola was added on top of the
of airflow can also be incorporated in the m above the concrete floor and the bin roof canola already present in the bin when
mathematical models if the distribution of had a 26.7 slope. The effective vertical using the conical spreader. The arrange
foreign materials in storage bins can be height for storage of canola was 5.26 m. ment of various components ofthe unload
estimated. Chang etal. (1981, 1983, 1984) Canola, variety Tobin {Brassica camp- ing system is shown in Fig. 1.
studied the distribution of foreign material estris), harvested during the fall of 1984 Stephens and Foster (1976, 1978) and
incorn and wheat bins. Toourknowledge, was used to fill the bin. Canola was first Chang et al. (1981,1983,1984) studied the
studies on thedistribution of foreign mate swathed and then combined at moisture distribution of fines in corn, wheat and
rials in other stored crops have not been contents of about 6-10%. The low sorghum. They used a multicompartment
reported in the literature. moisture contents were due to hot dry probe to sample the bin at grid points
Canola is an economically important weather at harvest. The amount of foreign spaced 300 mm radially and 280 mm ver
oilseed crop for Canada. In 1984, about material in canola depends on many fac tically. This method of sampling was not
used in the present study for several rea sampling and was reintroduced after
sons. First, it was decided that the opening unloading the next truck. Four truck loads
through which the sample enters the probe were used for the first bin fill and five truck
is small and could prevent entering of the loads were used for the second bin fill.
chaff due to bridging. Second, if a probe The profile after filling the bin usingthe
with a large opening and diameter is used conical spreaderwas not a conebuthadthe
it disturbs the sampling area during inser shape of a doughnut. The sampling grid
tion; also, the difficulty in inserting the for the spreader-filled bin is shown in Fig.
probe increases significantly with probe 3. In preliminary tests, filling with the
diameter. From studies of Stephens and conical spreader did not give an axisym-
Foster (1976,1978) and Chang etal. (1981, metric fill, rather canola sloped in a heap
1983, 1984) it can be assumed that the from one side ofthe bin wall. The problem
distribution of foreign material is sym was in the design of the spreader. The
metrical about the vertical axis of the bin opening between the two cones was too
when bins are filled using spreaders or a wide. Therefore, a slight angle from the
centralspout. Therefore, sampling of a bin vertical ofthe dropping canola guided it to
alongone radiuswas considered sufficient one side. The spreader was modified so
instead of along the bin diameter. The that the top cone was nearly full during
method of sampling a bin along its radius emptying. Also, thehanging portion ofthe
is described below. This method could be dropping-spout was tied so that it dis
extendedto samplealong the bin diameter. charged vertically downward. The modi
The first truck load (about 81) of canola fications resulted in axisymmetric dis
was transferred into the bin using the tribution of foreign material. Many
unloading system shown in Fig. 1. A farmers may not notice this problem and
plywood platform (250mmwide and2500 the conical spreader would result in bins
mm long) was hung from an inside ladder having all the chaff accumulated on one
in the bin. The platform was parallel and side.
close to the canola profile in the bin but
was not touching the canola surface. This Analysis of Samples for Foreign
arrangement allowed easy sampling of Material
canola at all sampling locations spaced Collected samples were subdivided
372 mm apart along the bin radius. The using a Boerner divider into samples of
Radius (r),m 250 g each. The 250-g sample was frac
samples were collected using a circular
sampler 100 mm in diameter and 145 mm Figure 2. Sampling points for spout-filled tionated into three components using two
long.The sampler wasinserted atthepoint bins. Tylerwoven wiremesh sieves (No. 12 and
of samplingand then wascoveredusingan 16). The opening of No. 12 sieve was 1.70
aluminum sheet at the bottom and lifted. 600 g. The samples near the bin wall mm and of No. 16 was 1.18 mm. The
The disturbance caused by sampling at one weighed less because of the presence of sieves were 51mm in depth and 203 mm in
point was not transferred to the adjacent higheramounts of lighter chaff. As shown diameter. The samples were shaken for 5
sampling point. The amount of sample in Fig. 2, the grid of sampling points in min using a Ro-Tap shaker. The shaking
collected at each point was about 800 g this method followed the profile of the time was determined based on preliminary
except nearthebinwall where itwas about canola. The platform was removed after tests in which material on sieves was
2-5 -1 -5
-2 -6
3 -7
20 4
8
9
0 I
Radius (r), m II 1-5
Figure 3. Sampling points for a conical
a>
spreader-filled bin.
10
o
recorded at 2-min intervals. The dif e
ference between amounts of material
retained on the various sieves after 4 and 6 !o-5
min was negligible. Hence, 5 min was
arbitrarily chosen as the shaking time.
The material retained on the No. 12
00
sieve was defined as chaff, the underflow
of No. 12 and overflow of No. 16 was
00 0-5 10 I-5 20 2-5
defined as canola and the material that Radial Distance, (r), m
passed through No. 16and was retained on Figure 5. Distribution offines along the radius ofabin which was filled axisymmetrically using a
the pan was defined as fines. Canola at this central spout. The radial distances were measured from the bin center. The numbers 1to
point contained a small amount of needle 4 are for first bin fill and 5 to 9 are for second bin fill.
likeparticles of plantresidue. Theseparti
cles wereseparated as overflow using rid of the bin were used to estimate the fines in the mass of canola for spout-filled
dle screen No. 000 with a feed rate set at 3 weighted average of chaff and fines for bins are shown in Figs. 4 and 5, respec
on a Carter Dockage Tester (Simon Day each truck load. The volumes of canola in tively. The lines were drawn through the
Ltd., Winnipeg, Man.). This small frac the bin represented by each sample were mean values of chaff and fines. The means
tion was added to the chaff because it taken into account when means of chaff were calculated from triplicate samples
looked similar. The masses of fines, can and fines were estimated. collected 372 mm apart (radially), except
ola and chaff were recorded and were used The individual amounts of chaff and for the sampling location, farthest from the
to give the distribution of foreign material fines were divided by the mean values of center, where only duplicates were used.
in thebin as affectedby the method used to chaff and fines to determine the nor The quantity of chaff and fines were nor
fill the bin. malized chaff and fines. malized by dividing the measured value at
each point by the mass averagequantity at
Normalization of Foreign Materials RESULTS AND DISCUSSION
the same level. The plottedlinesrepresent
The amounts of fines and chaff deter Distribution in Spout-filled Bins two bin fills. The lines numbered 1 to 4 are
mined at seven locations along the radius The radial distributions of chaff and for first bin fill and 5 to 9 are for second bin
ormalized P-bai
| 1.5
2.0 -
-
a-Spout fill tion:
o
filled bin; Cs = chaff at radius r for spout-
filled bin; and r = location along bin
v radius measured from central axis (m).
k ^
Chang et al. (1981, 1983) have reported
very high concentrations of fines around
the center of the bins filled with corn and
wheat. We did not observe very high con
centration of fines near the center of the
o.qL X J_ bin. Rather, we observed that the con
centration of fines in canola bins near the
0.0 0.5 1.0 1.5 2.0 2.5
bin center was almost equal to that near the
Radial Distance,(r) , m sides ofthe bin. The minimum fines were
Figure 6. Distribution ofaverage chaff and best fit curves for spout- and spreader-filled bins. The found about midway between the center
radial distances were measured from the bin center. and the bin wall. The canola particles are
very small, and hence distinction between
canola and fines is not as markedly defined
2.5r- as in the studies of Chang et al. (1981,
a-Spreader fill 1983) with corn and wheat. The distribu
a-Spout fill tion of fines in spout-filled canola bins can
2.0 mathematically be given by:
FJFS = 1.093 - 0.474 r + 0.241 r2 (2)
(correlation coefficient = 0.88)
1.5
where Fs = mean fines for spout-filled
0) bins; and Fs = fines at radius r for spout-
=5 ,0 w^r-^-T--^irr: filled bins.
E
o Distribution in Spreader-filled Bins
z 0.5 Normalized distribution of chaff and
fines for spreader fill are given in Figs. 8
and 9, respectively. The lines were drawn
o.oL through the mean values of chaff and
0.0 0.5 1.0 1.5 2.0 2.5 fines. The means were calculated from
triplicate samples collected 372 mm apart
Radial Distance ,(r), m
(radially), except for the sampling loca
Figure 7. Distribution ofaverage fines and best fit curves for spout- and spreader-filled bins. The tion, farthest from the center, where only
radial distances were measured from the bin center. duplicateswere used. Linesnumbered 1,2
and 3 represent the samples taken after
fill. The variation among replicates for fill can be given. Most of the chaff was transferring first, second and third truck
each location was less than 5% for 80% of concentrated near the wall ofthe bin (Fig. loads of canola. The detailed data are
the samplesof chaff, and for all samplesof 6), and minimum chaff occurred between given elsewhere (Jayas 1987). The varia
fines. The detailed experimental data are 0.6 and 0.8 m from the center of the bin. tion among subsamples was less than 5%
givenin Jayas(1987). Asadditional canola This trend was also confirmed by visually for most samples of chaff and for all sam
was added in each filling, the drop height observing the canola profileafter the addi ples of fines. In the case of spreader fill, a
decreased for successive additions. The tion of each load. The cleanest spot was change in drop height had no significant
effect of drop height on radial distribution not the center of the bin but rather it was effect on radial distribution of chaff and
of either chaff or fines is negligible. the annular area between 1.2 and 1.6 m fines. The mean radial distributions of
Hence, the chaff and fines were averaged diameter of the bin. It was expected that chaff and fines for spreader fill are shown
for all samples at each sampling point minimum chaff should occur in the center in Figs. 6 and 7, respectively. The radial
along the bin radius. of the bin. During filling using the central distribution of chaff in spreader-filled bins
The mean values for chaff and fines are spout, particles larger than canola have a is described by:
shown in Figs. 6 and 7, respectively. In tendency to slide along the slope, but some CJCC = 0.960 - 0.866 r + 0.520 r2 (3)
calculating mean values, data points of these particlesdid not have enough time (correlation coefficient = 0.94)
marked "5" were not used. No explana and were trapped in the falling stream.
tion for such a distinct behavior in trans This trapping was probably the main rea where Cc = mean chaff content for
ferring the first loadduring the second bin son for higherchaff at the centercompared spreader-filled bins; and Cc = chaffcon-
a
if tionated to find chaff and fines means for
that bin. By substituting values for radial
-
^ 10 " ><
/'' an appropriate equation, the distribution of
-'/'' foreign material in the bins can be pre
dicted. In the event a representative sam
---'/'
"^* "~ "* ** ** -. ^v ^Z ../ ple cannot be obtained and analyzed, an
|o-5h
estimate of total foreign material in canola
can be used. The estimated foreign mate
rial can be divided by 2 to obtain chaff and
I 1 1 i 1 . 1 . 1 fines means.
00
Equations 1, 2, 3 and 4 are developed
00 0-5 10 1-5 20 2-5
from experimental data for 4.6-m-diame-
Radial Distance, (r), m ter bins. The use ofthe equations to predict
distribution of foreign material should be
Figure8. Distribution of chaffalongthe radiusof a bin whichwas filled axisymmetrically using a
limited to the bins having diameters of
grain spreader. The radial distances were measured from the bin center.
about 4.6 m. The main reason to limit this
study to a 4.6-m-diameter bin was its
2-5 availability and its commonness on most
farms in western Canada. An extensive
1
study with different diameter bins would
2 require substantial cooperation from the
20- 3 farmers and a significant amount of finan
cial resources. If funds become available a
general equation should be developed for
1-5 prediction of chaff and fines not only in
a>
canola bins but also in bins filled with
other cereals and oilseeds.
| 10
CONCLUSIONS
o
The distribution of chaff and fines in a
z
0-5 4.6-m-diameter bin can be modelled using
quadratic equations relating the fraction of
chaff or fines to the radial coordinate.
Contrary to general belief, fines were not
00 1
highly concentrated in the center of the bin
00 0-5 10 1-5 20 2-5 for either of the fill methods. The use of a
Jayas, D. S., S. Sokhansanj, E. B. Moysey, and E. M. Barber. 1987. The effect of airflow direction on the
resistance of canola (rapeseed) to airflow. Can. Agric. Eng. 29: 189-192.
The resistances to airflow of Tobin and Westar canola, of foreign material usually found in harvested canola, and of
several mixtures of Tobin and foreign material were measured experimentally for two airflow directions. The resistance to
airflow of individual samples for the horizontal airflow direction was 0.5-0.7 the resistance to airflow for the vertical
direction. From a practical point of view, the resistance for horizontal airflow direction can be estimated as 0.6 the
resistance for vertical airflow direction.
INTRODUCTION of mixtures of canola and foreign mate of 720 mm. A square (230 mm x 230
Resistance to airflow of various agri rials. mm) airflow straightener made from steel
cultural products has been studied for at wool was installed in the pyramidical sec
least 55 yr by many investigators including MATERIALS AND METHODS tion to diffuse the air uniformly. The air
Stirniman et al. (1931), Shedd (1951), Bar- Experimental Method was then forced through the canola col
rowman and Boyce (1966), Bakker- The effect of the direction of airflow on umn. Two possible directions of airflow
Arkema et al. (1969), Agrawal and Chand resistance to airflow was studied in a cubic (horizontal and vertical) were used. The
(1974), Bern and Charity (1975) and box (Fig. 1). Compressed air from a cen box was always filled from the top as
Abrams and Fish (1978). (An extensive list tral compressor was regulated by a ball would be the case in bins filled on the
of these studies is given in Jayas (1987).) valve and a pressure regulator. The air was farm. The direction of airflow was
In most of these studies on resistance to filtered before entering the pressure reg changed by rotating the air delivery sys
airflow of agricultural products it has been ulator. The air was then forced through a tem. The pressure drop, to an accuracy of
assumed that resistance to airflow is equal calibrated orifice plate. The measured vol 0.25 Pa (0.025 mm of H20), was mea
in all directions. Therefore, the reported ume of air was introduced into a pyra- sured across 250 mm of a canola column in
results were measured only for the vertical midical air plenum having an apex height both directions. The study was conducted
airflow direction. Many mathematical
models (Brooker 1969; Marchant 1976;
Haque et al. 1981) have been developed to
predict pressure patterns in grain bins.
.--<r
These models are also based on isotropic
behavior of agricultural products. Lamond
and Smith (1982) pointed out the pos 250
AIR
sibility of different resistances to airflow
based on airflow direction since most ker
nels (such as wheat and barley) lie with
their major axes horizontal. Recent inves
tigations by Kumar and Muir (1986) con
firmed the variability of resistance to
airflow with airflow direction for wheat
and barley. Canola seeds are spherical in
shape, hence the effect of direction of air
flow on resistance to airflow of clean can
AIR STRAIGHTENER 720
ola seeds might be expected to be
negligible. The resistance to airflow of
foreign material usually found in har
vested canola is expected to be different
for different airflow directions. To our
knowledge, no study has been reported on
the effect of direction of airflow on resis
ALL DIMENSIONS ARE IN MILLIMETERS
tance to airflow of canola, of foreign mate
rial usually found in canola and of
mixtures of canola and foreign materials.
The objectives of this study were to deter
mine the effect of direction of airflow on AIR
resistance to airflow of canola, of foreign Figure 1. Elevation view of the apparatus used to study the effect of direction of airflow on
material usually harvested with canola and resistance of various samples. The apparatus was supported on a light metal frame.
The material then was dropped approx Westar 0 0 Sprinkle Horizontal 725.0 34.4 344.6 0.8151 0.998
imately another 800 mm in a rain-like Westar 0 0 Sprinkle Vertical 725.0 34.4 221.8 0.8608 0.997
manner resulting in a packed fill. The 0 100 0 Spout Horizontal 406.6 68.8 2071.3 0.8643 0.995
0 100 0 Spout Vertical 406.6 68.8 2089.8 0.7788 0.997
sprinkle-fill method was used to simulate
0 100 0 Sprinkle Horizontal 481.9 63.0 1037.9 0.8809 0.996
grain spreaders used in farm bins. The 0 100 0 Sprinkle Vertical 481.9 63.0 986.0 0.7811 0.998
design of a scale-model grain spreader was 15 0 Spout Horizontal 638.0 44.9 505.2 0.8431 0.994
Tobin
not considered a viable option. Tobin 15 0 Spout Vertical 638.0 44.9 271.8 0.8821 0.997
Canola Samples Tobin 25 0 Spout Vertical 604.5 48.5 384.2 0.8521 0.997
190
CANADIAN AGRICULTURAL ENGINEERING, VOL. 29, NO. 2, SUMMER 1987
TABLE II. MEAN PRESSURE DROPS AND STANDARD DEVIATIONS AT VARIOUS AIRFLOWS extent, these factors. Considering all these
ACROSS 1 m OF 6.5% MOISTURE CONTENT TOBIN CANOLA FOR TWO AIRFLOW DIREC factors, a variation of 5-10% in experi
TIONS AND TWO METHODS OF FILL
mental data is justifiable, and data with
Pressure drop (Pa/m) such variability have an excellent practical
Horizontal Vertical value.
Spout Sprinkle Spout Sprink e
Sprinkle fill gave consistently higher
Airflow
(m3/(sec-m2)) Mean SDt Mean SD Mean SD Mean SD
bulk densities by approximately 7%, as
can be seen from Table I. Resistance to
0.0158 80.0 10.6 166.8 10.3 120.0 10.1 200.0 0.0
0.0217 105.2 6.1 210.0 11.0 158.8 7.9 280.0 0.0 airflow was also significantly affected by
0.0298 158.8 8.3 303.2 19.7 225.2 6.0 400.0 0.0 fill method. In Fig. 2, difference in airflow
0.0454 258.8 9.2 480.0 17.9 352.0 11.3 613.2 20.6 resistance due to fill method appears to be
0.0624 382.8 10.1 713.2 30.1 514.8 4.1 893.2 20.6
small, because it is drawn to a log scale; in
0.0857 573.2 16.6 1063.2 38.8 757.2 6.5 1300.0 17.9
733.2 24.4 1353.2 30.1 954.8 4.1 1653.2 20.6
fact it is a two to one ratio.
0.1032
0.1178 880.0 36.7 1613.2 30.1 1133.2 8.3 1926.8 16.3
0.1305 1000.0 28.8 1853.2 20.6 1288.0 7.2 2206.8 16.3
0.1419 1110.8 53.6 2070.0 32.9 1437.2 10.9 2456.8 15.0
Analysis of Data
0.1523 1240.0 55.2 2273.2 37.2 1576.0 13.4 2700.0 17.9 Equation 1 in its linear form was fitted to
0.1619 1336.0 41.6 2480.0 31.0 1716.0 9.8 2910.0 11.0 mean data for each of the 22 test condi
0.1709 1442.8 46.2 2666.8 41.3 1840.0 7.2 3140.0 17.9 tions shown in Table I. The estimated
tSD is standard deviation and is based on sample size of six. coefficients A and B along with correlation
coefficients for each sample are given in
00 Table I. The correlation coefficients for all
_
D O a O
tion. Statistically, the mean values of the B
-
transfer properties due to direction of air BARROWMAN, R. and D. S. BOYCE. 1966. MOHSENIN, N. N. 1970. Physical properties
Air distribution from lateral ducts in barley. of plant and animal materials. Vol. I. struc
flow. If properties either encourage faster
J. Agric. Eng. Res. 11(4): 243-247. ture, physical characteristics and mechan
removal of moisture or maintain the same
BERN, C. J. and L. F. CHARITY. 1975. Air ical properties. Gordon and Breach Science
level of moisture removal for horizontal flow resistance characteristics of corn as Publishers, New York.
airflow direction compared to vertical air influenced by bulk density. ASAE Paper no. STATISTICAL ANALYSIS SYSTEM INC.
flow direction, then development of new 75-3510. Am. Soc. Agric. Eng., St. Joseph, 1982. SAS users guide: Statistics. SAS,
aeration systems will be worth pursuing. Mich. Inc., Raleigh, N.C.
BROOKER, D. B. 1969. Computing air pres SEGERLIND, L. J. 1983. Presenting velocity-
CONCLUSIONS sures and velocity distribution when air pressure gradient data for use in mathe
Based on this study it can be concluded flows through a porous medium and non matical models. Trans. ASAE (Am. Soc.
that resistance to airflow of canola and of
linear velocity pressure relationships exist. Agric. Eng.) 26(4): 1245-1248.
Trans. ASAE (Am. Soc. Agric. Eng.) 12(1): SHEDD, C. K. 1951. Some new data on resis
mixtures of canola and foreign materials
118-120. tance of grains for airflow. Agric. Eng.
for the horizontal airflow direction is 0.6
HAQUE, E., D. S. CHUNG and G. H. FOS 32(9): 493-495, 520.
the resistance to airflow for the vertical TER. 1981. Pressure and velocity field in SHEDD, C. K. 1953. Resistance of grains and
airflow direction. airflow through packed bed of corn mixed seeds to airflow. Agric. Eng. 34(9):
with fines under non-Darcy flow conditions. 616-619.
REFERENCES Trans. ASAE (Am. Soc. Agric. Eng.) 24(6): STIRNIMAN, E. J., G. P. BODNAR, and E.
ABRAMS, C. F. and J. D. FISH. 1978. Resis 1595-1599, 1604. N. BATES. 1931. Tests on resistance to pas
tance of sweet potatoes to airflow. ASAE JAYAS, D. S. 1987. Resistance of bulk canola sage of air through rough rice in a deep bed.
Paper no. 78-4523. Am. Soc. Agric. Eng., oilseed to airflow. Ph.D. Unpublished The Agric. Eng. 12: 145-148.
Contribution no. I-8241 and no. 14012, received 30 May 1986, accepted 10 October 1986
Phillips, P. A. and D. Fraser. 1987. Design, cost and performance of a free-access, two-level pen for growing-
finishing pigs. Can. Agric. Eng. 29: 193-195.
Precast concrete panels and a ramp were installed on a steel support frame in a conventional 2.6 x 4.9-m slotted-floor
pen to produce a free-access two-level pen for growing-finishing pigs. The area ofthe two-level pen was about 1.7 times
that ofthe single level pen at an added material cost of approximately $70 per m2of upper level. Daily gain, feed conversion
efficiency and carcass grade index of pigs reared from 25 kg to market weight in the two-level pen compared favorably with
pigs reared in a one-level pen of similar total floor area.
INTRODUCTION to market weight in two-level and con Canadian standards for pigs of 75 - 100 kg
Two-level pig housing is sometimes ventional pens. per pig (Anonymous 1984). Uniform
adopted as an economical method of using groups were selected from the Animal
barn space. Two-level "decks" are nor METHODS AND MATERIALS Research Centre's specific-pathogen-free
mally used for pigs only at the weaner Pen Description Yorkshire herd. Each group was drawn
stage, and they offer few advantages A two-level pen (Fig. 1)was constructed from approximately six litters maintain
beyond the economy of housing more ani to fit into an existing pen measuring 2.6 x ing, as closely as possible, a balanced sex
mals per unit of floor space. 4.9 m. The 0.84-m high upper platform ratio. The pigs weighed about 25 kg aver
The use of a ramp providing free access consisted of solid panels of precast rein age body weight.
to both levels creates some new pos forced concrete 1.83 m long, with a 2.6-m For the trials, a fortified pelleted corn-
sibilities in pig housing and makes it feas slotted panel (16% porous) along the back barley-soybean meal diet was provided ad
ible to house heavier growing-finishing ofthe pen. The panels were supported by a libitum from two, two-hole feeders on the
pigs in two-level pens. In such pens, the frame of 76 x 76 x 8-mm steel angle. upper level. Water was available from
two levels provide a choice of thermal Other important features of the pen, three nipple drinkers on the lower level and
environment, and the exercise gained in detailed in Fig. 1, include the extended one on the upper level located over the
using the ramp could be beneficial to pig partitions, the protective barrier on the porous floor near the top of the ramp. On
health and condition. The upper level platform, and the ramp. The ramp frame of the day after the pigs were put into the pen,
might be used to increase the proportion of 25-mm square tubing was covered with a an attendant scattered feed on the ramp to
solid floor area available to the pigs, and double layer of expanded metal and cross- attract any pigs that had not yet climbed to
the choice of levels could provide avenues laid with 18xl8x3-mm angle steps the upper level. Out of 96 pigs, four such
of escape from aggressive pen-mates. spaced at 100 mm. The ramp was hinged at animals were identified; after these were
Hence, such designs could provide a the top end so it could be lifted out of the chased up the ramp once, they appeared to
means of addressing concerns over the way. If ceiling height is sufficient, a one adapt to the system.
welfare of intensively housed pigs. piece ramp is used; otherwise the ramp can Body weights of the pigs were recorded
A previous study (Fraser et al. 1986) be made to fold by adding another hinge at at the start of the trial, on day 7, and on
used small prototype pens, 2.4 x 2.4 m, its center with adequate support provided shipping day when a market weight of
to test the principle of free-access, two- under the hinge. 90-105 kg was reached. Total feed con
level housing for pigs. The animals were The total area provided in the two-level sumption was recorded for each group.
studied for relatively short periods (8 wk) pen was about 21 m2, the bottom level Mean carcass grade index for each group
over a weight range of about 20-60 kg at a being totally slotted and 12 m2 (allowing (based on weight and thickness of backfat)
stockingdensity of 0.53-0.63 m2per pig. for space lost because ofthe ramp), and the was computed from packing house
The results showed that pig weight gain, mainly solid top area 9 m2. Manure col receipts. The work extended from June
health and feed conversion efficiency were lected under the slotted floor where it was 1984 to December 1985, each group
at least equal to that of similar pigs housed aerated and circulated by means of an requiring approximately 3 mo to attain
in conventional, single-level pens offering oxidation ditch. Methods of moving pigs market weight.
the same total floor space. into and out of the two-tiered pens are Comparative data were available from
The work reported here tested a modi described by Fraser et al. (1986). four similar groups of 24 pigs each, housed
fied design intended to suit more normal in an adjoining room in a conventional
commercial requirements. The new pen Experimental Methods single-level pen measuring 1.83 x 11m.
was designed to accommodate the entire The two-level pen was tested with four This pen gave about 20 m2 (0.84 m2/pig)
growing-finishing phase and to fit the com consecutive groups of 24 pigs each in a of floor space, with 60% slotted, similar to
mon 11-m-wide hog barn. This paper heated, insulated barn. This stocking rate the two-level pen in total area and solid-to-
describes the design, discusses costs, and provided about 0.88 m2 of floor space per slotted ratio.
compares performance of pigs housed up pig, in accordance with recommended To determine how pigs used the two-
.^a%^^ 0fcf
IF ' FEEDERS
. _ ' *
^rrnnrrir
T$y\i TrrrrrrrrrrTPrripnprpnnr1"
mi 111II11111111111111! ij
ii.m.UAim
i
UIZ
SH
W\ W
/"'' iS^^^Pl'
By 1 LLi
S s s 13a a 3 oatfaMiiejsssoi'S
E
n O)
> t
R
ffiwS^^^^^I
SECTION A-A
z
> B 0.76 m |- B
; .;
A_ 1" "1-
>
z C c
> PROTECTIVE
c BARRIER
n
c fV
73
> ii
r
c c
z
g FEEDERS
z
m ! II I I
J'^
z
p SECTION B-B
*
1.83 m t-
< A
c
-
z
c Figure 1. The design and dimensions of the free-access, two-level pen.
us
a
2
2
en
90
TABLE I. PERFORMANCE VALUES (MEAN TABLE II. PERCENTAGE OF TIME SPENT The final installed cost of the second
SD) OF 96 GROWING-FINISHING PIGS IN ACTIVE AND RESTING IN THE TWO- level under farm economic conditions is
A FREE-ACCESS, TWO-LEVEL PEN, LEVEL PEN
difficult to estimate, but certainly must
COMPARED WITH SIMILAR GROUPS IN A
CONVENTIONAL PEN Time (%)t include labor, transportation, and extra
costs for small tools and fasteners and a
Level Active Resting Total
Pen design profit for the installer. As a first approxima
Upper 12.6(10.4-14.3) 31.4(12.2-46.3) 44
Variablet Two-level Conventional tion, these latter items can reasonably be
Lower 9.4 (5.5-14.0) 46.6(37.8-59.4) 56
Total 22 (15.9-28.3) 78 (71.0-84.4) 100
expected to result in an installed cost
Days to market 153.3 154.6
Initial weight (kg) 25.64.35 24.43.01
which is double the materials cost (i.e.
tMean of four groups, and range of group means.
Final weight (kg) 99.25.74 96.75.84 $140 per m2). This compares favorably to
Gain (g/day) 878 82.2 849 94.4 $230 per m2 for purchasing new swine
0- to 7-day gain extrapolated to a 12-h period of daytime building space (Bunn and Kains 1986).
(kg) 5.461.56 5.090.84 activity, suggest an average of five to six Where stocking density of a barn is
Feed conversion
ramp uses per pig per day. However, some increased by installation of two-level pens,
efficiency 2.70 0.095 2.72 0.054
Carcass index 100.7 3.82 100.63.11
pigs use the ramp more than others, while design ventilation rate would have to be
a few did not use the ramp at all until forced increased, and the air inlet configuration
tStandard deviations with n = 96 pigs (except feed
to do so. Ramp design features need to be might have to be altered to suit the two-
conversion efficiency, where n = 4 pens) are shown
to illustrate individual variability in performance in improved so that all pigs will use the ramp level system. Studies of ventilation
the two pen types. spontaneously. Work is underway to deter requirements are needed under full-scale
mine pig preference for ramp design fea conditions and with a variety of manure-
level pen, observations were made tures such as slope, width, step spacing handling systems.
between 0730 and 1000h for 1day approx and illumination.
imately every 2 wk during each trial. The Pigs dunged only occasionally on the
number of pigs active and resting and num upper level, and the dung did not accumu CONCLUSIONS
bers on each level were recorded at 10-min late because it was usually pushed off the Performance results indicate that two-
intervals. The number of trips up or down platform by normal pig traffic. level housing has no adverse effects on
the ramp was counted throughout the 2.5-h In this study, the pigs rested on the average daily gain, feed conversion effi
observation period. upper level much more than in our pre ciency and grade index of pigs grown from
vious study (Fraser et al. 1986) which 25 kg to market weight (100 kg) at a stock
RESULTS tested the smaller 2.4 x 2.4-m, two-level ing density of approximately 0.88 m2/pig.
Pigs performed very well in both the pen. The nearly twofold difference proba The cost of new floor space provided by
two-level and the conventional pens bly can be attributed to the much larger adding a second level is about 60% of the
(Table I). Days to market were similar for undisturbed area available for resting in cost of new building floor space. Further
both groups, averaging 153.3 for the two- the longer pen. This larger space seemed studies are needed on the ventilation of
level and 154.6 for the conventional pen. better suited to the pigs' resting habits as two-level housing systems, and to identify
Final market weights averaged slightly the animals appeared to prefer huddling in and quantify possible improvements in the
lowerfor the conventional group (96.7 kg) groups of six to ten rather than in groups of welfare of pigs housed in free-access, two-
as compared to the two-level group (99.2 two or three as required by the smaller pen. level pens.
kg). The two-level group had slightly bet The cost of adding the upper level was
ter average daily gain, feed conversion approximately $70/m2 for materials. The ACKNOWLEDGEMENTS
efficiency and 7-d weights, but the dif precast concrete panels are the main com We wish to acknowledge the helpful cooper
ferences were small and not statistically ponent and account for 70% ofthe material ation of the staff at the Animal Research Centre
significant. Seven-day weights indicate cost. Next, and about equal to each other in pig unit, and the Engineering and Statistical
how well pigs adapted initially to the two- cost, are the support structure for the plat Research Centre Shop staff for constructing the
level system. Carcass grade index was the form and the extra material required to two-level pen.
same for both pen systems. expand the partition to 1.55 m in height. In
Pigs distributed their time fairly evenly certain cases, the pen partitions could REFERENCES
between the upper and lower levels serve to support the platform. For exam ANONYMOUS. 1984. Recommended code of
(Table II). Because the feeder was on the ple, concrete pen partitions to a height of practice for the care and handling of pigs.
upper level, there was proportionally more 0.76 m could support the concrete panels, Agriculture Canada, Ottawa, Ont. Publ.
activity there than below. The lower level with spindle partitions on top. 1771/E.
was used mainly for resting, drinking and The barrier and ramp must each be BUNN, A. J. and F. A. KAINS. 1986. Sum
dunging. During the interval from 0730 to costed against the second level and, mary of pork production and costs, February
1986. Pork News and Views. Ontario Minis
1000 h the ramp was used an averageof 31 together, they account for 10% ofthe mate
times.
try of Agriculture and Food, London, Ont.
rial cost. A small area of the lower level is
Mar.-Apr. issue, p. 3.
required for the ramp and support posts FRASER, D., P. A. PHILLIPS, and B. K.
DISCUSSION and, strictly, the cost of this lost space THOMPSON. 1986. A test of a free-access
Pigs used the ramp willingly at all stages should also be assessed against the upper two-level pen for fattening pigs. Anim. Prod.
of growth. The behavioral observations, level. 42: 269-274.
Thacker, P. A. and E. M. Barber. 1987 Use ofplanned drafts inanattempt toreduce preweaning mortality inbaby
pigs. Can. Agric. Eng. 29: 197-200.
A total of 69 litters from crossbred sows (Large White x Landrace) of mixed parity was used to evaluate the
effectiveness ofafarrowing crate attachment (Medata Blowaway Unit) designed to reduce baby pig mortality by directing
an air stream directly under the standing sow. Use ofthis attachment was hypothesized tocondition piglets toavoid the
areaunder the standing sowandthereby reduce theirchances of being crushed if the sow moved. All sows received the
same treatment during the experiment except that a Blowaway Unit was attached tohalf ofthe farrowing crates a minimum
of3days before parturition and remained there until 10 days postpartum. Sows farrowing in crates fitted with a Blowaway
Unit crushed significantly (P < 0.05) fewer of the total number of piglets born alive than did sows farrowing incrates
without a Blowaway Unit (5.3 vs. 2.3%). However, mortality occurring as a result of starvation andother miscellaneous
causes was slightly higher for sows farrowing in crates fitted with a Blowaway Unit than without (11.5 vs. 10.6%).
Therefore, under the conditions of this experiment, preweaning mortality was not significantly reduced by using the
Blowaway attachment.
INTRODUCTION mine its effectiveness in reducing pre away Unit was attached to their farrowing
Preweaning mortality is a source of weaning mortality. crates a minimum of 3 days before parturi
majoreconomic loss to the swine industry. tion and remained there until 10 days
It is estimated that an average of 15-25% MATERIALS AND METHODS postpartum. Two Blowaway Units were
of all pigs farrowed die before weaning This experiment was conducted from utilized during the experiment and were
(Fahmy et al. 1978). Most of these deaths April to October, on a 60-sow, farrow to moved from crate to crate as required.
result from crushing by the sow, chilling, finish operation located near Vilna, The Blowaway Unit (Fig. 1) comprised
starvation, weakness or disease (Bille etal. Alberta. The farrowing facilities consisted a small squirrel cage fan (76 mm diameter,
1974; English and Smith 1975). In con of two 4.6 x 12.2-m farrowing rooms 64-mm-wide blade) powered by a 115-V
trast, some production units have a pre located on either side of a 1.8-m hallway. motor rated at 0.7 amps. The fan directed
weaning mortality below 10% (English et Each farrowing room contained seven air into a 0.9-m downpipe constructed
al. 1982). Therefore, it may be possible to 1.5 x 2.1-m wire-floored farrowing crates from rigid plastic tubing with outside and
save as many as one pig per litter through (Faromor Inc., Waterloo, Ont.) located inside diameters of 60.4 and 51.7 mm, re
improved facilities and management. over a sloped gutter (280-480 mm deep) spectively. The downpipe was joined to
Preweaning mortality can be reduced by which was flushed with fresh water twice a the center of a horizontal 1.4-m length of
providing a heated creep area for the baby day to remove fecal material. The farrow plastic tubing which had 13 equispaced
pig (Adams et al. 1980). This supplemen ing crates were equipped with heated front 20-mm-diameter holes drilled in it at 15
tary source of heat encourages the piglets creeps maintained at 27C with heat and 30 angles off a horizontal plane.
to lie away from the sow and therefore provided from hot water pipes located in The horizontal tube of the Blowaway
reduces their chances of being crushed the concrete floor of the creep. Unit was mounted on the lowest bar of a
when the sow moves (Titterinton and Incoming air was drawn from the attic farrowing crate. The unit was secured to
Fraser 1975). Unfortunately, the newborn of the swine barn and entered the farrow the crate by means of an aluminum box-
pig tends to lie against the sow's udder for ing room through an adjustable styrofoam section bracket and two clamping screws
the first 2 or 3 days of life even when baffle located on the interior wall. Air was fitted with thumb screws. This mounting
heated creep areas are provided (Petherick exhausted through two 305-mm-diameter, system enabled a Blowaway Unit to be
1983) and it is during these first 3 days that variable-speed fans located on the exterior installed on a farrowing crate in less than
the majority of mortalities occur (English wall. Hot water heating was utilized to 5 min.
and Smith 1975). maintain the temperature of the farrowing An infrared sensing device was situated
A farrowing crate attachment (Medata room at222C. on the top of the downpipe. The infrared
Blowaway Unit, Medata Sales, Elmira, A total of 69 litters, obtained from beam of the Blowaway Unit was adjusted
Ont.) has been developed recently which crossbred sows (Large White x Land- so that it passed just above the sow's back
is designed to reduce baby pig mortality by race) of mixed parity, were tested during when it was lying down. A 7.6-cm reflec
creating a draft directly under the standing this experiment. Sows were moved into tor disc, mounted on the top bar of the
sow. It was hypothesized that this draft the farrowing room on day 109 of gestation opposite side of the farrowing crate, re
would condition the piglets to avoid the at which time they were alternately turned the beam of light to the receiver.
area under the sow and to utilize the creep assigned to either the control or the experi Whenever the sow stood, the beam of light
as their sleeping/resting area. The purpose mental group. Sows in the experimental was broken, which activated the blower
of this experiment was to evaluate the group received the same treatment as those and directed a 12.45 m/s air stream (range
blowaway attachment in order to deter in the control group except that a Blow 9.5-15.0 m/s) under the sow's body.
HEAT LAMP
(in creep area)
When the sow lay down, the air stream TABLE I. PERFORMANCE OF SOWS AND LITTERS
was automatically switched off. Use of a Blowaway SEt
Control
solid partition between farrowing crates
prevented piglets in adjacent crates from Number of litters 35 34
Average sow parity 4.42 5.02 0.51
coming in contact with the airstream. 0.15
Pigs born dead per litter 0.91 1.05
Air temperature at the Blowaway inlet Pigs born alive per litter 11.25 11.26 0.43
and at the discharge holes was measured Birth weight (kg) 1.40 1.41 0.03
using a Fluke multimeter and temperature Pigs weaned per litter 9.45 9.71 0.30
Weaning weight (kg) 7.30 0.18
sensing attachment and was the same at 7.35
Mortality (%) 15.97 13.76 2.47
both locations (22C). Total air flow, esti
mated using an aspect ratio of 0.6, 13 t Standard error of the mean.
outlet holes 20 mm in diameter and an
average air velocity of 12.45 m/s, was treatment as necessary. Heat lamps were and immersing them into water.
30.5 1/s. provided in the creep area for the first The experimental data were tested for
During gestation, all sows were housed week following parturition. Fresh creep normality by measuring skewness and kur-
individually and received between 1.8 and feed (18% crude protein) and water were tosis following which it was subjected to
2.7 kg of feed per day depending on con provided after the piglets reached 10 days analysis of variance according to the pro
dition. During lactation, the sows were fed of age. cedures outlined by Snedecor and Cochran
ad libitum the same ration as they received Records were kept of the number of pigs (1967). The incidence of specific causes of
during gestation. Chemical analysis ofthe born alive, born dead and weaned. Birth mortality were analyzed by Fisher's exact
gestation-lactation diet (Association of weight and weaning weight were also method (Keeping 1962).
Official Analytical Chemists 1975) indi recorded. In addition, mortality from birth
cated that the ration met or exceeded to weaning was tabulated. Piglet mortali RESULTS
National Academy of Sciences-National ties were subjectively grouped into pre- There were no significant differences
Research Council (1979) recommenda agreed categories by the barn manager. (P> 0.05) in the number of pigs born alive
tions for all nutrients. Causes of piglet mortality included overly or weaned between sows farrowing in
Pigletswere individually weighed with ing (piglets which were obviously crushed crates fitted with or without a Blowaway
in 24 h of birth and at weaning (average by the sow), starvation (showing signs of Unit (Table II). Sows on both treatments
weaning age 26 days). Routine herd man emaciation), genetic defects (ruptures, farrowed an average of 11.3 piglets born
agement procedures were followed, in spraddle legs and other deformities), born alive per litter. Control sows weaned 9.4
cluding clipping teeth and tails and ear weak (low piglet vigor at birth) and other piglets per litter while sows farrowing in
notching at day 1,administering ironinjec miscellaneous causes. Stillbirths were crates fitted with a Blowaway Unit weaned
tions and castrating at day 3 and antibiotic confirmed by removing the piglets lungs 9.7 piglets per litter. Total preweaning
198
CANADIAN AGRICULTURAL ENGINEERING, VOL. 29, NO. 2, SUMMER 1987
TABLE II. INCIDENCE OF PREWEANING different between the Blowaway and the crushing losses accounted for only 33% of
MORTALITY DUE TO SPECIFIC CAUSES t control group. Brabbs and Hogberg (1985) piglet mortality. Therefore, with a poorer
Control Blowaway also found a reduction in crushing losses environment and a lower level of manage
Crushed
but no significant difference in total pre ment, it is possible that more pigs would
5.32a 2.34b
Starvation 2.53 3.39
weaning mortality in an experiment have been saved as a result of using the
Genetic defect 0.76 1.04 designed to test the effectiveness of the Blowaway attachment. However, under
Born weak 4.82 3.65 "Portable Pig Sitter", an apparatus based the conditions of this experiment, pre
Miscellaneous 2.54 3.39 on a principle similar to that of the Blow weaning mortality was not significantly
t Percentage of totalpigsbornalive(394bornalivein away Unit. reduced.
the control group and 383 in the blowaway group). The fact that there was an increase in
a,b Means followed by different letters differ preweaning mortality due to starvation and ACKNOWLEDGMENTS
(/><0.05).
other miscellaneous causes for piglets far This study was supported by the Farming for
mortality averaged 15.97% for control sows rowed in crates fitted with a Blowaway the Future Program of Alberta. The authors
and 13.76% for sows farrowing in crates Unit suggests that at least some of the wish to thank Mr. Richard Scoffom for his
fitted with a Blowaway Unit (P > 0.05; piglets that were saved from crushing by excellent work in the care and management of
Table I). the Blowaway Unit later died due to other the experimental animals.
There were no significant differences causes. Therefore, if the full potential of
(P > 0.05) in the birth weight or weaning the Blowaway Unit is to be realized, addi REFERENCES
weight of piglets farrowed in crates fitted tional attention must be given to under ADAMS, K. L., T. H. BAKER, and A. H.
JENSEN. 1980. Effect of supplemental
withor without a Blowaway Unit (Table I). privileged pigs to ensure that they do in
heat for nursing piglets. J. Anim. Sci. 50:
The average birthweight of piglets for both fact survive until weaning and are not sim 779-782.
treatments was 1.4 kg while the average ply saved from crushing so that they can be ASSOCIATION OF OFFICIAL ANA
weaning weight of piglets was 7.3 kg. lost at a later date from other causes. Man LYTICAL CHEMISTS. 1975. Official
Sows farrowing in crates fitted with agement techniques that would appear to methods of analysis. 12th ed. AOAC, Wash
a Blowaway Unit crushed significantly have potential for use in combination with ington, D.C.
(P < 0.05) fewer piglets than sows far the Blowaway Unit would include cross- BILLE, N., N. C. NIELSEN, J. L. LARSON,
rowingin crates without a Blowaway Unit fostering (transferring piglets to another and J.SVENDSEN. 1974. Preweaning
(Table II). Sows farrowing in standard sow (Gadd 1985)), supplementing diets mortality in pigs. 2. The perinatal period.
Nord. Vet. Med. 26: 294-313.
crates crushed 5.3% ofthe total number of with commercial milk replacer (Hogberg
BRABBS, E. K. and M. G. HOGBERG. 1985.
piglets born alive compared with 2.3% in 1979) and creching (use of a supplemen
Effectiveness of the portable pig sitter on
crates fitted with a Blowaway Unit. tary rearing box (Gadd 1984)). baby pig loss. Michigan State University
However, mortality as a result of starva Piglets farrowed in crates fitted with a Report of Swine Research. 469: 26-28.
tion, genetic defects and other miscel Blowaway Unit became conditioned to ENGLISH, P. R. and W. J. SMITH. 1975.
laneous causes was slightly higher in avoid the air stream within the first 12 h of Some causes of death in neonatal piglets.
farrowing crates equipped with a Blow life and, thereafter, they moved to the Vet. Annu. 15: 95.
away Unit compared with the control creep area as soon as the Blowaway Unit ENGLISH, P. R., W. J. SMITH, and A. Mac-
(Table II). was activated. Even after the Blowaway LEAN. 1982. The sow Improving her
There were no problems with the Blow Unit was removed, the piglets continued to efficiency. Farming Press Ltd., Wharfedale
away Unit during the experiment. Care move away from the sow's body whenever Road, Ipswich, Suffolk.
FAHMY, M. H., W. B. HOLTMAN, T.M.
was taken to ensure that dust did not it stood. Since the majority of piglet losses MacINTYRE, and J. E. MOXLEY. 1978.
accumulate on the infrared reflector as this occur during the first 3 days of life Evaluation of piglet mortality in 28 two-
would have caused the fan to blow contin (English et al. 1982), it is likely that the breed crosses among eight breeds of pigs.
uously. Under normal use, it is unlikely Blowaway Unit could be removed from Anim. Prod. 26: 277-285.
that a build-up of dust would occur, since the farrowing crate approximately 4 days GADD, J. 1984. Creching is coming. Pig
most commercial farrowing rooms are not after parturition without reducing its effec Farming (Dec.) pp. 38-39.
excessively dusty and the Blowaway Unit tiveness. Therefore, one Blowaway Unit GADD, J. 1985. Fostering saves you piglets.
would be cleaned prior to installing on a could be used for about six farrowings a Pigs 1(3): 10-13.
clean farrowing crate. month which means that one unit would be HOGBERG, MG. 1979. Reducing baby pig
required for every 30 sows in the herd. losses through supplementation of sow's
milk with a milk replacer. Michigan State
DISCUSSION In assessing the potential of the Blow
University Report of Swine Research. 386:
The overall results of this experiment away Unit as a means of reducing pre 5-7.
indicatethat use ofthe Blowaway Unit can weaning mortality, it is important to point KEEPING, E. S. 1962. Introduction to statis
result in a significant (P < 0.05) reduction out that this experiment was conducted tical inference. Van Nostrod Co., Toronto,
inpreweaning mortality due to crushingby under excellent management and environ Ont.
the sow. Crushing accounted for 33% of mental conditions with the result that even NATIONAL ACADEMY OF SCIENCES
piglet mortality for litters farrowed in the control sows had a preweaning mor NATIONAL RESEARCH COUNCIL.
crateswithouta Blowaway Unit compared tality significantly below the level found in 1979. Nutrient requirements of domestic
with only 17% of piglet losses in crates most commercial operations. In a recent animals. No. 2. Nutrient requirements of
fitted with a Blowaway Unit. However, survey of eight commercial swine units swine. 8th rev. ed. NAS-NRC, Washington,
D.C.
more piglets were lost due to starvation located throughout the Canadian Prairies
PETHERICK, J. C. 1983. A note on nursing
and other miscellaneous causes in crates (VIDO Swine Technical Group 1985), termination and resting behaviour of suck
equipped with a Blowaway Unit than in crushing losses accounted for an average ling piglets. Appl. Anim. Ethol. 9:
the control. As a consequence, total pre of 42.2% of all piglet mortality whereas in 359-365.
weaning mortality was not significantly the control group of this experiment, SNEDECOR, G. W. and W. G. COCHRAN.
Douglas S. McGinnis
Engineering andStatistical Research Centre, Research Branch, Agriculture Canada, Ottawa, Ontario K1A 0C6
Contribution no. 1-875, received 16 September 1986, accepted 9 March 1987
McGinnis, Douglas S. 1987. Automated line-heat source system for the measurement ofthermal conductivity and
diffusivity. Can. Agric. Eng. 29: 201-207.
A system was developed for the rapid, automatic, and simultaneous measurement of thermal conductivity and
diffusivity of food materials and, with knowledge of the food density, for determination of its specific heat. The system
utilizes the line-heat source method for conductivity and diffusivity determination, and operates under the control ofa
microcomputer which is interfaced with a data logger for temperature and voltage measurements, and with external
electronics linked to a line-heat source apparatus. Experimental measurements are made automatically after the user has
entered the necessary experimental conditions and desired experimental precision to the microcomputer. The system was
found to be extremely simple to use, and capable of providing accurate measurements, together with statistical precision
statements, in 6 min orless for agar-water gel. The system software was programmed in IBM-BASIC (version 3.0). The
software was written to provide maximum user flexibility, and contains an optional algorithm (subroutine) for automatic
calibration of the diffusivity measurement apparatus.
INTRODUCTION nent must be identified on the basis of tem with these attributes and to demon
Thermal diffusivity values of foods are largest size and/or lowest thermal strate its usefulness in making precise
used in a variety of food engineering diffusivity. measurements of thermal conductivity and
applications involving heat and mass A number of methods for thermal diffusivity automatically and rapidly.
transfercalculations for freezing, cooling, diffusivity determination have been
or other thermal processing operations. reported in the literature. These methods THEORY
The thermaldiffusivity (a) of a food mate generally require solving equations The history and development of a math
rial depends upon its composition and describing a transient temperature field ematical basis for the line-heat source
physical properties and is related to con resulting from an applied constant or con method of conductivity and diffusivity
ductivity (k), specific heat (c), and density trolled heat source or sink. Most of the determination is presented extensively in
(p) by the relationship methods are indirect in that the thermal the literature (Chang et al. 1980;Mohsenin
a = k/(p-c)
conductivity, specificheat, and density are 1980; Nix et al. 1967). The essential ele
measured separately, sometimes under dif ments of this theory are given here for
Thermal property values are reported in ferent experimental conditions, and a is discussion purposes.
the literature for a number of selected food calculated. A popular method for food Considering only radial conduction heat
materials, but for an ever increasing inven materials involves using a thermal con flux in an infinite cylinder, the tem
tory of processed and/or manufactured ductivity probe (line heat source) in con perature-rise at any point within that cylin
food products, thermal property data are junction with one or more temperature der following a suddenly initiated,
not available or are simply not reported. sensing probes at fixed radial distance(s) constant heating rate, line-heat source is
For foods undergoing thermal steriliza from it, as described by Mohsenin (1980) given by the following equation (Lowan
tion or pasteurization processes, the for and others. This technique has the advan 1940):
mulation of the product and the thermal tages of reasonably short test duration,
properties of its components comprise a small temperature change of the product, 7X6) = -ln0
(-Q"P2"1
set of critical factors which determine the and the ability to obtain the conductivity
2ttR 2 (2/i)(!)J (1)
thermal response characteristics of the and specific heat under the same test con For points close to the line-heat source
product under the applied heating condi ditions. The technique is well suited to such that p ^ 0.16, it has been found that
tions. Singh (1982) and others have dis biological materials, especially those Eq. 1 can be expressed to better than 1%
cussed the importance of using the which are significantly altered by changes error by using the first two terms (Hooper
component thermal diffusivity values in in temperature. and Lepper 1950), hence:
numerical or analytical solutions to heat Despite the short test times for this
transfer in composite foods, as opposed to method, extensive computations for the
the potentially erroneous use of "effec determination of the thermal diffusivity
7X6) = JL\zK
2-jtA'l 2
-Inp] (2)
tive" or "apparent" thermal diffusivity must be carried out using a trial and error The change in temperature for this point
values. It is therefore important that the method on a digital computer, and transfer between times 0, and 62 is, therefore:
thermal processing specialist know the ring data to the computer can be time con
thermal properties of each component suming. It was recognized that a 7(0.) - 7Xe2) = (fi/(4ir*)) ln(0,/e2) (3)
when applying that knowledge in the significant time saving could be realized if The conductivity may then be deter
design of processes to ensure food safety control of the experimental apparatus, mined from Eq. 3 knowing the heat source
and optimum quality. This is particularly gathering of the data, and subsequent data strength and the temperature versus time
true in the practical development of ther analysis could all be performed by a single relationship. Use of a time correction fac
mal processes for heat-sterilizable foods desktop computer. The objective of the tor has been widely accepted as a means of
wherein the slowest heating food compo work presented here was to develop a sys compensating for the finite diameter and
TEMPERATURE CONTROLLER
AND PUMP
OIL RESERVOIR fr
WATER OUT
Figure 2. Schematic layout of automated system for conductivity and diffusivity measurement.
can give accurate thermal conductivity put to a relay (SPDT, R10S-E1Y1-J10K, temperature pairs used in the regression
measurements in the shortest possible time Potter and Brumfield) which closes the may be specified by the user to better esti
(e.g. =^10 sec), with the advantage of mini heating circuit. The circuit and software mate the straight-line portion ofthe curve.
mizing temperature changes in the product were designed to permit setting of the In the final phase of the program, the user
near the heating probe, yielding a closer probe voltage prior to probe energization. is requested to provide values of param
estimate of the k value associated with the A simplified logic flow chart ofthe com eters by which the diffusivity can be deter
mean sample temperature during the test. puter program is presented in Fig. 3. Ini mined by nonlinear least squares re
A shorter test time may also be advan tially the user is prompted to input the gression, fitting Eq. 1 to the observed pas
tageous for some products which undergo desired experimental conditions. The pro sive probe temperature rise values. The
irreversible changes in thermal properties gram next monitors temperature condi sample density, number of regression data
at the prescribed test temperature. tions until a state of thermal equilibrium, points to use, the acceptable resolution
However, to obtain a sufficient temperature as defined by the user, is achieved between (maximum estimation bracket size) of the
rise at the passive probe location in such a the two probes and between the probes and best fit solution (a1), and a starting esti
short time for accurate thermal diffusivity the constant temperature fluid. Once equi mate of the specific heat for the iterative
determinations, considering temperature librium is attained, the user is prompted to solution, are supplied by the user. The
measurement precision (resolution initiate probe heating through key entry. "Golden Section method" as described by
0.1C) and error, would require that the The heating probe element is energized, Himmelblau (1972) is used for the uni-
probes be located very close together (e.g. and data are logged until the user-specified dimensional search algorithm for a1.
r ^ 1cm), and/ or that the heating power be test duration is completed, at which time Confidence limits for the experimental
increased substantially (e.g. Q 2^ 30 the probe is de-energized. The raw data are curve fit of Eq. 1are developed by a second
W/m). The finite diameter ofthe two pro then transferred to temperature, time, and Golden Section search algorithm. This
bes (e.g. 1.7 mm), and their combined voltage arrays. The resistance ofthe heat algorithm finds a value of a such that:
influenceon heat flow in the vicinity ofthe ing probe is calculated for each probe tem
probes imposes a limitation on the probe perature and used with the known voltage (T(a) =a(a')-(V 1+'(F(l,
m \
m- 1, 1 7))
spacing. In addition to this, the probe to calculate the average probe heat strength (5)
spacing is limited by the particle size and (Q) for the test (the constantan heating
uniformity of the product. For these rea element resistance varies slightly with where
sons, the potentially high rate of tem temperature, in a linear fashion). The slope
o-(a) = X (T -/(&, a))2,
perature data acquisition possible using of the probe temperature rise versus
this configuration was considered to yield ln(6 60) curve is determined by linear and
no significant advantage for this applica least squares regression, and the thermal
tion. conductivity is calculated (Eq. 4). Since a1 = value of a such that a is minimized.
Heating of the conductivity probe is ini heating probe temperature rise values dur The probability level (7) in Eq. 5 is only
tiated by the user from the computer key ing the first few seconds following the start approximate where/is nonlinear (Draper
board. This operates by way of a of probe energization can be dramatically and Smith 1981). However, Eq. 1 was
LABSOFT command to the A/D, D/A affected by the initial temperature found to be highly linear (linear/?2 = 0.95
converter producing a latched binary out response of the probe, the range of time- 0.99) for extended segments of the
NO YES
CALCULATE THE PROBE N.2 TEMPERATURE RISE FOR THE LAST TIME VALUE
USING THE SPECIFIC HEAT ESTIMATE; COMPARE THIS VALUE TO THE RECORDED
VALUE, AND REPORT THE ESTIMATION ERROR
INCREASE NO.
DATA POINTS FOR
REGRESSION
Figure 3. Logic flow diagram ofprogram used to control the thermal property measurement system and to analyze the experimental results.
204
CANADIAN AGRICULTURAL ENGINEERING, VOL. 29, NO. 2, SUMMER 1987
TABLE I. NOMENCLATURE
curve following the first 60-120 sec for the
Variable Definition Units selected range of experimentaldesign con
E Euler's constant = 0.5772157 ditions (i.e. probe spacing, heat source
F Frequency statistic of the F-distribution strength, etc.).
f Predicted temperature-rise using Eq. 1 C The confidence limits so developed
k Thermal conductivity W/(mC)
m Sample size provide the user with a measure of the
n Positive integers uncertainty in the results owing mainly to
Q Heat source strength (W/m) the imprecision of the temperature mea
r Distance from line heat source m surements. In order for the user to resolve
R2 Coefficient of determination (0-1)
S Slopeof temperature-rise vs. In (0 - 60) and correct for systematic errors, however,
for heating probe C
the user must first calibrate the apparatus
T Temperature rise C using a substance of known thermal con
Observed (/th) temperature-rise C ductivity and diffusivity. In this regard an
f Sample arithmetic mean temperature-rise C optional part of the computer program
Thermal diffusivity m2/s allows the user to automatically determine
Bestfit solution of thermal diffusivity using
Eq. 1 the time correction factor for the probe
m2/s
P r/(2(a0),/2) apparatus. In this operation the user
7 Probability level for null hypothesis provides the true (known) thermal con
(0.05 for 95% confidence limits) ductivity of a test material such as 0.5%
e Time from initiationof line heating
e0 Time correction factor
s
agar-water gel and the program finds, by
s
a Sum of squares oc2 trial and error, the time correction factor
for the probes that permits a match, to a
user-specified precision (0.01%, say) of
the measured and true thermal con
ductivity.
TABLE II. PROGRAM OUTPUT FILE PRODUCED BY THERMAL CONDUCTIVITY/
DIFFUSIVITY MEASUREMENT SYSTEM
Equation 1 is highly sensitive to probe
separation distance, and it will be neces
THERMAL PROPERTY REPORT DATE: 86-08-20 sary to either correct or check for minor
Test sample: 0.5% AGAR Calibration run
Datafile: A:TEST1.DAT
disturbances from true parallel alignment
Applied pressure (kPa) = 0.00
of the probes, or for off-center position of
RAW DATA: the thermocouples within each probe. The
Heating Passive corrected probe separation distance can be
Data set Time Voltage Heat Bath probe 1 probe 2 7-Rise obtained by trial-and-error calibration
No. (sec) (V) (W/m) (C) (C) (C) (C) using a homogeneous substance of known
0 0. 0.000 0.000 28.1 28.2 28.2 0.0 thermal diffusivity. Systematic or random
6 30. 9.990 28.920 28.1 40.7 28.2 0.0
12 60. 9.980 28.868 28.1 42.9
errors which cannot be corrected directly
28.3 0.1
18 90. 9.970 28.813 28.1 44.3 28.4 0.2
through calibration must be estimated by
24 120. 9.960 28.758 28.1 45.2 28.6 0.4 the usual experimental replication and sub
30 150. 9.950 28.702 28.1 45.9 28.9 0.7 sequent calculation of a mean and standard
36 180. 9.950 28.703 28.1 46.4 29.1 0.9 error for a and k.
42 210. 9.940 28.646 28.1 46.9 29.4 1.2
48 240. 9.930 28.590 28.1 47.3 29.6 1.4
As a means of assessing the relative
54 270. 9.930 28.591 28.1 47.7 29.8 1.6 degree to which the experimental passive
60 300. 9.920 28.534 28.1 48.0 30.0 1.8 probe data fits Eq. 1(for a), and the degree
66 330. 9.920 28.535 28.1 48.3 30.2 2.0 to which the heating probe data fits Eq. 4
(for k\ a coefficient of determination (R2)
Averages 9.949 28.695 28.1
value is calculated for the nonlinear and
SUMMARY:
Conductivity (W/mC) = 0.6115
linear regressions. Of the eight alternative
Density (kg/m3) = 996.49 and uniquely different expressions for R2
Specific heat (J/(kg C) = 4161.84 identified by Kvalseth (1985), the follow
(4137.80-4186.17) 95% CL* ing recommended form was selected for
Diffusivity (m2/sec) = 0.1474E-06
both regressions:
(0.1466E-06 - 0.1483E-06) 95% CL*
* CL = Confidence limits
Conductivity parameters
i - X if, - Tf-/Z (t - T)2 (6)
2.1-
2.0-
1.9-
5 1.7H
<
QC
i.bH
5
1.5-
1.4-
1.3-
1.2-
m ri 111TtiniM i ii m 11 m i rnii 1111 ii i rji 11 m 11111 fi m int 1111111 riij 1111111 ii j11111111111 ii 11 m ii 11 n mi 11 j11111 m 11 j11 ni 111 i j11111 n 11 j111111 n 11 n 11 n 1111
200 210 220 230 240 250 260 270 280 290 300 310 320 330 340 350 360
TIME (s)
Figure 4. Experimental and predicted temperature rise of the passive probe as a function of heating time.
However, the optimum heating time could predicted temperature rise curve. Figure 4 damental electronic components into a sin
vary significantly depending upon the pro shows the two temperature rise profiles gle package, with the attendent advantages
bes used, their separation distance, the (dashed lines) for the upper and lower 95% of lower system cost and portability.
heating power applied, and the substance confidence limits of a. These curves are However, the system can be easily
being measured. Table II reports all rele not to be confused with 95% confidence assembled from components which may
vant information pertaining to the calcula contours for the temperature-rise versus already exist in many well-equipped ther
tion of the thermal conductivity and time data. The approximately uniform pat mal process development laboratories.
diffusivity respectively. The R2 value for tern of the temperature-rise error residual
the heating probe temperature versus loge (measured-predicted) indicates that Eq. 1 ACKNOWLEDGMENT
(time) curve was found to be no less than provides a good fit to the data, and The author is grateful for technical support
0.998 for the tests on 0.5% agar water gel provides evidence of good equipment per provided by Dave Dow of the Engineering and
Statistical Research Centre (E.S.R.C), and for
using 20 data points. Because the tem formance.
assistance provided by associates and staff of
perature rise for the passive probe is low It was found that approximately 12itera E.S.R.C. on this project.
compared to that of the heating probe, and tions are required to obtain a to within
because ofthe limited precision ofthe tem 0.1% of the "best-fit" solution using 50
perature measurements (0.1C), the terms of Eq. 1 and 20 temperature-rise vs. REFERENCES
number of data points required for accurate time data pairs. CHANG, C. S., F. S. LAI, and B. S. MILLER.
diffusivity determination was found to 1980. Thermal conductivity and specific heat
increase with decreasing test time or with SUMMARY
of grain dust. Trans ASAE (Am. Soc. Agric.
Eng.) 23: 1303-1306, 1312.
decreasing overall temperature rise of the An automatic system for the measure
DRAPER, N. R. and H. SMITH. 1981.
passive probe. Table II provides the ment of thermal conductivity and dif Applied regression analysis. 2nd ed. John
residual sum of squares value and 95% fusivity has been successfully developed. Wiley and Sons, Inc., Toronto, Ont.
confidence limits for this regression. Fig The development system reported here HIMMELBLAU, D. M. 1972. Applied non
ure 4 shows the last 25 measured tem could be further refined and redesigned linear programming. McGraw-Hill Book
perature rise values from the data towards production of a more compact Co., Toronto, Ont.
presented in Table II, together with the commercial system by combining the fun HOOPER, F. S. and A. R. LEPPER. 1950.
R. P. Rudra
Rudra, R. P. 1987. A curve-fitting program to stress relaxation data. Can. Agric. Eng. 29: 209-211.
A computer program has been developed for the analysis of stress relaxation data using the method of successive
residuals. This technique eliminates the use of graph paper and eyejudgement for developing a multi-term exponential
stress relaxation equation. This program hasthepotential to select best fit multi-term exponential equations to exponential
decay data.
INTRODUCTION terms of exponential functions. The most stress values for equal time intervals. The
Viscoelastic materials show a stress commonly used method is that of suc spline function is also available in a sub
relaxation phenomenon which is one ofthe cessive residuals. This method, however, routine form in the International, Mathe
most important parameters in characteriz is rather time consuming, and the accuracy matical and Statistical Library (IMSL)
ing materials (Ferry 1961). In stress relaxa using graphical technique depends upon package. The incremental slope or change
tion the test specimen is suddenly brought personal judgement. It is the purpose of in stress values and time arrays are scan
to a given strain, and the stress required to this study to develop a suitable computer ned backward to arrange them in a
hold the deformation constant is measured program to be used to analyze the stress decreasing order of time. Equal incremen
as a function of time. The results are ex relaxation data by the method of suc tal slope or change in stress values for
pressed in terms of a time-dependent mod cessive residuals. constant time intervals will detect the
ulus^/) in tension or compression, G(t) in straight line portion ofthe stress relaxation
shear, or K{t) in bulk compression. These DESCRIPTION OF THE curve. The computer, however, performs
parameters can be used either in direct COMPUTER PROGRAM very accurate analysis resulting in very
form or in their reciprocal form called If the stress relaxation is expressed in large numbers of a and t. Since, for most
compliance in viscoelastic analyses. the general form of practical purposes, three exponential func
One rheological model representing tions are adequate for representing stress
stress relaxation is the generalized Max a(r) 0\. + 2/ 0~n
(3) relaxation data, the problem is solved by
well model as illustrated in Fig. 1. If this taking the difference in the adjacent incre
model is subjected to constant strain e0 at then the computer program begins by mental slopes or change in stress values
time t = 0, the total stress a in the model is looking at either slope or change in stress and comparing them with a very small
given by values corresponding to equal time inter value of tolerance. After obtaining the
<T = CT| + 0"2 + O3 + + CTn + (Tc (1)
vals in order to obtain the straight line points for the straight line portion of the
portion of the log-stress versus time curve log-stress versus time curve, the data
where ae is equilibrium stress. to compute an and Tn of Eq. 3 from the points are arranged in an increasing order
The decrease in stress with time can be slope and ordinate intercept, respectively. of time, and the equation of the straight
represented by The program takes the stress relaxation line portion is obtained by simple linear
v(t) = e0 [Edl e_,/T| + Ed2 e",/T2 + ... data and converts the stress or stress relax regression. The slope of the straight line
ation modulus into a logarithmic form. portion gives time constant t, and the
+ dn e",/Tn + Ec] (2) The semi-log curve is then divided into a intercept on the ordinate gives the coeffi
where t,, t2, ... Tn are relaxation times, large number of equal time segments. The cient of the exponential term a,. The first
TreI, corresponding to various elements in cubic spline function (Erh 1972) is used to residual is obtained by taking the dif
the model; E0 is the initial modulus; Ed is look at the incremental slopes or change in ference between the original stress versus
thedecaymodulus (E0 - e); and Eeis the
equilibrium modulus.
To predict stress at any time, /, the stress
relaxation curve must be represented by a
mathematical expression depicting the
behavior of the material under constant
strain. The plot of the logarithm of stress
versus time is generally nonlinear. It has
been found that a three-term exponential
equation is usually a reasonably good fit
for practical purposes (Mohsenin 1978). mi1?, igg^ igi^ nig
Various graphical techniques outlined by
Mohsenin (1978) and numerical analysis
-!*
methods given by Moore (1974) can be
used to express the relaxation curve in Figure 1. Generalized Maxwell model representation.
18.0 288.0 CO
W
300 _
20.0 280.0
H
where DS = incremental slope, and zero value of equilibrium stress. The data 0.0 344.20 0.0 81.89
DY = stress value. representing the stress versus time rela 0.41 275.60 0.36 46.99
0.82 219.50 0.72 18.85
(8) Compare the difference (DIFF) tionship were divided into 50 equal time
1.63 149.80 2.15 16.32
with tolerance. segments. With a tolerance value of 0.005, 2.86 99.26 2.87 15.19
(9) Take the points of log-stress and the program generated values of 385.82 3.67 79.45 3.22 13.49
time for which DIFF or DIF is less than and -65.469 for a, and t,, respectively. 4.90 65.02 3.94 9.30
7.27
tolerance to obtain the points for the The corresponding stress values for the 5.71 48.62 4.30
6.53 39.70 4.66 5.45
straight line portion of the log-stress ver first residual are given in Table II. The 7.35 31.85 5.02 3.83
sus time curve. repetition of the procedure resulted in 8.57 21.94 5.37 2.44
(10) Compute the values of a and t by values of 262.29, 82.218, -3.4464 and 9.39 16.81 5.73 1.30
simple linear regression. -1.4166 for a2, a3, t2 and t3, respec 10.20 13.22 6.09 0.43
11.43 10.36 6.45 0.17
(11) Compute the differences (STMD) tively. The stresses associated with the sec
12.65 8.19
between the original stress and the stress ond residual are also shown in Table II. 13.47 6.60
Na obtained from the straight line to The curve and equations developed from 14.69 4.48
obtain the first residual. this analysis are shown in Fig. 2. Com 15.51 3.29
parison with the experimental data yielded 16.73 1.68
STMD(I) = a (I) - Na (I) 17.55 0.76
very good agreement with the fitted
(12) If STMD is less than or equal to values, with a determination coefficient
zero, then omit that value. (r2) or 0.991. Statistical analysis indicated was used by manipulating the equilibrium
(13) Compute the natural logarithm of no significant difference between the fitted stress or modulus (ae or Ee) in a(0 or E(t).
the STDM values. values and experimental curve at the 0.05 Analysis performed on wheatdough witha
(14) Repeat steps 4-11 two more times level. value of 40 g for ae and a value of 0.005 for
to obtain terms for the second and third Some food materials do exhibit equi tolerance predicted the following equa
exponentials. librium stresses. In such cases this method tion.
210
CANADIAN AGRICULTURAL ENGINEERING, VOL. 29, NO. 2, SUMMER 1987
the technique of fitting the computer curve
to the stress relaxation data, based on the
method of successive residuals developed
in this study, is an easy and reliable
method for determining the exponential
functions representing given stress-relaxa
tion data. This method can handle any
number of exponential terms depending
upon the precision and accuracy required.
This method can also be used to obtain the
best exponential curve fit by calculating
the linear determination coefficient
between the values obtained for a different
number of exponential terms and the
experimental data, comparing the fitted
values with the experimental data, taking
the one with the highest determination
coefficient and not significantly different
from the experimental data as the best fit
curve. The method is simple, accurate and
less time consuming.
Air to air heat exchangers can be used to 87-116 HIGH PRESSURE VENTILATION
preheat ventilating air and hence reduce INLETS FOR ANIMAL HOUSING
heating costs for livestock barns. However,
frost accumulation is a major problem in this J.J. Leonard and F. Kloster, Dept. Agric.
application. Currently available frost control Eng., University of Alberta, Edmonton, Alberta
systems are based on some combination of T6G 2G6.
time, pressure loss, core temperature or
exhaust air temperature. They do not result in In order to obtain suitable trajectories of cold
an optimal rate of heat transfer, independent air jets for minimum ventilation during winter,
of barn temperature and relative humidity. high air velocities are required. Generally
The development and testing of a frost these velocities cannot be achieved using
control strategy is presented, based on a conventional exhaust fans and inlet designs.
measured instantaneous rate of heat transfer. One possible alternative is to use a high
The temperature rise of the supply air is pressure centrifugal fan to provide minimum
measured and the rate of heat transfer is ventilation during cold weather. Laboratory
controlled by positioning a damper to regulate measurements of air flows and distribution
the flow rate of cold supply air. Experiments using such a fan and a radially discharging
were conducted using a 470 1/s counterflow ceiling inlet are described. The results of
metal plate heat exchanger. Tests were these measurements are used to assess the
performed for a cold air temperature of -25C, feasibility of practical systems, and to suggest
and for return temperatures and relative possible designs.
humidities of 13 and 25C, and 40 and 75%,
respectively. The prototype controller was 87-117 PRODUCTION OF BIOGAS IN A 50 m3
confirmed to operate properly as predicted. DOWNFLOW FIXED FILM DIGESTOR
Three control parameters were identified as
Two groups of interrill sediment transport K.A. Tan, R.P. Rudra, W.T. Dickinson,
capacity experiments were conducted using a D.E. Elrick and G.J. Wall, School of
0.7 by 1.5 m long laboratory flume, artificially Engineering, University of Guelph, Guelph,
roughened with a loamy sand soil glued onto Ontario NIG 2W1.
the inner surface. In the first group, uniform
laminar flows were produced from a constant Rainfall infiltration experiments into a three
head tank, while in the second group, flows layer soil system were conducted in the
were produced by a single nozzle rainfall laboratory under simulated rainfall conditions.
simulator centered over the flume. Three A low permeability layer was simulated using
slopes of 2, 9 and 20% and three discharges a metal sheet having a known number of holes
or rainfall intensities of 0.5, 3.0 and 10.0 to transmit water at a low rate. Tests were
1/min were used in a three by three factorial performed using two rainfall rates of 16 and
complete block design with two replications. 39 mm/h, three depths to the low permeability
The loamy sand soil was injected into the layer of 20, 50 and 100 mm and two types of
flow from an overhead hopper, and flow low permeability layers having conductivities
velocity, water and sediment discharge and of 0.015 and 0.008 of the conductivity of the
transported sediment size distribution were overlying layer. A Green-Ampt infiltration
measured. model was applied to this system, with the
E. Ouellette, G.S.V. Raghavan, Dept. Agric. The technical feasibility and economics of a
Eng., and R. Reeleder, Dept. Plant Sci., ground water heat pump system for a
Macdonald College of McGill University, Ste. temperature controlled potato storage facility
Anne de Bellevue, Quebec H9X ICO. was studied. A prototype heat pump system
was installed and tested. A mathematical
Given favorable conditions, some vegetable model was then developed and used to
storage diseases have the potential of simulate the proposed system, to obtain the
producing extensive losses in both the quality optimum performance parameters. The
and quantity of products, if they are not optimum economic heat pump power, defined
detected on time. Today, disease detection in terms of total cost to own and operate the
still relies on mostly visual and olefactory system, was determined using climatic and
signs, and on the ability of the storage economic data.
manager to recognize them. Monitoring
techniques that could identify and measure 87-410 C 0 M P A RIS 0 N BETWEEN A
abnormal crop stresses are needed so that CONVENTIONAL COOLING SOURCE
corrective measures can be applied to AND STORED WINTER COLDNESS
minimize losses. FOR PRECOOLING VEGETABLE
The volatile profiles emanating from stored PROCESSING
Insects can be grown as livestock to produce P. Gitelman, President, U.F.L. Foods Inc.,
animal protein for human food. More than a Toronto, 0 ntario.
million species are available in this taxonomic
Events at the Annual CSAE Meeting together with the Engineering Institute of Canada
Canadian Engineering Centennial convention, May 18 - 22,1987,
le Palais de Congres, Montreal, Quebec
CSAE affiliates officially with the ASAE. CSAE and ASAE Eric Norris is Master of Ceremonies at the CSAE Annual
Presidents, Don Allen and Bill Johnson sign a letter of agree Banquet, McGill University Faculty Club, May 20.
ment at the CSAE Council meeting, May 19, 1987.
Ron Britton is the incoming President of CSAE. He also was Walter Bilanski of the University of Guelph was presented the
presented with the Canadian Sheet Metal Building Institute Maple Leaf Award.
Award.
ASAE President, Bill Johnson, gives a thorough speech on the Phillipe Savoie, formerly the Quebec Regional Director, be
needs and challenges for agricultural engineers in the future. comes Vice-President (Technical).
31st Canadian
Agricultural
Organised by the
Canadian Society of Engineering
Agricultural Engineering
as part of the
Conference
Agricultural Institute of
Calgary, Alberta
Canada Annual Conference
August 21 -24, 1988
Sessions are planned in the following areas
Soil and Water Structures and Environment
Coordinator: Dr. D. Chanasyk Coordinator: Mr. R. Borg
University of Alberta Alberta Agriculture
Department of Soil Science 4920 - 51 Street
Edmonton, Alberta T6G 2G6 Red Deer, Alberta T4N 6K8
Papers presented by graduate students are eligible for the graduate student paper competition (if eligible, please
indicate specifically in a covering letter).
Authors will be notified of acceptance or rejection of papers by March 1988.
Send abstracts of approximately 250 words of the proposed papers, before November 1, 1987
to : Dr. P. Savoie
C.S.A.E. Vice-President, Technical
Agriculture Canada
Departement de genie rural
Universite Laval
Sainte-Foy, Quebec G1K7P4
JOURNAL OF AGRICULTURAL
ENGINEERING RESEARCH
Editor: D. J. White
Ministry of Agriculture, Fisheries and Food, London, U.K.
The Journal ofAgricultural Engineering Research reflectsthe broad spectrum of interdisciplinary interests
inherent in this field. It publishes mainly original research papers but also includes in-depth review
articles and short research notes.
Published for The British Society for Research in Agricultural Engineering by Academic Press London
Recent Contents: Ventilation of Livestock Buildings by Natural Convection, M.P. Foster, MJ. Down. A
Resistive Probe Moisture Sensor for Tropical Root Crops and Vegetables, G.O.I. Ezieke. The Thermodynamic
Performance of a Continuous-flow Fluidized Bed Grain Disinfestor and Drier, G.R. Thorpe. Modelling the
Performance of a Cross-flow Grain Drier, M.E. Nellist. Physical Changes in Stored Bulk Rice, M.C. Gough,
H.S. Cheigh, S.K. Kim, T.W. Kwon. The Compaction of a Partially Dried Lucerne, D. Nason.
Address
Date
The Editorial Board will assess suitability table, graph, chart or photograph. In general, illustration or group of them should be planned
and essential detail of papers submitted for pub follow the Council of Biological Editors Style to fit, after reduction, onto a space 58 mm wide
lication in Canadian Agricultural Engineering. Manual, 5th ed., published by the Council of (one column) or 122 mm wide (two columns).
One or more reviewers will be used. Their Biology Editors, Inc., Bethesda, MD 20814. The original should be not more than three
comments and suggestions will be compiled times the size of final figure. The number of
and submitted to the author. The review will References engravings required, and therefore the cost to
ensure that: List references alphabetically by authors at the author, can be reduced if two or more draw
the end. Include year of publication, title in ings can be grouped together. For identifica
1. A research paper presents a piece of re
lower case except first letter of first word, and tion, the figure number and author's name
search carried to a well-defined stage of
source, with volume and page numbers where should be written on the lower left corner with
advancement and the conclusions are ade
applicable. Names of periodicals should be soft pencil.
quately supported by the experimental
results. abbreviated in the form given in BIOSISList of Line drawings should be carefully made in
Serials with Title Abbreviations (Bioscience India ink on white drawing paper or tracing
2. A technicalpaper presents a clear, concise,
Information for Service of Biological linen. Letters, numerals, labels and axis cap
and factual outline and interpretation ofthe
Abstracts, 211 Arch Street, Philadelphia, Pa, tions should be made in capital size with a
development, design, test, or analysis under
19102). Material in press, with the name ofthe stencil or lettering set, not typewritten. Letters
consideration and that it is a contribution in
journal, may be used as a reference. Private and numerals must be at least V/2 mm high and
the field of agricultural engineering.
communications and unpublished reports preferably 2 mm high in final form. Curves on
3. A general paper on education, research, or should be referred to in parentheses in the text. graphs must be 0.3 mm wide after reduction.
extension is pertinent to major changes in Avoid the use of footnotes. Use the author-date Axes and grid lines should be clearly visible but
curriculum, research, or extension or to for system in the manuscript when referring to arti inconspicuous; a width of 0.2 mm after reduc
ward-looking developments in these areas. cles in the Reference section. tion is suggested. Figure numbers and captions
4. A technical note, of one journal page or should be typed on a separate page, not on the
Tables original illustrations. When a paper is sub
less, on equipment development, technique
of measurement, or method of analysis will Designate tables at the top by table number mitted for publication, the original illustrations
have an application for other workers in the (Roman numerals) and title, all in capital let need not be provided so long as the copies are of
field of agricultural engineering. ters. All headings and other information in such quality that reviewers can understand
tables are to be in lower case except first letter them. Original drawings must be provided
MANUSCRIPT
of first word. Keep the table compact and place when the paper is accepted for publication.
The manuscript should be typed double- it across the page wherever possible. Do not use
vertical lines.
spaced on paper 8Vi x 11 inches (21.6 x 27.9 DISCUSSIONS
cm) with margins not less than Wa inches (3.3 Discussions may be submitted on any paper
Measurements
cm). The first page should contain only the or technical note published in the Journal for a
title, authors' names, addresses (including Use the metric system (SI) in the text.
Tables, charts and graphs should be given only period of not more than four months following
postal codes), and contribution number where publication. Discussion of a paper or technical
in metric units.
applicable. Tables and captions for illustrations note is open to anyone who has significant
should be on separate pages, placed after the comments or questions about the content of the
text. Manuscript paper with numbered lines is Equations
Equations and formulas must be set up paper/technical note. A discussion will not be
preferred. The original and two copies are accepted for publication if it contains material
required. clearly. Use capitals for symbols as much as
possible and lower case for superscripts and readily found elsewhere, is purely speculative,
The title of the paper should be capitalized subscripts. Greek and other characters should introduces personalities or otherwise falls be
and centered on the page; it should give an be identified clearly. Equations should be num low the standards of a technical paper in a
accurate description of the article, using key bered on the right-hand margin in large num professional journal. Authors will be given an
words that can be used for computer-indexing. bers and in line with the center ofthe equation. opportunity to reply to discussions.
The format for discussions differs from those
ORGANIZATION Abbreviations of papers in that figures are to be identified by
This paper should be organized to conform For commonly used terms, consult the CBE capital letters to avoid confusion with the orig
with present Journal practice. Research and Style Manual. Abbreviate units of measure inal paper. The discusser should refer to him
Technical papers must include a short abstract only when used with numerals. Do not use self as "the writer" or "I" and to the author of
section of about 200 words. abbreviations in the title. Normally, numbers the original paper as "the author." The first
Major headings Center on the page with less than 10 should be spelled out, e.g. six. page shows the title of the original paper with a
all words in capital letters. footnote to identify the author, volume, page
Paragraphs and date. Name and address ofthe writer ofthe
Subheadings Start at left-hand margin, If paragraphs are to be numbered, designate discussion follow the title.
capitalize first letter of major words. by Arabic numerals. Designate sub-paragraphs Discussions will be reviewed by the Editorial
Sub-subheadings Start at left-hand mar by lower case letters in parentheses. Board and possibly the reviewers ofthe original
gin, in lower case except first letter of first paper. The length of a discussion is restricted to
word, and underline.
ILLUSTRATIONS one journal page. Lengthy discussions will be
Technical and detailed information should Either original drawings or clean, glossy returned for shortening, or the writer may be
be included only in the form of description, photographs are acceptable for illustrations. An encouraged to submit a paper or technical note.
Electrical Power and Processing
SIMULATION OF ADSORPTION DRYING OF CORN, WHEAT, BARLEY AND OATS USING SODIUM
BENTONITE
K.C. Watts, W.K. Bilanski, and D.R. Menzies 173
Food Engineering