Experiments on the spontaneous : oil have stow that the reaction is °: comnon hydrolysis; the so-called catalysed monomcle- cular hydrolysis; that is, the formula of the instan- taneous speed of reaction does rot include 2 factor linked to the moisture concentration so long as tne latter is close to saturation. However, in practice, if the moisture content is appreciably below the solubility lizit, the syeed of reaction of the hydrolysis is considerably reduced. Practically, when moisture content reaches tne level of 0.1% and if there is no moisture availacle to compensate for the water utilised in the hydrolysis, it may be considered that the acidification process is almost completely stopped. This is due on the one hand to the fact that for moisture. contents around 0.1% the hydrolysis proceeds at a velocity markedly lower than thet prevailing at or close to saturation level and, on the other hand, to the fect that any tendency to hydrolysis entails a depletion of moisture and consequently a slower reaction speed (confer Chapter YI). The problem of oil storage is therefore primarily one of drying and secondly one of preventing the absorp- tion of atmospheric moisture. The study waich has been undertaken consists of the following: 2. Study of the solubility of water in pala ofl. 2. Determination of the vapour pressure of the dissolved water. 3. Drying througs iry air pubbling. 4. Industrial drying of oil or dehydration. 5. Hygrosecpic characteristics of the oil ‘that is the rata cf aoisture absorption in humid surroundings. 6. Speed of ucisture adsorption under industrial conditions, i.e. reaydration.ee een - 18 - These points are examined in detail below:— SOLUBILITY oF OIL . This hes been determined if relation to temperature, free fatty acid content and free hydroxyl radicals. The determination is carriea out by centrifugation of the oil to which has first been added an excess of water. 4 special container is used. The water content of oil is determined after prolonged centrifugation. In fact, the solubility corresponds to the asymptot of the dis solved water contents recorded after multiple centrifugations. HATER The determination of water in oil is carried out using acetyl chloride (33). The solubility curve of a few samples of oil-is given in Figure 37. The solubility increases with increasing ?.F:i content and with higher temperature. Comparative teste on refined oil to which fatty acid has been added and on naturally acidified oil, that is possessing a high hydroxyl value, show that the free OH radicals play a part in the solubility of water in oil but a lesser one than that of temperature or free fatty acids. | The determination of the solubility of water in oil can only be carried out above a temperature of 40°C. In order to determine the solubility of water in oils which remain fluid at 00C, corn oil at 0.28% free fatty acid and soya bean oil at 0.11% F.F.a. were tested. The results which are shown in Table 58 show that the pattern is the same as for palm oil. Table 58 ‘ Solubility of water at various temperature (expressed _in % on ofl | Temperature Oil oe | 0% | 08% Corn oil 0.145 | 0.255 0.375 Soya bean oi] 0.090 0.200 0.290Solubility of water in palm oi} versus: | | ! | Tenporatude gl 0 50 30 70 ED ous results makes it 1; % the solubility of governed oy the following ‘whe examination possible to state +e: water in vegetable ot. equation: Ss = a+B( ¢ where S = the solubility in g * of water. A = a constant specific 2 oil considered. Bs factor of increase for each percentage unit of F.P.a ¢ = the factor of increase for each degree of ‘temperature. FPA = free acidity in g % of palmitic acid. + = temperature in degree cen In the case of palm oil of Congo origin, particularly within the range of temperature 50° te 80°C and the range 0 to 7% P.F.Awy the adove equation is expressed cy the following S = 0.198 + 0.0112 (FFA) + 9.00467 (+-50) The equation enables an as: sment tc te made cf the solubility in routine conditions for a naturally acidified oil. In other words, it provides an imate of the maxi- mum moisture content likely to be found in oil at the outlet of a perfectly set industrial centrifugal purifier in which no dehydration takes place es a result of evapora- tion through contact with air.It is necessary to yo_at out that the two equations are only rough approximations. They are however sufficiently accu yr routine practice prowided they are used witl specified range cf tempera- ture and acidity. particulars, the effect of temperature must be checked through the use of the general law on solubility equilibriv ains | AE “t+ EN where: S = Solubility T = Absolate tenperature A Heat of solubilisation R = Constant of perfect gas By intevration and conversi decima? logarithm thé following is obteiz log. S= -K B+ OC By plotting log $ against the reciprocal of temperature (for two values of 7), a straight line is obtained which mikes it possible to read off solubility at any temperature, provided AE remains fairly constant o ide range of temperature. VAPOUR PRESSU tg oil to improve its The need to dehydra: xeepability makes iz nse Inowledge of the vapour That knowledge permits of industrial dryine. ary to have a detailed essure of dissolved water. establish the feasibility The determination of the wate> vapour. pressure in oil has therefore veen undertaken. The apparatus (Figure 38) consists of a 250 cc round flask equipped with a mercury manometer and Zitted with a lateral tube which is sealed after the flask has been com- pletely evacuated. ‘The flask is then immersed in 2 thermostatically controlled water bath accurate to the 1/l0th of a degree. The changes in mercury in the manometer are ~ecordéd. The moisture conten’ is determined by the acetyl chloride aethed (33) at the end of a series of recordings at various temperatures. The results are corrected for the following: a) The residual air estimgted by a measure- ment carried out at O°° on the assumption that at that temperature the atmosphere is saturated with moisture. (Tn our tests, the correction yas always lower than 55 um of mercury).- 184 - of mercury at various ratures (Tae correction is very f 1 mm of naticn of moisture without correction ae accuracy is of the order of the volumesof the vapour phase 2x id phase ere respectively 50 and 200 co. The eventual correction is smaller than the titretion scouracy. The results content of: oil for the various The results are show: in Table 59 and are plotted in Figures 39 and 40 (vapour pressure against moisture content). solubilities as determined in Sectio. 1, taking the vapour pressure of free water as cor esponding vapour pressure, are also plotted on the % Moisture Water alors 0.16% 0.10% 0.01% The results follow Henry's law: "The amount of a gas of low. solubility which dissolves in a given mass of liquid, at a defined temperature, is directly proportional tc tie partial pressure of the gas" possitie tv establish "a priori” the equilibriun tare content, that is the maximum limit of drying that may be attained witn air at a given temperature ani e known degree of hygrometry. This is because the 2quilivrium aoisture content is equal to the solumlisy multiplied by the hygrometric This makes degree 0” the eir hathemasical expression is as follows it = s*K where Heg® = vu moisture content at t° t se temperature t° hye ee of air at t?- 18¢a = Figure 38 . \ as. Apparatus for the determination of the vapour pressure ¢f water dissolved in400 300 200 100 400 300 200 100 - 184d - Figures 39 £ 40 Water vapour present in om of mercury Measured vapour saunas — 905 ob0 1S COTE | ater vapour pressure in am of mercury ° pe i | | E_(oin at 2.446 PoP A.) 7 i , | i I i A sonppansty | Heagurea vater Prpesure | * 0.05 0.10 oa 5 Rois: r content Vapour pressure of zater dissolved in oil within the concentration range from nil to solubility limitFor technical reasons, no determination of sclubility and vapour prescura was carried out at 1000¢. However, it may be found by extrapolation of. the above results taat at a temperature of 1000C il within the F.?.a. range 1 to 7? reaches equili- Sriun for air with water vapour pressure of about 30 mm of mercury (maxinua in Congo) when the moisture ontent of the oil is within the range 0.017 to 0.020%. The latter must te considered as the limit obtainable for oil at 100°C and air in the Congo. It could be lowered (wnich does not seem to be of practical value) ‘ey increasing the temperature of oil or tarough desiccation of air THE DRYING OF OIL THROUGH DRY AIR BUBBLING The conclusions of Section 2 give the conditions of equilibrium governing the limit of drying cut they provide no indication on the speed with which the jimit is attained or at least approached. They how- ever make it possible to assimilate the problem of Sxchanges water-oil and 2ir to one of exchange between a liquid phase (012) and a low solubility gas (water vapour), the air playing no part in’ the exchange reaction apart as a support and dilutant of the water vapour in the gas phase. A series of experiments were carried out to ascertain the speed of drying of oil under various conditions. The experimental results are given in fable 60. The apparatus used is outlined in Figure 41. In tests 1 to 14 the apparatus was used exactly as shown in the The only variable was the diameter of the + nozzle and its depth in felation to the oil level. In tests 15 to 19, the air stream flowed through a 7 mm diameter nozzle set parallel to the surface of the oil. A 200 rpm mechanical stirrer was used in tests 17 to 19. The stirrer caused a shallow vortex only. All experi- nents were carried out on 400 g of oil, representing a height of approximately 10 om in the round flask. Samples were drawn at periodical intervals for moisture content determination vy the acetyl chloride procedure. The calcium chloride moisture ‘absorbing colum is used as a check on moisture content determinations since it gives an indication of the amount of moisture removed. Wen the experimental results are fitted to a formula of the type of the mass transfer equation, the following applies:- ag = = KC at where C the moisture concentraticn at a given tine. + = the time elapsed since the start of the experiment. K = 8 constant governed by the experi- mental conditions.- 186 - ‘The equation means that at any given time the speed of water removal is proportional to the water concentration, all other conditions being equal’ The constant K is therefore a measure of the suit- ability of the experimental conditions. By inte- graticn end conversion to-decimal logarithm the following exp. 2ssion is obtained: log $2 = kt A plot of the legarithmof moisture content against time yields a straight line the slope of which is K. The law applies with good approximation to the experimental results. ‘The various factors likely to have an effect on the velocity of moisture removal, that is on the value of K in the above equation, are discussed below. A. Temperature Tests 4, 5, 6 and 7 carried out under identical conditions and. the results of which are plotted on Figure 44 lead to the following values of K . at 80°C (test 6) 0.040 at 100°C (tests 4 & 7) 0.062 at 120°C (test 5) 0.079 This corgesponds to a factor of 1.22 + 0.03 for every 10°C. Relative humidity of air The pairs of experiments 7 and 8, 9 and 10, 11 and 12, 13 and 14 show that the velocity of moisture removal is hardly affected at all when the water vapour pressure of air is low: (see Figure 42). C. Diameter of the air bubbles depth of the nozale Tests 7, 9, 11 and 13 show clearly the restricting effect of bubbles of increasing diameter and also of reducing the distance between the tip of the nozzle and the oil surface (Figure 43 refers). D. ° Stirring Finally, *tests 15 “and 16 show that the renewal of the air above the oil surface is not sufficient to ensure drying but that stirring aseociated to air renewal over the oil surface leads to an efficiency of the same order of that obtained by air bubbling (tests 17 and 18).- 187 - It should be noted that the frequency of air changes is not critical since variations in air throughput in the ratio of 1 to 10 litreshave no effect on the value of K. 2.50 [Tog otf 10” x Vaistans cbatent | +326 | | : | | 2.25, 4 1 t 1 180 i | 1 i | 2.00 t | 1 +100 | | r” 1.75 > 056 ! (aie | °. 1.50 | (ate be 120°} | ose | ie at rN | | h——| Lit | | | sco} | Ni 1.00 |-tarreck oF the Feapgratize X t t of the dry 4 air) | t : 0.75 i LN (lime of afying in|minutes)), 5 10 15 20 25 30 35, 2.50 7 Tojatane oamtent | 1 0.316 aaa i Test rd.13) At @ om frog the 0.180 y i | | Test fq. 9) off surtacd. t | | test ro.12) At 10 cm todo the ead Pl Test Fd. 7) off, surfacd. 0.100 i | | 1.75} 1 + 0.056 3.50 + : (0.032 : i Réfect gf the diameter 1.25 }or-tne |dzying air 0.018 bubbled. i : i \ 1 + 200 T ’ Tmo wetnie) j0+010 i i i i | ot are i | [pe of saying tn fimutes) | oO: 5 10 15 20 25 30 35 40 Speed of dehydration of a m 1 samples188 9,001 ‘OgIz 4® Lto"d rE0"0 69t"0]Sot"o auraayae TTA poqeanyes Ste so Soastt ot] 6 rs £00" v20"0 €g0r0| Lot*o Suraatgs WON 9,001 s19 Arp go ort tT] og ss 700" ‘720° €go"o| Lot+o Suyzrt39 HTH 9,001 z¥v Axp 30 soxatt ot | L - zat" cetro ectro] eto Buraat98 ON 9,00T 110 fp yo sea4tT OT] 9 sorsane ay of eer’ Leto Geto] gto] TeTTERrd MOTE ate SuTEITaS On 9,001 24% Axp go soxytT OT D,001 foot z20"o| 6roro 060°o| g2t*o] SLt*o ye poyvanes x18 30 oxftT T orz'o| Ts0°0 160"0) €zt-ofolt*o sorgans om oTeq mo Zu | are Ary Jo oxatt T 9,001 ‘0,12 toro! 9200 060°0| zot*o| P1T"o| MOTT 9 FO worzoq wu] av pesoangee 278 30 ext 1 os oto" yr0°o| 6¥0"0| £60°0| EL1-0 ABOTE OM JO worgoq wm ) 9,001 4¥P AXP Fo OXATT T zzoro zvoro 190"0] ¥2t"0] PLT*O eorgins ona KoTeG BO ZMH T | O,0OT ITE posesravs Jo exATT T ep g00"o 6t0"0 660"0| €or"o|zlt-o eorgans y2 KoTeQ Wo Z wu T 9,00 *¥P Axp Jo matt T 2,001 ‘9,12 £z0"o| 9€0"0 290° 0|OTT*0| PLT*O. AOLTE ous FO woyg0q mu T 40 poqeanges ar@ go esftt t | 29 £00"0 otoro Stor0|060°0|z11*0 ASETE OW FO BOIgOq MU T 9,00 #10 Are go exter r | L ov Sto'o| z90"0 Sotro| pez*o ABETE ANA Jo BorZ0q WM T 9,08 te Arp zo omstt tT | 9 ol 900°9] tt0°o| 9£0"0|S60°0| e¢2"0 HEPLE OM FO woRg0q BU T o,oet are Axp go emt t |S 2°9 |€ooro 100" | 600° Storo agro HECTE om FO m0930q oH T oot sre Arp goomrt tc | F ye |ttoro} © | Lto"o| €¢o"o| £90°0, 9ST" BETZ yy JO wozz0q uM G+ 9,06 x7 Axp go eaatt Sto | € ASETE oR Jo O330q ore |rzoro $60°0| 260°0 servo gLE‘o| pur doy uosmyeq fem sTey um g 9,00t trv Arp go omatt t | z z°€ |rtoro 020"0| Leo"o 6Lo"0 g9t*o HSPTE ONE FO wog0q aE g 9,00 s¥e zp 30 exast T zoty |_ov |_se |_oc |_oz |_ar |_ot |_¢ [Lo ‘eTzz0u agnuyu ced sox9tT UT somite Uy Say em Jo uoryrs0g y uorsuowrg | nots go ener tte pue eanqrredues OF STATetch of the drying apparatus 1. from compressed air reservoir 2. capillary flow-meter 3. silica gel desiccation’ tube’ 4. 600 mi flask 5. thermostatic oil bath 6. calcium chloride desiccation tube. The drying by a stream of air parallel to the oil surface can be treated as a problem of water evaporation at the interface of oil and air. This problem presents two cases:- 1). Oil is not stirred: As a working hypothesis it may be assumed that the water concentration at the interface is renewed only ty the diffusion effect. In that case the speed of the whole process is governed by diffusion. 2) If oil is stirred sufficiently to ensure that the composition of the oil at the surface is con- stdntly identical to that of the bulk of the oil phase, the only factor playing a part in the process is the speed of evaporation, provided the renewal of the gaseous phase is sufficiently fast. In both cases the process is governed by an equation of the type ac = - Ke at . but the nature and the value of the kinetic constant K are likely to differ considerably from those given in Table 60. In practice, drying is often carried out under conditions which are intermediary between cases 1 & 2 above. It is therefore necessary to take into account the frequency of surface renewal, that is the efficiency of agitation.- 190 -" This factor can be measured empirically only and should be determined for each individual experi- mental condition. Summarising, the problem can be regarded as one of mass transfer. The fact that no account is taken of heat transfer does not entail a significant error considering the small amount of moisture to be removed. 4. INDUSTRIAL DRYING OR DEHYDRATION A Dehydration by ventilation Long before research work was undertaken the HoB (Huileries du Congo Belge) had developped an oil arying opparatus which reduced the moisture content to 0.05%. It was observed, in the course of service, that the equipment reduced moisture content to less than 0.1% under normal conditions only if the initial value was not higher than 0.4 to 0.5% Beyond a figure of 500 kg/hour an increase in throughput affects dehydration adversely. Compensa- tion may be achieved to some extent by using a higher temperature. The governing factor in oil drying whether in the laboratory or industrially appears to be the degree of dispersion of/oil, that is the size of the contact surface between air and oil. The degree of dispersion can be increased by allowing the hot oil to drop fron a height over a plate in the drier. Lene Dehydration by means of cascading in thin layers appears less effective than by spraying. Oil flowing on a disc driven by an electric motor and passage of air through the resulting film also provide effective drying. The capacity of the equipment is thus increased to 2 tonnes per hour (Figure 44 refers). Figure 4 (Se sprayi dise i \Confirmation of the excellent results obtained through aeration of an oil film is provided by the fact that moisture content can te reduced to less : li, that is appreciably less than the limit of solubility, simply by centrifugation associated to the aeration of the finely divided oil at the . outlet of the centrifuge bowl. A specific case concerns a large mill in the Congo where drying is carried out solely by natural aeration during centrifugal purification. 4 moisture content of 0.05 to 0.06% is obtained at a throughput of appro- ximately one tonne per hour. At a temperature of 100°C the centrifuge bowl shown in Figure 45 (De Laval VIB 1919) leads to moisture contents lower than 0.1% even when through- put is as high as 2 tonnes per hour. If natural Ventilation is increased through the addition of a suitable fan at the outlet of the bowl, moisture content can be further redyced to avout 0.04% (output of the fan 0.26 m3/sec). Vacuum drying Vacuum drying has been achieved in the laboratory by finely spraying hot oil into a flask evacuated by means of a vacuum pump. . With oil containing 0.65% moisture and heated up 40 10090, it is possible to go as low as 0.07% moisture. Under the same conditions of drying but for an initial value of the moisture content of 0.15%, the final level may be as low as 0.01%. The experimental output was 35 to 48 litres per hour. The laboratory results are therefore excellent. Unfortunately those obtained in an industrial apparatus are less encouraging. Another industriel development operating with a barometric colum has made it possible to dry oil but a serious drawback of the equipment is the carry over of oil into the water of the venturi ejector even when numerous baffle plates are fitted in the oil/water separation drum. Priming and stopping of the equipment and sampling constitute relatively complicated operations. An outline of the apparatus is given below:~- 192 - Figure 4% Centrifuge bowl with outline of air flow xeTeH TKO raieE PO- 193 - A few results are given below:~ Moisture content Throughput Before ‘After tonne/hbur 1.18 0.22 2.3 1.18 0.15 2.3 2.39 1.37 2.3 85 0.26 0.08 34 95 0.26 0.07 34 95 0.32 0.11 9.0 85 0.35 0.17 9.0 The study was undertaken chiefly to find an alternative to drying by ventilation because of the possibility of a deleterious effect of air on oil “in the latter system. It was however observed that aeratign of oil, even for a temperature in excess of 100°C did not affect quality either immediately after drying or after ageing of the oil (oxidation, bleachability, isomerisation). a In conclusion, drying by ventilation possesses a number of advantages over vacuum drying. The chief ones are:~ - low capital & maintenance costs - low space requirement - very low power requirement. Oil oxidation resulting from aeration will be discussed in the next chapter. No isomerisation resulting from storage at high temperature appears to take place. The di and triene content shows no increase as indicated by the resulta given below. (Determination by ultraviolet spectrophotometry) i= Control ., Dey_ O41 Diene content (%) , 0.089 0.062 Triene content(Z) 0.005 0.004. The above oil was dried by bubbling air during 7 minutes in the Swift apparatus.- 194 - HYGROSCOPICITY ~ VELOCITY OP WATER ABSORPTION OF ‘The absorption of moisture by thg dry,cil has, been evaluated in saturated air at 40°, 60° and 60°C using 125 mm thick oil laygrs. Te results show that absorption is faster at 60°C & 80°C. Moreover, it seems that absorption is more uniform at the higher temperatures, presumably as a result of convection currents. At 40°C these currents are not detectable and part of the oil crystallises. At that tempera- ture, the limit of rehydration appears to be lower than'the solubility as determined previously. The rehydration curves of 2 6.5% F.F.A. oil at the 3 temperatures mentioned above are given in Figure 47. hen the experiment is carried out in air much drier than its saturation point, the process of water absorption is considerably slowed down, in particular at 40°C. . The governing factor of the speed of rehydration is the thickness of the layer of oil in contact with air. It can be shown by increasing the thickness of the oil layer that there is an inver'se proportionality between thickness and moisture absorption. For instance the increase in weight can be detected on a laboratory balance for samples of a few millimeters thickness exposed to air for 20 to 30 minutes. If the oil layer thickness is as much as 25 to 30 centimeters, the limit of solubility is reached only after approxi- mately 40 days. With layers several meters thick rehydration may require several months. Table 61 and Figure 48 show the results recorded in a rehydration experiment carried out on aprroximately 50 kg of oil stored in.an open drum and exposed to air in approxi- mately 28°Cand 65 to 95% relative humidity. Table 61 Moisture content Initially 0.040% After 4 days 0.049% After 6 days 0.0538 After ll days 0.0764 After 17 days 0.083% After 27 days 0.095% After 28 days 0.105% | After 33 days 0.120% After 39 days 0.120% | After 44 days 0.115% t295 Figure 47 Bffect of time and temperature of storage on the ration of oil in 2 laboratory experiment o St ot g 9 S$ b € 2 kept ogh of ove 26t OT o2t 96 22 ah _snou fz Oo GT ART. AP ET HATO: FH ATTA CS Taedet om serdube pus da puod pxexey Tyo pus] zo wloagoq Teatyoodspa pexey sordan ‘eam prob out zo Sy o6elo s + jaa fous Gin fe ppauda ai ira [ous | zon oz" of". oro oso.Higure 48 Rehydration of oil. Storage in open 0,10 0.08 0.06 0.04 40 3035 40 5 10 i> 20 25 aays _ These test results were confirmed by the study of rehydration of oil stored in drum,but in indirect contact with the ambient eir through an oil valve. The object of the valve is to stop atmospheric moisture from diffusing towards the air in contact with the oil. The drums were painted black and were exposed to sun and weather. The results obtained with a number of valves are assembled in Table No. 62. A sketch of one of the valves is given in Figure 49. Figure 4 "Yilleboat" Valve- iT - Table 62 Rehydration of oi? in drums fitted with Iicagel | Vetieboat] RCE. [OCCT | Control) "trap ‘valve |valve | valve _| 1 Tnitially | 0.036% | 0.024% | 0.030% |0.040%| 0.0508 | | After 19 days| 0.035% | 0.071 | 0.057% }0.060%) 0.087% | after 35 days] 0.075% | 0.090% | 0.100¢ |o.070¢| 0.082% After 52 days| 0.072% | 0.055% | 0.076% |0.064%| 0.052% | a Accuracy * 0.015 REHYDRATION AT INDUSTRIAL SCALE The previous experiment, although carried out on 250 kg of oil, is not representative of what happens in tanks of several hundred tonnes capacity in which the ratio of surface to volume is very different. A large scale survey was carried out ‘in the installations of the 8.C.C.B. Approximately 140 tonnes of dry oil from the Bosondjo factory were stored in lorry tankers (6 tonnes approximately) for e period of ll days. The initial moisture content was determined. Further tests were carried out periodically. The temperature and the relasive humidity of air were also recorded daily. Thirteen days after the filling of the tankers, the oil was transferred into barges for transport to Leopoldville. An analysis was carried out. Moisture content was eventually determined in Leopoldville, 15 days after the loading. The results are given in Table 63. Table 63 Rehydration in Tanks (cime [Bamber oF Woistare content of |Moisture content (da daya|2O7Y tankers) oi1 in lorry tankers) of oil in the | YS) received and tank analysed 1 2 0.08 - 0.05 | : 2 3 0.10-0.08~0.07 | - 3 3 0.07-0.07-0.05 { - 4 - - |0.05 - 0.07 | 5 4 0.08-3.07=0.05-0.005} - 6 2 0.06 ~ 0.09 | 5 7 2 0.04 - 0.06 | - 8 | 2 0.10 ~ 0.08 0.08 - 0.10 9 | 2 0.05 = 0.07 - 10 1 0.08 | - lu 1 lo.o7 = ; Average 22 0.069 |- 198 - Average moisture content : ‘at the time of production: 0.076% At day 13th: loading into barges. Moisture content in deep tanks: 0.071% It may be concluded that if suiteble precautions are taken, oil rehydration is a very slow process. | 1f abnormal increases in moisture content are observed, they certainly can be ascribed to accidental ingress of water. ‘The reception slip of the oil by Congopalm, mentions the following results recorded at Leopoldville after a 15 days long trip:- Deeptank No.1 0.06% Moisture Deeptank No.2 0.067 Deeptank No.7 0.07% " Deeptank No.8 0.06% d ‘A recora of the observed data is given below: Air maximum temperature in tank 52°C Air minimum temperature in tank 18°C Average temperature in the morning(6-7 am) 20.6° Average temperature at noon (12-7 pm) 45.2° Average temperature in the evening (5.00 Fu)45.0° Average relative humidity of air in tank: Morning 100% oon 358 Svening 36% Over the 15 days of observation, condensation appeared on the tank sides during 2 days. Industrial results Although studies on dehydration and rehydration date back to 1952-1953, several years were required to generalise the drying of oil in the producers’ factories. Gt present, most of the oil delivered to Congopalm is dried in the producing factory (79%). For a long time, @ number of producers who did not succeed in delivering low moisture content oil to Leopoldville incriminated rehydration during storage, transport etc..... Q comprehensive survey showed that sharp rehydration was invariably due to accidental ingress of water (steam coils, pumps...Ean ee eee NanenI ne ncn NTE SeenTTn TEETER Dine - 199 = Using thousands of results, it is possible to summarise the question of rehydration by stating that palm oil is undeniably hygroscopic but that under the present conditions of storage and bulk - transport the water absorption is such a clow process that it should not be perceptible between the area of production in the Congo and the place of utilisation in Europe or in America. The improvement resulting from industrial dehydration can be best illustrated by the record of acidification of oil (between Matadi and the ports of discharge) exported by the cooperative during the last 4 years: Year Increase _in F-F.A. 1952 0.74% 1953 0.59% 1954 0.29% 1955 0.16%