Determination of Tramway Wheel and Rail Profiles To Minimise Derailment

Rail Te~h~~l~~~ l~l~~t
at
Manchest er Met ropolit an Universit y
Det erminat ion of Tramway Wheel and Rail P rofiles
t o Minimise Derailment
Dat e: 12t h February 2008
RTU Ref: 90/3/A
C lient : ORR
Aut hors: Dr P aul Allen
Senior Research Engineer
Tel: 0161 247 6251
E- mail: p. d. allen@mmu. ac. uk
Dr Adam Bevan
Senior Research Engineer
Tel: 0161 247 6514
E- mail: a. bevan@mmu. ac. uk
, ;
oFFacE o~ aa~~ a~cu~arioN
Determination of Tramway Wheel and R ail P rofiles to Minimis e Derailment F inal R eport
P roj ec t Title Determination of Tramway Wheel and R ail P rofiles to
Minimis e Derailment (OR R /CT/338/DTR )
P roj ec t Manager Dr. P aul A llen
Client OR R
Date 12/02/2008
P roj ec t Duration 6 Months
Is s ue 1
Dis trib ution Dudley Hoddinott (OR R )
David Keay (OR R )
P DA /A B/SDI/JMS (R TU)
P roj ec t file
R eport No. 90/3/A
R eviewed b v: P rof. Simon Iwnic k i
Contac t: Dr P aul A llen
Senior R es earc h Engineer
Tel: 0161 247 6251
E- mail: p. d. allen@mmu. ac . uk
s i
!Yw.
2n'. -
^y. . yy. : m'~ ~
4'~: ~~
. !fit' . . ~'
. y, . l. :
CONF IDENTIA L
Summary
As the firs t phas e of a three s tage proj ec t, the O ffic e of R ail R egulation (O R R ) c ommis s ioned
a wide ranging s tudy to review c urrent tramway s ys tems and their wheel and rail profiles
within the UK. Completed by the H ealth and Safety Exec utive (H SE) Labs , the work was
reported under the P has e 1 O R R s tudy doc ument, entitled À s urvey of UK tram and light
railway s ys tems relating to the wheel/ rail interfac e' ~ ' ~ .
P has e 2 of the work, pres ented within this report, analys es this initial s tudy and extends the
work through the applic ation of wheel- rail c ontac t analys is tec hniques and railway vehic le
dynamic s modelling to determine optimis ed wheel and rail profile c ombinations whic h
minimis e derailment ris k and wear.
The key findings of the work are s ummaris ed below:
Three profile ' Sets ' have been identified; for new light rail and tramway s ys tems ; for
exis ting s ys tems with 1 0mm gauge c orner rails and for `tram- train' s c hemes . Thes e
optimis ed profile s ets offer relatively low c ontac t s tres s es and values of Ty (wheel- rail
wear index) when c ompared to many of the profile c ombinations in c urrent UK operation.
The rec ommended profile s ets repres ent fundamentally c ompatible wheel and rail
profiles , however, the c harac teris tic of any one partic ular s ys tem may dic tate detail
c hanges to the s ugges ted profile forms . F or example a s ys tem with a number of mid-
range c urve radii may benefit from an inc reas ed level of c onic ity
New wheel and rail profiles have been rec ommended whic h are geometric ally c ompatible
in the new c ondition, this prevents problems with high initial wear rates of wheels and rails
and will generally provide profiles whic h are more s table in terms of s hape c hange
I t has been s hown that a c onventional wheel profile with a s ignific antly inc reas ed flange
root radius will not provide s uffic ient gravitational s tiffnes s forc e to improve independently
rotating wheel (I R W) wheels et guidanc e
The c onc ept of adopting profiles whic h generate s ingle point flange c ontac t, partic ularly
for I R W equipped non- powered axles , has been s hown to offer benefits in terms of
dec reas ed flange wear rates , thes e benefits als o apply to c onventional axles
The Manc hes ter Metrolink (MML) wheel profile in its new c ondition was s hown to offer a
lower level of protec tion agains t flange c limb relative to the other profiles s tudied.
~ R ~ ~
~ +
' 1 a' i.~ y:.lii ~ ~
CO NF I DENTI AL
However, the MML wheel profile in the new c ondition is not c ons idered as a derailment
ris k and relatively how levels of wheel flange wear has been s hown to s ignific antly improve
the wheel's flange c limb protec tion
I t was demons trated that flange des ign methods s hould inc lude analys is of wheel- rail
c ontac t angle and wheel lift as an indic ator to flange c limb protec tion
With the exc eption of the MML wheel profile, all other wheel profiles s tudied s howed good
derailment protec tion
Wheels et fit analys is has demons trated that c ons ideration needs to be given to the
groove width, es pec ially in s mall radius c urves , if premature k eeper rail wear and
s ubs equent replac ement is to be avoided
V ehic le c onfiguration and highway/ pedes trian s afety s hould be a c ons ideration in the
s elec tion of groove width, that is , a balanc e needs to be ac hieved between s uffic ient wear
allowanc e and c learanc e for the wheels et in tight c urves and maintaining ac c eptable
levels of s urfac e adhes ion/ entrapment hazard for other road/ s urfac e us ers
A light rail and tramway wheel- rail ` B es t P rac tic e' guide has been developed whic h further
expands on the prac tic al is s ues related to the findings of the s tudy and is pres ented in the
bes t prac tic e guide, titled ` A Good P rac tic e Guide for Managing the Wheel- R ail I nterfac e
of Light R ail and Tramway Sys tems ', R TU referenc e 90/ 3/ B ~ 9~ .
A c k nowledgements
The authors would lik e to thank Jim Snowdon (Tramtrac k Croydon Ltd. ) , A ndy Steel
(Hy B ras ail P roj ec ts ) and the proj ec t team (David Keay and Dudley Hoddinott, OR R ) for their
c ontribution to the work ing group meetings held during the proj ec t. The authors would als o
lik e to thank s thos e who c ontributed data to the proj ec t inc luding; Midland Metro and
Manc hes ter Metrolink (Clive P ennington) .
**
_ ~ ,~ . . ; ~
iii ti: ,-; ~ ,f,,
,N ,~ li~
CON F I DEN TI A L
Contents
Summary
ii
1 . Introduction
1
1 . 1 P has e 2 Work Overview 1
1 . 2 Org anis ation of R eport 2
2. R eview of P has e 1 R eport
4
2. 1 Derailment Summary 5
2. 2 Grooved R ail Summary 6
2. 3 Non-g rooved R ail Summary 8
2. 4 Wheel P rofile Summary
1 1
2. 5 V ehicle Summary 1 5
2. 6 R eview of P has e 1 R eport Summary 1 6
3. Wheel- R ail P rofile Analys is 1 8
3. 1 Contact Analys is 1 8
3. 1 . 1 Contact P os ition 1 8
3. 1 . 2 R olling R adius Difference 21
3. 1 . 3 Contact Ang le and Wheel L ift 27
3. 2 Grooved R ail Wheels et F it 28
4. P rofile Selection for Simulation 32
4. 1 Selected Wheel P rofiles 32
42 Selected R ail P rofiles
33
4. 2. 1 Grooved R ail P rofiles 33
4. 2. 2 Non- Grooved R ail P rofiles 34
5. Generic V ehicle Model 35
5. 1 The V AMP IR E P ackag e 35
5. 2 Generic V ehicle Model 35
5. 2. 1 Generic V ehicle Overview 36
5. 2. 2 The V AMP IR E Model 37
5. 2. 3 Motor Bog ie Model 38
5. 2. 4 Trailer Bog ie Model 39
rs +
IV
CONF IDENTIAL
6. Dynamic Simulation Derailment Study 40
6.1 Derailment Study 41
6.2 Derailment Study R es ults 44
6.2.1 Derailment Study R es ults Summary 44
6.2.2 Derailment Study Wheel L ift 49
6.2.3 Derailment Study Contact Angle 53
6.2.4 Derailment Study Angle of Attack 54
6.2.5 Derailment Study Y/Q R atio 55
7. Dynamic Simulation Curving Study 58
7.1 Curving Study Overview 58
7.2 Curving Study R es ults 60
7.2.1 Curving Study Tgamma 60
7.2.2 Curving Study Contact Stres s 64
7.2.3 Curving Study Angie of Attack 70
7.3 Curving Study Summary 71
8. Dynamic Simulation L ateral Stab ility 72
8.1 Conventional Axles 72
8.2 Independently R otating Wheels (IR Ws ) 72
8.2.1 IR W Simulation Overview 73
8.2.2 IR W P rofiles offs et R unning 74
8.2.3 IR W P rofiles F lange Wear 76
9. Wheel- R ail P rofile R ecommendations 7~
9.1 Selection Methodology 78
9.2 Definition of ` P rofile Sets ' 80
9.2.1 P rofile Set 1 (New Sys tem] 80
J .2.2 P rofile Set 2 (E xis ting Sys tem) 81
9.2.3 P rofile Set 3 (Tram-Train) 83
9.2.4 IR W Wheel P rofiles 86
9.2.5 Wheel- R ail P rofile Selection Matrix 86
1 0. Conclus ions 89
R eferences 92
~T~~
:i
~!
CONF IDE NTIAL
Appendix Al Wheel- R ail Analys is P lots
A1.1 Contact P os ition P lots 93
A1.2 R olling R adius Difference P lots 95
A1.3 Contact Ang le and Wheel L ift P lots 97
Appendix A2 Wheel- R ail I nterface Summary Sheets
A2.1 BS 113A R ail Section 102
A2.2 BS SOA R ail Section 105
A2.3 S49 R ail Section 108
A2.4 R i 59 Grooved R ail Sections 111
~~~~
~~
.~.~, ~
CONF I DENTI AL
Determination of Tramway Wheei and R ail P rofil es to Minimis e Derail ment F inal R eport
P age I ntentional l y Bl ank
s s " s
~~ v ii , ~ns ~~,
, , _ ~. . ~.
CONF I DENTI AL
1. Introduction
L ig ht rail and tramway s ys tems operating in the UK have not adopted s tandard wheel or rail
profile des ig ns . In particular the wheel cros s s ectional profile has a very s ig nificant effect on
the running behaviour of the vehicle, the forces g enerated between the wheel and the rail and
the lik elihood of derailment. Unlik e the mainline railway, which is controlled by R ailway Group
Standards , there is no clear g uidance to eng ineers res pons ible for des ig ning and maintaining
the critical wheel- rail interface and as a res ult there is potential for s afety critical problems to
aris e. This is exacerbated by the potential, in the abs ence of s uch controlling s tandards , for
wheels , rails , s witches and other components orig inally des ig ned for heavy rail applications to
be us ed inappropriately in lig ht rail applications . A further is s ue is the tendency for wheel- rail
interface is s ues to be overlook ed at the s pecification and des ig n s tag e as they tend to fall
between the remits of the vehicle builders and infras tructure contractors .
The OR R commis s ioned a s tudy with the primary aim of optimis ing wheel and rail profile
des ig ns , to reduce the ris k of derailments as s ociated with profile incompatibilities . Guidelines
will als o be developed to ens ure g ood wheel- rail interface practice, for both exis ting and new
s ys tem procurement.
The work des cribed in this document reports on P has e 2 of the three- part OR R initiated
s tudy, which aims to addres s the is s ues des cribed above. P has e 1 of the s tudy carried out by
the HSE, was a review of lig ht rail and tramway s ys tems operating in the UK, the rang e of
wheel and rail profiles us ed and als o information on derailments and near mis s es . An OR R
report was produced s ummaris ing the P has e 1 finding s and this document is reviewed as part
of the P has e 2 work des cribed within this document.
P has e 3 which is anticipated to follow the work outlined here, propos es controlled trials of the
optimis ed wheel and rail profiles determined within P has e 2 to verify the finding s of the s tudy.
1.1 P has e 2 Work Overview
The following s ection outlines the orig inal propos ed work s chedule to complete P has e 2 of
this OR R proj ect. The overall aim of the proj ect, in reducing derailment ris k and wheel and rail
profile wear will be prog res s ed throug h completion of the following k ey tas k s :
R eview of exis ting data recorded in OR R report; À Survey of UK Tram and L ig ht
R ailway Sys tems R elating to the Wheel/ rail Interface' . The review will identify the
mos t relevant wheel and rail profile combinations and as s is t in targ eting the
' ~+ `
tiìr.~ het~ .r
~ ~ .~ y~ J ll ~ ~
CONF IDENTIAL
Determination of Tramway Wheei and R aii P rofil es to Minimis e Derail ment F inal R eport
s ubs equent anal ys is and s imul ation work through devel opment of a profil e s imul ation
s et.
Us ing the identified s imul ation s et, as s es s ment wil l be c arried out with res pec t to
derail ment ris k and wear, us ing wheel- rail c ontac t and vehic l e dynamic s imul ations ,
both in- hous e s oftware and the VAMP IR E vehic l e dynamic s s oftware pac kage wil l
be us ed. S imul ations wil l util is e a repres entative, generic vehic l e model .
Us ing the output from the s imul ation work, two or three wheel and rail profil e s ets wil l
be devel oped, thes e wil l be optimis ed for s pec ific s ys tem c harac teris tic s in order to
reduc e the ris k of derail ment and reduc e wear.
It is antic ipated that in addition to appl ic ation to c urrent tramway s ys tems , the profil e
s ets wil l inc l ude profil es whic h wil l be s el ec ted as ` bes t prac tic e' profil es and thes e
c oul d pos s ibl y be rec ommended for us e by promoters in the tender/proc urement
s tages of new s ys tems .
Through c ons ul tation with the OR R and UK tramway operators , one or more of the
profil e c ombinations inc l uded within the newl y devel oped profil e s ets may be s el ec ted
for trial . It is l ikel y that a s el ec tion c riteria wil l be required to optimis e the c os t benefit
of the trial phas e and this wil l be devel oped through the c ons ul tation proc es s ,
together with agreed output meas ures to as s es s the performanc e of the profil es with
res pec t to the initial proj ec t aims .
The produc tion of a ` B es t P rac tic e Guide' for the operation of l ight rail s ys tems with
res pec t to the wheel- rail interfac e for both ex is ting and new s ys tem proc urement
1. 2 Organis ation of R eport
S ec tion 2 pres ents a review of the P has e 1 report outl ining key findings whic h are of partic ul ar
rel evanc e to the obj ec tives of the P has e 2 work, inc l uding a s ummary of the rail s ec tions ,
wheel tyre profil es and vehic l e c onfigurations c urrentl y us ed on UK tramway s ys tems .
S ec tion 3 detail s the prel iminary anal ys is work whic h was undertaken to anal ys e eac h of the
wheel and rail profil e c ombinations . This inc l udes anal ys is of the c ontac t c onditions , in terms
of c ontac t pos ition, rol l ing radius differenc e, c ontac t angl e and wheel l ift. In addition, for the
grooved rail s ec tions , the fit of the wheel s et within the gauge is al s o pres ented, the obj ec tive
being to determine the c l earanc es between the wheel s et and the grooved rail s ec tion in the
new ~ onditiar.
~ ~ ~ ~
:. *
e~ ~ , . ~
- ~ . :~ , < :~ 2 ~ , :~ + , , . s- . ~
~ . . ~ N- . 1~ 1 , .
CON F IDEN TIAL
Section 4 des cribes the profiles which were identified for fu rther analys is work, tog ether with
reas ons for their s election bas ed on the analys is carried ou t in the prev iou s s ections .
Section 5 prov ides a des cription of the VAMP IR E railway v ehicle dynamics s imu lation
s oftware u s ed to as s es s the performance of the s hort-lis ted profiles (s ee Section 4) with
res pect to derailment protection and minimis ation of wheel and rail profile wear. The ' g eneric'
v ehicle model g enerated to repres ent the maj ority of v ehicle config u rations identified in
Section 2 is als o des cribed in this s ection.
Sections 6, 7 and 8 detail the res u lts of the dynamic s imu lation s tu dies u ndertaken u s ing the
` g eneric' v ehicle model and the s hort-lis ted profiles . Section 6 s u mmaries the res u lts from the
derailment s tu dy whereas Section 7 des cribes the res u lts from the cu rv ing analys is . Section 8
pres ents khe res u lts with reference to the lateral s tability of conv entional and IR W ax les .
Section 9 des cribes the methodolog y and definition of the s elected wheel- rail profile s ets .
Section 10 pres ents the conclu s ions from the work. Appendices Al and A2 prov ide a
cflmplete s et of analys is and s imu lation res u lts g enerated du ring the work.
sr
,,4.,`,
CONF IDENTIAL
2. R eview of P has e 1 R eport
The P has e 1 report entitled À s urvey of UK tram and lig ht railway s ys tems relating to the
wheel/ rail interfac e' ~ ' ~ , was authored by Ed H ollis of the H ealth and Safety Laboratory (H SL).
The report details as pec ts of UK tram and lig ht rail s ys tems whic h are relevant to the wheel-
rail interfac e. The report c ontains tables and s ys tem data inc luding drawing s of wheel and rail
profiles whic h have been us ed during the P has e 2 s tudy. The following s ec tion des c ribes the
review of the P has e 1 report, outlining key finding s whic h are of partic ular relevanc e to the
P has e 2 obj ec tives . Detailed rec ommendations bas ed on the finding s of this review are
inc luded in the s imulation work, des c ribed in Sec tions 3 and 4 of the report, and in the `B es t
P rac tic e Guide' pres ented in the addendum to this doc ument [ 9 ]
The N ational Tramway Mus eum, Tyne and Wear Metro and B lac kpool Tramway s ys tem have
only been partially inc luded in the analys is as thes e s ys tems are s een as being too dis tant
from c urrent and likely future direc tion of lig ht rail to warrant s ig nific ant applic ation of the
proj ec t effort. I n the c as e of Tyne and Wear Metro the reas on for ex c lus ion being that it is
es s entially a `heavy rail' wheel-rail interfac e s ys tem.
201 Derailment Summary
Tramway N efinrork
T
N
.~
~
~
O
3
~ i
~
Y
C
~
~ ~ O
~
..
~
Y
~
Q'
L
O
tll
N
X
f6
N ~
w
h
LL
~ S
~
t6
~
J
~
~
~
N
W
~
7
~
~
~
w
O
p. ~ ? y
1_
C
~
~
~
~
t ~
~
e
~
Y
~ l6
Y
L
C N
~ 9
_
~
`, ~
~ S
~
ry)
r
lU6 t~ 6
m F
p
U
O
D
f6
~
~
~
O
Z
L
(n
~
H ~
Diamond 2 1 3 (6)
d Turnout 6 2 1 9 2 3 4 1 28 (52)
' ` Obs truc tion (Groove/ Trk) 1 1 2 4 (7)
d
~ Dis c rete Trac k F ault 2 2 (4)
0
y Damag ed/ Worn Keeper 2 2 4 (7)
v Overs peed 1 1 (2)
F lang e Climb 1 4 1 2 4 12 (22)
s a t~ ao}
~ ~ ` ~ :
s Tr h..~ ya' ...~ y.w 4
r .y: ii~ ~ ~ ~ .
CON F I DEN TI AL
Table 2 . 1 Derailment Summary
With referenc e to Table 2 . 1 above, it c an be s een that j us t over 50% of all derailments ,
reported within the P has e 1 work, c an be attributed to inc idents at turnouts . Detailed
information on thes e derailments was not available in all c as es , however, it was evident that
the maj ority of the turnout derailments were due to detec tion and c los ure failures rather than
wheel- rail interfac e is s ue. Cas es where the wheel- rail interfac e was at fault were g enerally
c aus ed by flang e c limbing or s triking of the s witc h tip.
At 2 2 % of c ited derailments , flang e c limb was the next hig hes t c aus e of c as ualties . Thes e
inc lude only inc idents outs ide turnouts and may be related to a c ombination of poor wheel- rail
interfac e c onditions and c ontributory fac tors s uc h as trac k twis t, poor bog ie s etup, hig h fric tion
c ondition or poor wheel finis h following turning , ag ain, not all details were available.
Damag ed or worn keeper rails attributed to 7% of derailments , as s ys tems bec ome older this
fig ure is likely to inc reas e, unles s s uitable maintenanc e meas ures are put into plac e, s uc h as
keeper renewal, c hang e of rail s ec tion or better c ontrol of g aug e.
Obs truc ted g roove and un- s pec ified is s ues at diamond c ros s ing make up the maj ority of the
remainder, with dis c rete trac k faults and over- s peed being lis ted at 4% and 2 % res pec tively.
I n s ummary, c los ure is s ues mus t be addres s ed and this would likely benefit from improved or
more s uitable s witc h blade ac tuation and detec tion c oupled with, where appropriate, better
protec tion ag ains t wheel flang e c limbing or running behind a partially c los ed s witc h blade.
F lang e c limb derailments (not within turnouts ) s hould in the mos t c as e be preventable throug h
fully c ompatible wheel and rail profiles whic h are des ig ned to operate on the c urvatures
pres ent on the s ys tem, c oupled with g ood fric tion manag ement, turning methods and c orrec t
bog ie s etup (no exc es s ive s tatic wheel load offs ets etc ) .
Keeper wear and ultimate failure is a relatively new problem as s ys tems reac h a c ritic al s tag e
of wear. Sys tem operators are beg inning to unders tand how to maintain and prevent
problems related to exc es s ive keeper rail c ontac t and res ultant wear. R elianc e on keeper rail
c ontac t for c urving c an and has res ulted in derailments . Meas ures to c orrec tly s pec ify new
g rooved rail s ec tions and als o to s afely maintain worn keeper s ec tion is a key part of the
P has e 2 output. I t is als o important to provide unders tanding to operators as to how keeper
rail running c an lead to derailment. There has s ometimes been a mis c onc eption that keeper
rails are c hec k rails and therefore c ontac t with the F lang ebac k is a feature of the des ig n.
~~~
~
fir
~L
CONF I DENTI AL
2.2 Grooved R ail Summary
The P has e 1 report identified that there are many different types of grooved rail profiles
c urrently us ed within the UK light rail and tramway s ys tems . Table 2.2 below s hows a review
of the grooved rail profiles , as identified in the P has e 1 report, us ed by eac h of the UK light
rail and tramway s ys tems .
Tramway Network
~ '~
a~i m
O Y ~ ~
~ ~ N ~
~C O ~ N
Q
N
'C
L L ~ ~ ~ E W a
~ ~ ~ ~
~
` ~
~
E v~
OA
n- 3
~
- o
to
~ c c
m
L
~
"O
~ w
U ~
~
~^ ~ f6
a
~
N rte+
00 V ~ ~ Z Z c A ~
BS 7 . 1
BS 8 1
d R i 59- R 1 0 1
R i 59- R 1 3
~ ~ ~" R i 59N
~ R i 60- R 1 0 . 1
R i 60- R 1 3
o R i 60N
~ ~
c 9 SEI 35G ~~ 4 0~ . . 4
SEI35GP 1
~~~~4 1 GP . 1
Table 2.2 Grooved rail s ummary ( ins talled vertic al unles s otherwis e s tated)
To unders tand the differenc es in the rail s ec tions identified in Table 2.2, the c ros s s ec tion of
eac h of the profiles ( ex c luding BS 7 and BS 8 profiles ) have been aligned and s uperimpos ed,
as illus trated in F igure 2.1 .
It c an be s een that the railhead profile for eac h of the rail s ec tions , ex c ept
for the R i 59- R 1 3( R i 59N) and R i 60- R 1 3( R i 60N) profiles , are identic al with the only
differenc es apparent in the geometry of the groove and keeper s ec tion of the rail. The
R i 59- R 1 3and R i 60- R 1 3profiles differ from the other profiles due to the rail head being
inc lined at 1 :4 0 and an inc reas ed gauge c orner radius of 1 3mm.
~~ + :+ y
Y..; m~.~.r, , ~ 6
, ~ ~, ~
CONF IDENTIAL
Grooved R ail Sections
70
60
50
40
30
20
- R i 5&R 10
- R i 5&R 13 (R i 59N)
- R i 60-R 10
10 -- R i 60-R 13 (R i 60N)
-- SEI 35G
~- SEI 35GP
-- SEI 41GP
0
0 10 20 30 40 50 60 70 80 90 100 110 120
F igure 2. 1 -Comparis on of the cros s s ection of grooved rail profiles
To s ummaris e the differences in the rail s ections the characteris tics of the profiles (groove
width, gauge corner radius etc. ) have been lis ted in Table 2. 3 below.
R ail
Groove Groove
Gauge
corner R ail head
Keeper height
R ail s ection
P rofile
Width* depth
radius inclination
rel. to rail
height (mm)
~mm) (mm)
(mm)
head (mm)
BS 7 28. 58 31. 75 7. 94 1:40 -4. 76 177. 80
BS 8 28. 58 36. 51 7. 94 1:40 -4. 76 177. 80
R i 59-R 10 42. 00 47. 00 10. 00 V ertical -5. 00 180. 00
R i 59-R 13
42. 00 47. 00 13. 00 1:40 -6. 00 180. 00
(R i 59N)
R i 60-R 10 36. 00 47. 00 10. 00 V ertical -5. 00 180. 00
R i 60-R 13
36. 00 47. 00 13. 00 1:40 -6. 00 180. 00
(R i 60N)
SEI35G 36. 00 40. 00 10. 00 V ertical -3. 00 152. 50
SEI35GP 36. 00 45. 90 10. 00 V ertical L evel 152. 50
SEI 41 GP 41. 00 45. 90 10. 00 V ertical L evel 152. 50
Table 2. 3 -Characteris tics of grooved rail s ections
~~f
, ~. , ~~~
CONF IDENTIAL
(`Groove width is meas ured at a dis tanc e of 9mm from the top of the rail)
Exc luding the BS 7 and BS 8 profiles us ed only by the N ational Tramway Mus eum, the
groove width typic ally varies from 36mm to 42mm. The us e of wider grooved profiles , s uc h as
R i 59 and SEI 41GP profiles , allow for inc reas ed s idewear to the running rail before c ontac t
with the keeper c an oc c ur but c an potentially inc reas e the s afety hazard to pedes trians ,
motoris ts and c yc lis ts that travel over them.
Generally grooved rails profiles were des igned with the intention of being ins talled vertic ally.
This is the c as e for the maj ority of UK tramway s ys tem with the exc eption of
Manc hes ter Metrolink, whic h has s ec tions of SEI 35G rail ins talled with an inc lination of
1: 40 , and the R i 59-R 13 (R i 59N ) / R i 60-R 13 (R i60 N ) profiles whic h have been des igned to
inc lude a 1: 40 inc lination in the rail head when the rail is ins talled vertic ally.
Typic ally grooved rail s ec tions inc lude either a 10 mm or 13mm gauge c orner radius . The
maj ority of UK tramway s ys tems us e profiles with a gauge c orner radius of 10 mm with the
exc eption Croydon Tramiink whic h us es R i 59-R 13 !R i 60-R 13 rail s ec tions employing a
radius of 13mm. The work des c ribed later in the report will look at the c ompatibility of profiles
with different wheel flange root and rail gauge c orner radii.
As identified in Table 2.3 and illus trated in F igure 2.1, the grooved rail s ec tions inc lude
different levels of keeper height. In s ome ins tanc es , s uc h as the SEI GP profiles , the keeper
rail is level with the running s urfac e of the rail whereas in others it is s lightly lower. Depending
on the level of wear on the railhead the keeper may eventually bec ome higher than the
railhead res ulting in a potential s afety hazard, The R i 59-R 13 / R i 60-R 13 profiles offer the
greates t c learanc e between the top of the keeper and the railhead (--6mm) due to the
rail head inc lination.
The grooved rail s ec tions identified above, along with a number of additional rail s ec tions not
c urrently us ed in UK light rail and tramway s ys tems , are defined in BS EN 14811-1 ~ 2~ .
2.3 N on-Grooved R ail Summary
Table 2.4 inc ludes a s ummary of the non-grooved rail s ec tions c urrently us ed within the
UK light rail and tramway s ys tems , as identified from the P has e 1 report.
It c an be s een that a large variation in profiles exis t, with Manc hes ter Metrolink us ing five
different types of non-grooved rail s ec tions , although s ome of thes e are ins talled over s hort
s ec tions of trac k or may have been c ompletely removed during rec ent rail renewal programs .
CON F IDEN TIAL
The mos t freq u ently u s ed non-grooved rail s ec tions inc lu de BS 80A, with b oth a 1:20 and
1:40 rail inc lination, and BR 113A s ec tions . BR 113A rail s ec tion appears freq u ently du e to
light rail and tramway s ys tems u s ing (or s haring) old Britis h R ail lines .
Tramway Network
w
c
0
m
3 Y
~ ~~ ~ vim, E o
a~
~
~
o ~ m
L L ~ ~ ~ w n
~ ~ ` ~ ~ ~ m
N
a 3 ~ m ~ = rn m ots ~
m m L o ~ a
c
o ~ c ~
m U D ~ ~ Z c ~ ~, o
BS 95R BH . 2
d
o BR 1091b ~ 1
a
~ BS 80A

~i:ao) (i:ao) (~:ao~ 5
~a
> BS 110A
o (BS 113A) ~ ~ ~ 3
t9
~
0
BR 113A . 6
z
S49 1
Tab le 2. 4 Non-grooved rail s u mmary (ins talled at 1:20 u nles s otherwis e s tated)
To s u mmaris e the differenc es in eac h of the non-grooved rail s ec tions the c ros s s ec tion of
eac h of the profiles have b een aligned and s u perimpos ed (ex c lu ding the BS 95R BH profile),
as illu s trated in F igu re 2. 2, and the defining c harac teris tic s have b een tab u lated in Tab le 2. 5
b elow.
~~~ ~
is s
. . ,~' .
CONF IDENTIAL
Non-grooved R afl Sections
50
45 _ __
35
25
~ {
I
1
1 5
1 0
BR 1 09
BS BOA
5 BS 1 1 3A
BR 1 1 3A
0
0 1 0 20 30 40 50 80 70 80
F igu re 2.2 Comparis on of the cros s s ection of non-grooved rail profiles ( vertical)
Crown
Gau ge
R ail s ection
R ail P rofile
radiu s ( mm)
corner
height ( mm)
radiu s ( mm)
BS 95R BH 304.80 1 2.70 1 45.26
BR 1 091 b 228.60 1 2.70 1 58.75
BS 80A 304.80 1 1 .1 1 1 33.35
BS 1 1 0A*
304.80 1 2.70 1 58.75
( BS1 1 3A)
BR 1 1 3A` 304.80 1 2.70 1 58.75
S49 300.00 1 3.00 1 43.00
Tab le 2.5 Characteris tics of non-grooved rail s ections
(*Note different des ignation for 1 1 3A rail s ections )
I t can b e s een that the maj ority of the non-grooved rail s ections inclu de a s imilar gau ge corner
radiu s , 1 2.7mm 1 3mm, with the exception of the BS 80A rail s ection which has a gau ge
corner radiu s of 1 1 .1 1 mm. The compatib ility of the gau ge corner radiu s with the wheel flange
root radiu s has b een inves tigated du ring the s tage 2 work and is des crib ed later in the report.
The rail crown radiu s of each of the non-grooved profiles varies from 228.60mm for the older
1 091 b rail to 304.80nm for the newer 80A and 1 1 3 pail s eciions . ideally this ~i-~ou lci b e
cons is tent with that of the grooved rail s ection in order to prevent trans ition prob lems s u ch as
~~~ 1 0
t ~W.I ~~
CONF I DENTI AL
u ndes irab le wear patterns in the rails , increas ed nois e lev els and general v ariation in wheel
rail contact conditions .
The effect of different rail inclination, as s een with BS 80A rail, has b een cons idered du ring
this work to inv es tigate the potential b enefits for a giv en wheel profile.
The different des ignations of 113A rail s ection s pecified in Tab les 2. 4 and 2. 5 hav e b een u s ed
to differentiate b etween the differences apparent in the old and new 113A rail s ections . The
BR 113A profile is b as ed on the original Britis h R ail des ign whils t the BS 113A profile is as
s pecified in Britis h Standard (BS 11) . The only difference b etween the two profiles is the
crown head width which is 12mm on the BR v ers ion and 19. 05mm on the BS v ers ion.
The non-groov ed rail s ections identified ab ov e, along with a nu mb er of additional rail s ections
not cu rrently u s ed in UK light rail and tramway s ys tems , are defined in part 1 and 4 of
BS EN 13674
~ 3, a~
2. 4 Wheel Tyre P rofile Su mmary
Us ing the information inclu ded the P has e 1 report, and the digital repres entations of the
wheel tyre profiles generated du ring the proj ect, a comparis on of characteris tics of each of the
UK light rail and tramway s ys tem wheel profiles has b een condu cted. Tab le 2. 6 b elow
s u mmaries the characteris tics of the different profiles ; inclu ding flange depth, flange
thicknes s , flange root radiu s and wheels et b ack-to-b ack dimens ion.
+s ~
_ , . , ~ .
CONF IDENTIAL
T
i4
N
C
~
3 ~ ~
.~
~
~
O
N
~
~
(L
O
C C
~ ~ p
Q
X
W
~
N
~
N
~
N
H
N
~~
J
t/ ~ , , ~.,
~
~
~
~
(6
c c Q- ~
m
a` ~i
t
a
c
~
~
a
m x s
- o
L
~ o
L ~
~C V
~
f0
~
- .~
U UQ-
O
D ~ ~
O
Z
_C
c~
~,
R eference name CR 4000 CT3 DR L5 MML-2
T69
(R ev-A)
P-3-
102639
Mod.
DIN25
P $
F lange
23.2 22.4 27.5 22.3" 23 22.81 23 29
thicknes s
(232) (22.4) (27) (22)* (23.02) (22.81) (23.02) (28.5)
Sd mm
F lange height 25.5 27 30 26.3 24 24 24 30
Sh (mm) (25.5) (27.0) (29.5) (25.9) (24.05) (24.0) (24.05) (29.3)
F lange gradient 4.9 6.43 11 8.5 3.0 3.4 3.0 12.5
qR (mm) (4.9) (6.43) (10.5) (8.3) (3.02) (3.4) (3.02) (12.1)
F lange angle
70 70 69.5 68 76 76 76 68
(deg)
F lange root
15 15 22 13 13 14 13 13
radius (mm)
Cone angle ~.20 1:26 to 1:19.0
1:20 1:40
1:40 to
1:40 n/ a
(grad) 1:10 8 1:20
F langeback dim.
1380 1380 1362 1362 1379 1380 1379 1362
(mm)
Wheel
New 630 630 740 740 680 660 670 740
diameter
(mm)
Worn 550 550 660 680 620 580 588 675
Taperline, L2
60 60 68 70 58 60 58 70
(mm)
Table 2.6 -Wheel profile s ummary (true flange thicknes s )
The flange height and thicknes s meas urements included in Table 2.6 have been calculated
with reference to the recommended taperline for each profile as pres ented in the table. F or
comparis on of the different profiles flange height and thicknes s meas urements have als o
been calculated at a nominal taperline of 60mm, thes e res ults are pres ented in parenthes es
in Table 2.6.
The current tyre profiles operated by Croydon Tramlink, Midland Metro, Nottingham Ex pres s
Trans it and Sheffield Supertram can generally be clas s ified as modified DIN-type wheel
profiles , with the tyre profiles us ed by Midland Metro and Sheffield Supertram being very
~~k~
s s
~, ...:r._s .~, :~
12 i, a..: F , n, :.~
CONF IDENTIAL
s imilar. Nottingham Expres s Trans it profile differs s lightly from the others as it inc lu des a
s ec ond angle of 1:20 on the ou ter edge of the tyre. Large flange angles in the range
70 76 c an be s een on the modified DIN-type profiles . The flange root radiu s of thes e
profiles v aries from 13mm to 15mm. A ll of thes e profiles are equ ipped with flat flange tips for
flange ru nning.
r~
~ _. _
--
_
_. ~ _ _-
_
__
- _ . . . ~, -
l:y5
F igu re 2. 3 DIN-type wheel tyre profiles , Croydon CR 4000 ( blu e) , Nottingham P -310239
( green) , Midland T69 R ev . A and S heffield Mod-DIN25 ( red)
The profile operated by Manc hes ter Metrolink differs from thos e operated by other
c onv entional tramway s ys tems as it inc lu des a fu ll flange geometry and ac u t-ou t in the
flangebac k of the profile. This c u t-ou t was introdu c ed to ac c ommodate the differenc e in
flangebac k s pac ing requ ired for both light and heav y rail trac k work s . This methodology has
been u s ed in other Eu ropean railways to allow light rail v ehic les to operate on other parts of
the railway network ( du al-mode light rail v ehic les ) and therefore has potential for UK
tram-train applic ations .
ma
___
~::
m. 0 ~j
i
~S
i
IF .
1aq
{
' .
1
au
, e
F igu re 2. 4 Manc hes ter MML2 wheel tyre profiles
The Tyne and Wear Metro and Doc k lands Light R ailway u s e mainline type tyre profiles with
the Tyne and Wear Metro u s ing a typic al B ritis h R ail ` P 8 '-tyre profile. Doc k lands DLR S profile
is a modified ` heav y rail' profile whic h inc lu des a large flange root radiu s to generate a
CONF IDENTIA L
Determination of Tramway Wheel and R aii P rofiles to Minimis e Derailment F inal R eport
s ig nific ant roiling radiu s differenc e, inc reas ing s teering performanc e arou nd tig ht c u rves . The
Tyne and Wear BR-P 8 and Doc klands DLR S tyre profiles are illu s trated in F ig u re 2.5 below.
, ~ ~ . _ _
, :
~ m
4
.. ..
L, z
a,
F ig u re 2.5 Tyne and Wear BR-P 8 (blu e) and Doc klands DLR 5 (red) wheel tyre profiles
The Croydon CT3 tyre profile is a prototype profile des ig ned to produ c e better wheel-rail
interfac e performanc e on the Croydon Tramlink network. The profile g eometry is s imilar to
that u s ed by KVB in Colog ne on their fleet of K-4000 vehic les , from whic h the CR-4000
vehic les operated by Croydon Tramlink are bas ed, bu t inc lu des a rou nder flang e tip and a
g au g e c orner radiu s of 15mm ins tead of 13mm on the KVB profile. G enerally, the CT3 tyre
profile has a ` worn' profiles s hape when c ompared to the c u rrent CR 4000 profile, with a
nu mber of blended radii. A c omparis on of the c u rrent Croydon CR 4000 and prototype CT3
tyre profiles c an be s een in F ig u re 2.6 .
n ..., .. .
_ .
~~~~~
a~ ~ f ~ ~ ~ ~
s .,
~ _ _
a_ ., . ._
------
_ _ _ _ _ _
_ ..._ . ..
_ _ _
_
.-
_
~ 1
._ ^ , a ~ u
~ . wdll ~ wir : ~ : i , ~ tea ids o
x ~ : .
F ig u re 2.6 Croydon CR 4000 (blu e) and ` prototype' CT3 (red) wheel tyre profiles
A variation in wheel diameter is als o obs erved in the P has e 1 report with valu es rang ing from
6 30mm to 740mm when new, as s hown in Table 2.6 .
CONF IDENTIAL
Determination of Tramway Wheel and R ail P rofiles to Minimis e Derailment F inai R eport
2.5 Vehicle Summary
Table 2.7 below illus trates the v ariety of v ehicle config urations currently operating in the UK.
I n this cas e, the N ational Tramway Mus eum has been neg lected as the v ehicles are
non-conv entional in the modern s ens e.
Tramway N etwork
.~
0
(p
3
i
~
3
~~
Y
~ v im, E o
~ ~ ~ o ~ m
~ E ' ~ w a
~
m
~ H
u'
~ ~
m
~
a 3 ~ m t c ~ ~a~i
~
otS a
m m o
~ 'o o ~
c
t
m
o
~ ^~ Z~
T C
c
0
I R W on trailer only ~ 2
L P owered I R W
1
f A ll Conv entional . 5
c
0
v Stewing ring ~ 5
Table 2.7 Vehicle Config urations
P res ently, the mos t common v ehicle config uration is a conv entional ax le s etup on all bog ies .
Generally a s tewing ring is us ed to allow almos t friction free rotation of the bog ie. I n the cas e
of the Croydon Bombardier CR 4000 tram, the arrang ement is not s trictly a s tewing ring and
has the addition of friction s idebearers to increas e bog ie rotational res is tance.
I t is relativ ely common practice to combine conv entional powered ax led bog ies with
independently rotating wheel (I R W) equipped trailer 'truck' s ections at articulations . This
config uration allows for a low floor s ection with ramps or s teps to allow acces s to the body
s ections abov e the ax les , which prov ides acces s ibility adv antag es but can res ult in
trailer-truck wheel wear is s ues and increas ed derailment hazard due to the nature of
operation of I R W's .
The only v ehicle operating powered I R W ax les is the N ET tram. The behav iour of powered
I R W's in terms of curv ing/wear and derailment is complex as they are neither pure I R W's nor
purely conv entional but us ually a mechatronic combination of both s ys tems . The effect of this
on the behav iour of the v ehicles was briefly inv es tig ated during this work, in collaboration with
the v ehicle manufacturers .
~~~~
~
.,. ~,..-~..d, ....
15 t~,~,: ~~s ....
~y I ~~
...~ -r'.1' .
CON F I DEN TI A L
F rom initial c ons u ltations with the v ehic le manu fac tu rers it appears that the c u rrent c ons ens u s
for fu tu re tram dev elopment is a mov e bac k towards c onv entional ax les , u s ing ramped floor
s ec tions to prov ide the req u ired ac c es s ibility. This s eems a s ens ible option as it redu c es the
potential problems whic h are as s oc iated with the lac k of g u idanc e of trailer IR W's and the
u n-c ertainty of the ` real-world' performanc e of powered IR W's . A pos s ible alternativ e direc tion
is the adoption of s teered IR W bog ies , bu t this is beyond the s c ope of the work.
2. 6 R ev iew of P has e 1 R eport Su mmary
The rev iew of the P has e 1 report has res u lted in a nu mber of obs erv ations relev ant to the
obj ec tiv es of the P has e 2 work. E s s entially the report forms a data library of nine UK tram and
lig ht rail s ys tems , with the addition of s u mmary tables and derailment his tories .
2. 6. 1 Derailments
50% of ail derailments , reported within the P has e 1 work, c an be attribu ted to inc idents at
tu rnou ts , the maj ority of thes e derailments were du e to detec tion and c los u re failu res
rather than wheel-rail interfac e is s u e.
22% of c ited derailments were c au s ed by flang e c limb, thes e may be related to a
c ombination of poor wheel rail interfac e c onditions and c ontribu tory fac tors s u c h as trac k
twis t, poor bog ie s etu p, hig h fric tion c ondition or poor wheel finis h following tu rning .
7 / a of derailments were attribu ted to keeper wear. 4% and 2% res pec tiv ely were c au s ed
by obs tru c ted g roov e and u n-s pec ified c au s es at diamond c ros s ing s .
2. 62 Groov ed R ails
Ail g roov ed rail, head profiles are identic al with the ex c eption of the R i 59-R 13 and
R i 60-R 13 profiles ; the rail head being inc lined at 1:40 with an inc reas ed g au g e c orner
radiu s of 13mm.
Groov e width typic ally v aries from 36mm to 42mm. Althou g h wider g roov ed profiles do
ex is t, s u c h as 67-R 1 (P h37a), as defined in BS E N 14811-1 ~ Z ~ .
R i 59-R 13 / R i 60-R 13 profiles offer the g reates t c learanc e between the top of the keeper
and the railhead (--6mm).
2. 6. 3 Non-g roov ed R ails
The mos t freq u ently u s ed non-g roov ed rail s ec tions inc lu de BS 80A, with both a 1:20 and
1:40 rail inc lination, and BR 113A s ec tions .
Non-g roov ed rail s ec tions hav e a s imilar g au g e c orner radiu s , 12. 7mm 13mm, with the
ex c eption of BS 80A rail whic h u s es a radiu s of 11. 11 mm.
i4s
e3, ~ , t. -~ . . ~ g _
~ ~ ry~ ~ i~ i~ . .
CONF IDE NTIAL
2 . 6 . 4 Wheel Tyre P rofiles
Wheel profiles operated by Croydon Tramlink, Midland Metro, Nottingham Expres s
Trans it and S heffield S upertram can generally be clas s ified as modified DIN- type wheel
profiles . Characteris tics of thes e profiles include;
o Large flange angles in the range 70 76 can be s een on the modified DIN- type
profiles
o The flange root radius of thes e profiles v aries from 13mm to 15mm
o A ll of thes e profiles are equipped with flat flange tips for flange running
o The tyre profiles of Midland Metro and S heffield S upertram are v ery s imilar
The profile operated by Manches ter Metrolink includes a full flange geometry and a
cut- out in the flangeback of the profile to accommodate the difference in flangeback
s pacing required for both light and heav y rail track works .
Docklands DLR 5 profile includes a large flange root radius to generate s ignificant rolling
radius difference, increas ing s teering performance in tight curv es .
2 . 6 . 5 Vehicle Configuration
P res ently, the mos t common v ehicle configuration is a conv entional axle s etup on all
bogies .
It is relativ ely common practice to combine conv entional powered axled bogies with
independently rotating wheel (IR W), but this can res ult in trailer- truck wheel wear and
increas ed derailment hazard.
The only v ehicle operating powered IR W axles is the NET tram.
~~
~
,. ~~. . . ,. .
~. _ - - 17
,,. . ~y,,,
,~ ~~~ ,.
CONF IDENTIA L
3. Wheel- R ail P rofile Analys is
Us ing the data generated from the review and analys is of the P has e 1 report, wheel and rail
profile combinations were s elected for fu rther analys is . I n general all profile combinations
were cons idered, with the ex ception of; repeated combinations , combinations where there is
ins ignificant adoption of a particu lar profile or the profile combination has been deemed to be
u nlik ely to contribu te s ignificantly to the later work .
The following s ection details the preliminary analys is work which was carried ou t on the wheel
and rail profile combinations . I n overview this inclu ded analys is of the contact conditions , in
terms of contact pos ition, rolling radiu s difference, contact angle and wheel lift. I n addition, for
the grooved rail s ections , the fit of the wheels et within the gau ge was analys ed, the obj ective
being to determine the clearances between the wheels et and the grooved rail s ection in the
new condition.
3.1 Contact Analys is
The contact analys is work formed the firs t s tep in analys ing the contact conditions between
new wheel and rail profiles for each of the tramway s ys tems . Us ing a combination of the
VAMP I R E s oftware contact pre- proces s or and in- hou s e developed MATLAB~ s oftware
rou tines , contact pos ition, rolling radiu s difference, contact angle and wheel lift were
analys ed. Thes e parameters were chos en as together they des cribe the commonality of the
profiles ( lik ely initial wear), their ability to s teer the vehicle and als o res is t derailment.
Du e to the nu mber of profile combinations , wheel- rail contact analys is has been condu cted for
each of the combinations identified in Table 2.2 and 2.4 with the aim of redu cing the nu mber
of s imu lations req u ired du ring the dynamic analys is , as des cribed in S ection 6, 7 and 8.
The res u lts pres ented in this s ection have been s elected to demons trate the more ex treme
variations in the pos s ible profile combinations ; a fu ll s et of contact analys is data is inclu ded in
Appendix A1 .
3.1 .1 Contact P os ition
Us ing the ou tpu t from the VAMP I R E contact pre-proces s or an in- hou s e developed
MATLAB s oftware rou tine was u s ed to indicate where contact was occu rring between the
wheel and rail fora 1 0 mm lateral s hift. With reference to the contact pos ition plots s hown in
F igu res 3.1 to 3.4, the blu e lines , drawn between the wheel and rail profiles , indicate the
points that are in conta~; ~n the wheel and rai! for a liven lateral dis placement of the ~vheels et
p +r
CONF I DENTI AL
with res pect to the track . The red line indicates the pos ition of the contact patch when the
wheels et is in its nominal mid- pos ition.
F igure 3 . 1 below s hows the contact pos itions for Croydon Tramlink CR 4000 wheel tyre
profiles when combined with R i 59- R 1 3 and BS 1 1 3 A rail s ections . This combination res ults in
non-conformal 2- point contact, with no contact pres ent between the gauge corner and rail
head. The nominal running pos ition (zero lateral dis placement) is s ituated on the crown of the
rail.
Non-conformal initial contact conditions in the new condition can caus e high contact s tres s es ,
res ulting in high initial wheel and rail profile wear, ideally the new wheel and rail profile s hould
have a more even dis tribution of contacts acros s the wheel and rail.
Croydon CR IOOD WMN P raR l~ and R i 69N R ail P rofile Creyden CR I000 Wh~~l P roAl~ and BS 1 1 3 A Wii P refii~
" HO
i i i i i i i
i i i i i i i
_ _ - I- - y _ _
_ ~. _ _
_ r _ _
_~_ - - 7 _ _ - T
i
i i i i
_ _ _i_ _ J _ _ 1 _ _ _ _ _ J _ _ _ 1 _
1 1
i ~ i i i i
i i i i i
- - - i - ~- r - - - -i- - - ~ - - - r - -
~ ~ ~ ~ ~
_ _ _ _ _ ~ _ _ _ _ ~ _ _ _ a _
i i0 i i i
i i ~ i i i i
i i i i i i i
i i i i i i i
i i i i ~ i
- i - - ~ - - - r - - - r - - - r - - 7 - - - r
i i i i i i
i i i i i i
ana inn ran yen ~F n inn xnn
~~
60
50
as
20
1 0
~ i i i i i
_ _ _ 1 _ _ _ J _ _ _ _ L _ _ _ 1 _ _ _ J _ _ _ _I_ _ _
1
1 I i
_ _ _ t _ _ _ , _ _ _ _ r _ _ _ r _ _ _ ~ _ _ _ _ i_ _
1 1
__ _i'______ ' _____I___
i i i i i
~- - + a- - - - ~- - - - - ~- - - - ~- - -
- - - - z- - - - r- - 0- r- - - ~ - - - i- - -
i i ~
___1 ~ 1 J ~___ ___ I____- ___ I ___ _ __'
i i i i i i
. - _ _ { _ _ - _ _ - _i_ _ _ _ } _ _ _ y _ _ _ _~_ _
~ i i i i ~
i i i i i ~
~ i i i i
i i i i i ~
_ ~
C X680 700 72a /40 760 780 B00 6Zv
F igure 3 . 1 Wheel- rail contact pos ition, Croydon Tramlink CR 4000 wheel and R i 59- R 1 3
(left) and BS 1 1 3 A (right) rail
The prototype CT3 tyre profile generates s imilar contacts to the CR 4000 profile on BS 1 1 3 A
rail s ection but more conformal s ingle- point contact on R i 59- R 1 3 , as s hown in F igure 3 . 2.
, . r ~ ~ ~ ~ ~
CONF IDENTIAL
s ir
. , ~ .
Cteyden CTS W h~s l P rofits and R i S9N R ail P rofile
i i ~ i i i i
i i i i i i i
___i___~___r___r___i___~___r __
i i i i i
i i i i
_I_ _J ''i __ ~' __J ___1_
i
''. ~ti 1 I 1
_I `___1
--I -I - I---I ---t ---
1
1
i ~ ~ o ~ ~ ~
--- ---------- --------- -
~ ~ ~ ~ ~ ~
___~ __~___r_'_r___~___~___r___
i i i i i ~
i i i i i i i
'. A fiAl1 ]AA ]7A ~df1 ]fill ]0 . ( I R f1( 1 A;
Croydon CT3 Wfi~s l P refila and BS 113A R ail P rofits
1__ J- L---1___J----I 7~ ___ - --- I-__-
i i i i i i
i i i i i i
60
_ _7 ___~ ____r ___r ___~ ____~ _ __
I
r~ _~ --__~_-_
I 1
1
1
~ - _i __ ~ ____~__ _-_~ ____1_--_
1
~ t ~ --r --0 . r ---~ ---r ---
~ ~ ~ ~ i
z o
--=---'----'---'----' '-
__ ___
~o
---+--~----~--*----~----~----~
o ~ ~ ~
_ ______________
i i i i i i
70
F AO 7tl~ 720 740 760 7B0 80 0 8:
F igure 3. 2 Wheel-rail c ontac t pos ition, Croydon Tramlink CT3 wheel and R i 59-R 13 ( left)
and BS 113A ( right) rail
In c ontras t, greater c onformality c an be s een in the c ontac t pos ition on the
Nottingham Expres s Trans it, partic ularly in flange root of the wheel and gauge c orner of the
rail, as s hown in F igure 3. 3. It is als o worth noting that the nominal running pos ition of the
wheel is s hifted towards the gauge c orner of the rail.
NET Wh~~i P rofile and SEI 41GP R ail P rofits
6
i i i i
i i i i i
________'__ ~} . I t. 1 ___
! ~
i i i i i
i 3 i i
20 ---~ - --~ -- -~----~----
i i ~ i i
i i i
~ ~ i i
f O '__I_
~ i i i i
i i i i i
-pp ___ ___J___ ~____1___J____~'_'_
~ i i ~ i
i i i i i
1 I I I 1 1
y~ _-_1_ __I__ _1____1___J____I____
1 1
i
~Q
650 680 70 0 720 740 760 780 60
YCT W1. . . 1P mF I~ a. M 0 . S MA /1inN1R ail P roR la
F igure 3. 3 Wheel-rail c ontac t pos ition, Nottingham Expres s Trans it P -310 2639 wheel and
SEI 41GP ( left) and BS 80 A ( right) rail
The Manc hes ter Metrolink MML2 tyre profile produc es c onformal c ontac t on the grooved rail
s ec tion with les s c ontac t with keeper due to the c ut-out on the flange-bac k of the wheel.
When c ombined with the BS 80 A rail s ec tion a large jump in the c ontac t pos ition c an be s een,
as s hown in F igure 3. 4, this c an res ult in high c ontac t s tres s es and ins tantaneous peaks in
wheels et forc es , thes e effec ts have been attributed to rolling c ontac t fatigue ( R CF ) in heavy
rail s ys tems . Generally, R CF does not oc c ur in light rail s ys tems as normal and tangential
s tres s es are too low.
+s
4~~~. ; hs s ~*~
. v. t:r4~ l~t
CONF IDENTIAL
Determination of Tramway Wheel and R ail P rofiles to Minimis e Derailment
F inal R eport
Maneh~aMr Yatralint YYh~s l Amll~ antl R iSB-R 10 R ail P rofile
i i i i i i i
i ~ i i i i ~
--~--- r---i--- -r---r-- ~---r---
i i i i i
i i i i i
_ _ ~ _ _ F _ _~_ _ _ ~ _ - _ F _ _
i i i i i
i i ~ i i
_ ' J ' _ _ _I_ _ _ 1 ' ' _ L _ _ _
~ i i i i
i i i i i
__i__ ~ ~ ~
I I i
__ ___r__~___1___~___~___r___
i i i i ~ i
t i i i i
__~ _~___}___I
__~___F ___
1 __ I I I 1
i i i i i i
raa ~m n yen ~F n inn nnn a
Manchags r Meholink Whaal P rofile antl BS BOA R ail P rofile
i i i
s o
- - +--- -+----
i
i i
- ---~----
i i i
40
_ _1_'
i
___~_ _1___J___'
i i
3p __ T_
1
___I_
1 1
__T___,____
20
___~
~
i
___i_
~
i i
___~__'-i_"'
i i
f0
___1__
i
~ ___~_
i
L"_1"_J___'
i i i
i
0 _ T _
_
i
_
i i i
T _ _ , _ _ _ _
~ I 1
10
___~___ ____i____~___
i i i
___a___'
i i i
20
_ _ _ ~ _ _ ~ ~_ _
1 I 1

1 _ 1
1 I i
3U
___r___ ____i____r___7___~__'_
i i i i i ~
4 F Aff inn ~~n ~~n ~s n ~xn eno r
F igure 3.4 Wheel-rail contact pos ition, Manches ter Metrolink MMLZ wheel and R i 59-R 10
( left) and BS 80A ( right) rail
As part of the P has e 2 s imulation work, contact pos ition plots as pres ented above, have been
us ed to analys e the propos ed new wheel and rail profile combinations to as s es s conformality
and pos ition of the contact on the rail head and the point at which flange contact occurs .
3.1.2 R olling R adius Difference
When a wheels et is running with s ymmetrical wheels on identical rails and the wheels et is
centred within the track gauge, then the rolling radius of the left and right wheels is eq ual. Due
to the conical form of atypical wheel profile, as the wheels et moves from the central pos ition
a rolling radius difference ( R R D) is generated, the rolling radius of one wheel increas es whils t
the other decreas es . An example plot of rolling radius difference fora 6mm s hift to both the
left and right ( pos itive to the right) is s hown in F igure 3.5 below.
The R R D between the left and right wheels for a given lateral s hift of the wheels et can be
us ed as an indicator of conicity. In general terms , the greater the s lope of the curve, the
greater the effective conicity. Increas ed conicity enables the wheels et to negotiate s maller
radius curves without flange contact. The des ign conflict in terms of conicity is that high
conicity can lead to lateral ins tability of the wheels et, known as 'hunting'. Us ually this is not a
common problem in light rail s ys tems , as maximum lines peeds tend to be relatively low
(=50mph maximum)
The R R D has been calculated for each wheel-rail combination identified in Tables 22 and 2.4
in Section 2. P rofile combinations which demons trate the greates t variation in characteris tics
are pres ented below. The complete s et of curves is included in Appendix A1.
F igure 3.5 below s hows the R R D graph for the Croydon Tramlink CR 4000 wheel profile and
the rail s ections on which it runs . It can be s een that a s imilar R R D is obtained with all rail
~~~
~ ~+
~. ~ --a-~w. gip-- 21 .,~~R ti... ~_
CONF IDENTIAL
s ec tions . The c ons is tenc y in the c ontac t c onditions is a b enefic ial featu re of the profile des ign,
as it s implifies the monitoring and maintenanc e of the tramway s ys tem. However it is als o
noted that the level ofi c onic ity generated is relatively low, s u gges ting that depending on the
c u rve radiu s dis trib u tion of the s ys tem, there c ou ld b e margin to improve s teering, with
b enefits in terms of flange wear and rail s ide and keeper wear. An Ins tantaneou s c ontac t
angle c hange c an b e s een b y the rapid inc reas e in R R D as the wheel goes into flange
c ontac t, this is a res u lt of the low level of c onformality b etween wheel flange and gau ge
c orner radiu s , this c an dead to high initial wear rates and pos s ib le degeneration of lateral ride
as there is little trans ition b etween tread and flange c ontac t.
V HIVIYIKt YIOI
F igu re 3. 5 R olling radiu s differenc e, Croydon Tramlink CR 4000
F igu re 3. 6 b elow s hows a c omparis on of the R R D generate u s ing the c u rrent Croydon
CR 4000 tyre profile with the prototype CT3 profile on R i 59-R 13 (R i 59N) and S49
(inc lined at 120) rail s ec tions . A s light redu c tion in the R R D c an b e s een when u s ing the CT3
profile with an inc reas e in allowab le lateral dis plac ement b efore c ontac t with the flange.
s y
. . ~ . . . . ,: . . . ~ . . ; . ~ ,. . y~
22 v, ,a. ,,
,~ . ,
CONF IDENTIAL
F inal R eport
v rirv i rirtc VIOL
F igure 3.6 R oiling radius difference, Croydon Tramlink current CR 4000 and prototype CT3
tyre profiles
In comparis on the wheel- rail combination adopted by Nottingham Expres s Trans it generates
good conicity with groov ed and non- groov ed rails , inclined at 1 : 40, as s hown in F igure 3.7 .
The profile does not s how the s ame cons is tency acros s the rail s ections us ed as the Croydon
profile but prov ides a s mooth trans ition between tread and flange contact, which can be
beneficial in terms of ride q uality and wear.
v Hrv irint rio[
F igure 3.7 R olling radius difference, Nottingham Expres s Trans it
~~~
~
R*r
. Y ~..., , , .d..y~ , y.; a
2.~ t!r.dv
CONF IDENTIAL
F ig u re 3.8 below s hows the rang e of R R D g raphs for the Manches ter Metrolink tramway
s ys tem. A v ery larg e v ariation in conicity is fou nd, with low conicity g enerated on BS 113A
and BS 80A rail profiiles , and mediu m to hig h conicity lev els on R i 59-R 10 and SEI 35G
embedded s ections . In the cas e of SEI 35G, the hig h conicity is du e to the redu ced 1:40 rail
inclination. Su ch a v ariation in conicity can lead to difficu lties in manag ing the wheel-rail
interface, as maintenance requ irements can v ary widely. As the conicity increas es to hig h
lev els , is s u es with v ehicle lateral ins tability may aris e. The Metrolink wheel, when combined
with R i 59-R 10 and SEI 35G rail profiles prov ides a s mooth trans ition between tread and
flang e contact, which is in contras t to the trans ition obtained with the other adopted rail
s ections . This hig hlig hts the difference in conformality between flang e root and g au g e corner.
F ig u re 3,8 R olling radiu s difference, Manches ter Metrolink
As des cribed prev iou s ly in the report, the wheel tyre profiles u s ed by Midland Metro and
Sheffield Su pertram inclu de v ery s imilar tread and flang e g eometry res u lting in s imilar contact
conditions when combined with their res pectiv e rail profiles , as s hown in F ig u re 3.9. Both
produ ce low conicity on BS 80A rail, ev en when inclined at 1:40, and a hig her conicity on
g roov ed SEI 35G rail s ection.
r
t*w
. ~ .
CONF IDENTIAL
F igure 3.9 R olling radius differenc e, Midland Metro and S heffield S uper~ram
Bas ed on the plots of rolling radius differenc e the eq uiv alent c onic ity of eac h wheel- rail
c ombination has been c alc ulated us ing the method generally adopted in the UK.
The eq uiv alent c onic ity is c alc ulated by as s uming that the wheeis et lateral s hifts will v ary
ac c ording to a normal dis tribution about the zero s hift pos ition. A weighted bes t- fit line is then
c alc ulated to the s lope of the rolling radius differenc e graph. The eq uiv alent c onic ity is giv en
by half the s lope of this line.
F igures 3.10 and 3.11 below s how the c omparis on of the eq uiv alent c onic ity v alues obtained
for groov ed and non- groov ed rails res pec tiv ely.
*y
CONF IDENTIAL
S heffield S upertram - S EI 35G
No# ting ham Expres s Trans it - S EI 41GP
Midland Metro - S EI 35G
Manches ter Metrolink - S EI- 35G inc. 1: 40
Manches ter Metrolink - R i59- R 10
Croydon i' ramlink - R i 60- R 13
Croydon Tramlink ; - R i 59- R 13
_0 . 20 1
0 . 135
0 20 0
- 0 . 147
0 . 0 52 (0 . 0 39)
0 . 0 51 (0 . 0 39)
0 . 392
F inal R eport
U. UU U. IU U. LU V. S V V. 4V U. ~U U. OU
Equivalent Conicity
F ig ure 3. 10 Equivalent conicity, g rooved rail s ections , ins talled vertically unles s otherwis e
s tated (' `Croydon CT3 values are in parenthes es )
Tyne &Wear Metro - BR 113A
S heffield S . Tram - BS 80 A inc. 1: 40
Notting ham Expres s Trans it - BS 80 A inc. 1: 40
Midland Metro - BR 113A
Midland Metro - 8S 80 A inc. 1: 40
Manches ter Metralink - BR 113A
Manches ter Metroiink - BS 80 A
DLR - BR 113A
DLR - BS 80 A
Croydon Tramlink ' - S 49
Croydon Tramiink ` - BR 113A
I0 . 0 25
0 . 0 26
o. 0 0 s
0 . 0 50
o. o~~
0 . 0 94
0 . 0 50 {0 . 0 39)
0 . 0 52 (0 . 0 39)
0 . 234
0 . 310
0 . 0 0 0 . 10 0 . 20 0 . 30 0 . 40 0 . 5fl 0 . 60
Equivalent Conicity
F ig ure 3. 11 Equivalent conicity, non- g rooved rail s ections , inclined at 1: 20 unles s otherwis e
s tated (*Croydon CT3 values are in parenthes es )
It can be s een from F ig ures 3. 10 and 3. 11 that a larg e variation in conicity occurs , with values
rang ing from 0 . 0 6 to 0 . 557. Generally the lower conicity values are s een on non- g rooved rail
s ections with BS 80 A producing the lowes t values .
~~'
+s
, _. . ~, . ~. . 4 a
26 . , . , . , n. ,
~.
, N. . ~
CONF IDENTIAL
I n terms of the P has e 2 ob j ec tiv es this v ariation is v ery pos itiv e as it demons trates that b y
c omb ining as pec ts of different profile c omb inations there is an opportu nity to c reate wheel
and rail s ets whic h c an meet the wheel- rail interfiac e requ irements of s ys tems with widely
differing c harac teris tic s . I t als o demons trates that at pres ent there does n' t appear to b e a
' des ign for c onic ity' philos ophy, with an apparently random dis trib u tion of c onic ity v alu es
ev ident. The ab ility to optimis e c onic ity for a giv en s ys tem c an lead to gains in terms of
redu c ed wheel and rail wear and lowered derailment propens ity.
3. 1. 3 Contac t Angle and Wheel L ift
A u s efu l indic ator of prox imity to derailment is the v ertic al lift or flange c limb of the wheel
relativ e to the rail, referred to as wheel lift in mm. V ertic al wheel lift indic ates the amou nt of
flange c limb s een b y the wheel. F lange c limb c an b e tolerated withou t nec es s arily c au s ing
derailment, s o long as it is not s o great as to reac h the zone of rapidly redu c ing c ontac t angle.
When this zone is reac hed the prob ab ility of derailment is far greater, as the c ontac t angle
c an b ec ome s mall. Therefore the relations hip b etween c ontac t angle and wheel lift c an b e
u s ed as gu ide to the protec tion agains t derailment offered b y a giv en profile c omb ination.
As an ex ample, F igu re 3. 12 b elow s hows a c omparis on of wheel lift and c ontac t angle, plotted
agains t lateral s hift of the wheels et. The plot inc lu des a DI N- type wheel profile
( S heffield mod. DI N25) and 4he more non- c onv entional Doc klands DL R S , whic h inc lu des a
large flange root radiu s des igned to allow c u rv ing down to a s mall radiu s withou t ex c es s iv e
flange c ontac t. Both wheels are ru nning on BS 80A rail s ec tion.
v Hiv iriKe riot
F igu re 3. 12 Contac t angle v wheel lift
~~
~
r
, ~ ~, ~
CONF I DENTI AL
Determination of Tramway Wheel and R ail P rofiles to ~ A inimis e Derailment F inal R eport
Cons idering F igu re 3 . 1 2 , which u s es a combined y- ax is of wheel lift in mm and contact angle
in degrees , it can be s een that there is a s ignificant difference in the contact angle and wheel
lift relations hip for the two profiles . The flange root ofi the DIN- type profile res u lts in a very
rapid ris e in contact angle, once flange contact occu rs . The flange angle remains high u p to
8mm of lateral s hift. However, s ignificant wheel lift is occu rring at this point, of approx imately
1 0mm. A ny movement beyond 8mm of lateral s hift res u lts in a rapid decreas e in contact
angle, whils t wheel lift continu es to increas e. A t 1 5mm lateral s hift the contact angle has
dropped from 75 to only 3 0 degrees for 1 7mm of wheel lift. A t this point derailment wou ld be
very lik ely.
In contras t, at 8mm lateral s hift, the DLR profile s hows almos t no wheel lift, the increas e in
contact angle is far more progres s ive. A t 1 5mm of lateral s hift the contact angle remains in its
initial ris ing s tage, s howing 60 degrees with only 8mm of wheel lift. It can be s een that for the
s ame lateral s hift, the DLR profile can generate a large rolling radiu s difference, together with
a high level of derailment protection, provided by a high contact angle
The contact angle and wheel lift relations hip of the DLR profile is a fu nction of its intended
des ign philos ophy of s teering the wheels et u s ing the flange root. It is an ex treme cas e bu t
demons trates the u s efu lnes s of this type of analys is .
Tram profiles generally ex hibit the contact angle and wheel lift relations hip s hown by the
DIN- type profile above, the fu ll s et of plots can be fou nd in A ppendix A 1 .
3 . 2 Grooved R ail Wheels et F it
A s all the UK tramway s ys tems u s e grooved rail s ections it is es s ential to u nders tand the fit of
the wheels et within the grooved rail s ection. The initial fit in the new ` des ign' condition will
dictate the life- cycle performance of the wheels et in terms of available flangeway clearance
and hence ability to generate rolling radiu s difference ( ability to s teer and redu ce angle of
attack ) bu t als o in terms of the degree to which s idewear can be tolerated before the k eeper
rail mak es contact with the fiangeback of the wheel. A dditionally once k eeper rail contact
does occu r the thick nes s of the k eeper s ection is lik ely to dictate the remaining life of the rail
s ection, or where s ignificant maintenance intervention will be req u ired.
The clearances identified in the following s ection s hou ld be viewed in the contex t of
max imis ing the time in terms of rail s idewear before k eeper rail wear occu rs . It is cons idered
that k eeper wear will u ltimately lead to the life ex piry of the rail s ection as it will no longer be
able to withs tand the loads being imparted on it by the flangeback of the wheel and will
therefore begin to pres ent a derailment hazard.
#~ M
. . ~ ~ :-, -a. . . ~ :,
2 8 ti` arnhc~ . :rr
CONF IDENTIA L
F igure 3.13 below s hows the wheels et fit for two tramway s ys tems with v ery different
clearances . The wheels et is s hown in its central pos ition within the track gauge in the upper
plot. The centre plot s hows the wheels et at the point of flange contact, the final lower plot
s hows the wheels et in heav ier flange contact with a s mall amount of wheel lift. The figures
adjacent to the centreline s how the wheels et lateral s hift from the central pos ition.
j _ _ _ A, .. j ' _ ._
s d, ~ . ~
~ a~
, ._ ._ .~ .~ ~
a_ _ _ ~ _ _ _
'y, n~
~ .~
~ ....1'
~ ~
Twi k ~ e
` ~ 1..._ 7 !
~ ~ 1 ~ _ Y ~ +rR
l ---r ~ ~ ,
~ ~ LA E
Croydon Tramlink R i 59N S heffield S upertram S EI 35G
F igure 3.13 Wheels et fit
With reference to the abov e figures , the difference in des ign clearances is ev ident between
the two s ys tems . With the wheels et in the central pos ition Croydon tramway s hows a
clearance of 16 .1 mm to the k eeper rail, this compares to 7 mm in the cas e of S heffield
S upertram.
When in flange contact Croydon tramway s hows a good lev el of clearance of 11.2mm, whils t
at S heffield S upertram the clearance drops to 2.3mm. With a little further lateral s hift and a
s mall amount of wheel climb (which has been witnes s es at S heffield S upertram), the
clearance to the k eeper rail is eliminated completely.
It is clear from the abov e s imple geometry manipulation that with a s mall amount of s idewear
the wheels et at S heffield S upertram will be wearing the k eeper rail and reducing the life of the
rail s ection, whils t in contras t, the Croydon s ys tem will hav e a greatly extended s ection life
before k eeper wear begins to caus e problems .
~ ~ ~ ~ ~
+*`
P M ~ 'tin'Y M1
'F U
29 4? ! ^ .: IfCS '.f'
~ ~ ~ ~ ~'i iy.~ lil ~ -
~ -ril' .
CONF IDENTIAL
I n addition to the keeper c learanc e, F ig u re 3.13 als o s hows the approximate wheel s afety
heig ht and width as des c rib ed in G u idelines ~ 5 ~ .
Tab le 3.1 b elow s u mmaris es F ig u re 3.13 with the addition of res u lts from the s ame analys is
applied to the other tramway s ys tems .
Tramway Network(G rooved R ail P rofile)
.N
c
m
Y
C
~ C
C
~
to
Y
C
Y
C
O O N N
~ ~ ~ c S O X
~
N
N
~ M
~
N
~ lh
L
~ N, O~
y
~
E Q.
~
7
to
p R ~ ' p ~
N
N ~
y ~
~ to
~
~ y~
~
~
~ {~
~ ~
A~ T (~ O V ~
C
V M
C
~ M ~ ~ ' ~ M
~
V ~ V ~ ~ ~ ~ ~ ~ ~ Z~ (~ ~
F lang eb ac k Dim./mm 1380 1380 1362 1362 1379 1380 1379
Trac k G au g e/mm 1435 1435 1435 1435 1435 1435 1435
Tie Bars No No No No No No No
F lang eway 5 .5 5 .5
3.6 3.6 5 .1 5 .0 4.7
Clearanc e/mm (6.7) (6.7)
Keeper Clearanc e/mm 16.1 10.0
15 .7 9.3 7.0 14.1 7.0
(Centered) (15 .5 )
(9.4)
Keeper Clearanc e/mm 11, 2 4.7
.~ Z6 5 .5 2.1 9.3 2.3
(F lang ing ) (10.4) (3.9)
*Safety Heig ht/mm
(16.9) (16.9)
13.1 13.1 17.8 16.4 18.2
*Safety Width/mm ~ 7: 6~ ~ 7: 6~ 6.8 6.8 7.0 6.5 7.4
Tab le 3.1 -G rooved R ail Wheels et F it
(Croydon CT3 valu es are in parenthes es )
Cons idering Tab le 3.1, the differenc es in g rooved rail b as ed s ys tems is apparent. SEI 35 G
offers low keeper c learanc e du e to its relatively s mall g roove width (s ee F ig u re 2.1) , res u lting
in low c learanc es at b oth Midland Metro and Sheffield Su pertram, whils t the wider R i 5 9-R 13
profiles ' wider g roove wid# h allows for g reater c learanc es . I t s hou ld b e noted that the s mall
keeper c learanc e does not pos e a g reat prob lem at Midland Metro as the s ys tem is relatively
s traig ht, this point demons trates the requ irement to des ig n for applic ation.
~ T
+faM
ti, , ~ > : n ~ ~
~ N, .r~ ,
CONF I DENTI AL
The maj ority of the tramway s ys tems have flangeback dimens ions nominally of 1380mm, with
the exception of Manches ter Metrolink, which operates at 1362mm. This difference is
accommodated in the relief of the back of the flange, effectively creating increas ed clearance
to the keeper.
None of the s ys tems s tu died u s e tie bars , this is s ignificant when cons idering that one of the
maj or cau s es of prematu re keeper rail is gau ge s pread du e to cu rving forces . The adoption of
tie bars wou ld s ignificantly increas e the life of the rail s ection in cu rved track.
F langeway clearances , the nominal clearance between the wheel flange and the rail gau ge
corner, is generally in the order of 5mm, with the exception of Manches ter Metrolink which
operates at only 3.6mm, des pite a s maller flangeback dimens ion. This lower valu e can be
attribu ted to the lack of conformality between flange and gau ge corner radiu s , res u lting in
early contact. A lack of conformality can lead to high initial rail s idewear rates , this will be
dis cu s s ed in more detail in Section 7 of the report.
~~
~ s r
,. ~..~..h.: ~ ~.w.
31 ~' 4 hS5` .I '
i..~N.,l) 1; ~
CONF I DENTI AL
4. P rofile Selection for Simulation Cas es
Due to the larg e number of pos s ible combinations of wheel and rail profiles , it was es s ential
for the later s imulation work that the number of s cenarios was reduced. The following s ection
des cribes the profiles which were identified for further analys is work, tog ether with reas ons for
their s election. The s election criterion is primarily bas ed on the analys is carried out in the
prev ious s ections .
4.1 Selected Wheel P rofiles
Table 4.1 below s hows the fiv e wheel profiles currently propos ed for the s imulation work.
Thes e profiles were approv ed at the working g roup meeting held on 13 ` h J uly 2007.
Cas e P rofile R eas ons for Selection
Al NET P -3 10263 9
Medium-hig h conicity on g roov ed /non-g roov ed rail
Conformal contact conditions
Dual cone ang le of tread, benefits conicity and conformality
A2 SST Mod DIN25
R eplaces Midland Metro as almos t identical
Different flang e root radius and cone ang le to NET DIN-type
Good conicity on SEI 3 5G
B Metrolink MML2
Us e of flang eback cut-out for potential tram -train application
R ound tip flang e des ig n
C Croydon CR 4000
Cons is tent contact conditions acros s a rang e of rail profiles
15mm flang e root radius
D CT3
Tramway optimis ed profile
Good contact and conicity conditions , 15mm flang e root
radius
Table 4.1 P ropos ed Wheel P rofiles
The propos ed wheel profiles offer a rang e of characteris tics which s hould, when combined
with a s uitable rail s ection, cov er the likely demands of many types of tramway s ys tem des ig n
layouts . The philos ophy of the P has e 2 work is that this will be demons trated throug h v ehicle
dynamic s imulation as des cribed in Sections 6, 7 and 8 of the report.
Table 42 below s ummaris es the profiles rej ected and the reas on for rej ection.
P rofile R eas on s for Non-Selection
Midland Metro Similar profile eomet and contact conditions to SST
T ne &Wear Conv entional mainline railwa rofile
Table 42 R ej ected Wheel P rofiles
+si
CONF IDENTIAL
4.2 Selected R ail P rofiles
R ail profiles were s elected primarily on the b as is of cu rrent frequ ency of u s e, av ailab ility and
als o whether they cou ld b e repres ented b y another profile, i.e. s ome profiles cou ld b e
eliminated as they are identical to other profiles other than for example groov e width and
depth.
The following s ection ou tlines the rail profiles , b oth groov ed and non-groov ed, which were
s elected for fu rther analys is work. The s u gges ted comb inations were agreed at the working
grou p meeting held on the 1 3 ' " J u ly 2007.
4.2.1 Groov ed R ail P rofiles
Throu gh analys is of the P has e 1 report, F igu re 4.1 (als o F igu re 2.1 ) was created which allows
the comparis on of head and groov e profile for effectiv ely all the groov ed rail s ections
commonly u s ed and av ailab le to tramways .
Groov etl R ail Sections
70
60
50
40
3 0
20
R i 5&R 1 0
R i 59- R 1 3 (R i 59N)
R i 60- R 1 0
1 0 - ~- R i 60- R 1 3 (R i 60N)
SEI 3 5G
SEI 3 5GP
SEI 4IGP
0
0 1 0 ZO 3 0 40 50 60 70 80 90 1 00 1 1 0 1 20
F igu re 4.1 Groov ed R ail P rofiles
With reference to the ab ov e figu re it can b e s een that es s entially there are only two types of
rail head profile, the R i 59/60- R 1 0 (1 0mm gau ge corner radiu s ) and the R i 59/60- R 1 3 , which
has a 1 3 mm gau ge corner radiu s and is inclined at 1 : 40, as oppos ed to v ertically for the other
profiles . Otherwis e the differences b etween the profiles are du e only to the groov e depth and
keeper rail clearance, thicknes s and height.
~r~~
~ `
,..._ .~~-,,..~..
3 3 t,a~: hs ,.,.
CONF IDENTIAL
Table 4.3 below s hows the grooved rail s ec tions s elec ted for the s imu lation s tage of the
P has e 2 work and reas ons for s elec tion.
Cas e P rofile R eas ons for Selec tion
A R i 59-R 10
I dentic al rail head profile to other grooved rail s ec tions
10mm gau ge c orner radiu s
B R i 59-R 13 (R i 59N)
I nc lined rail head profile, 1:40
13mm gau ge c orner radiu s
Table 4.3 Selec ted Grooved R ail Sec tions
As c an be s een from the above table, j u s t two grooved rail profiles c over the req u ired rail
head s hape. The wide variation of groove width will be ac c ommodated within the VAMP I R E
s imu lations throu gh variation of the trac k des ign parameter file. P arameters s u c h as groove
depth and keeper thic knes s will be pres ented in the ` Bes t P rac tic e' gu ide (9] .
4.2.2 Non-Grooved R ail P rofiles
The s elec tion of non-grooved rail s ec tion was more c omplex than that of the grooved rails as
there is a lac k of s imilarity in the head profiles . However, the profiles s elec ted and the reas on
for s elec tion are s u mmaris ed in Table 4.4 below.
Cas e P rofile R eas ons for Selec tion
A BS 80A (1:20 and 1:40)
F re u entl u s ed on UK tramwa and li ht rail s s tems
General) res u lts in low c onic it u nles s inc lined at 1:40
B BR 113A (120)
F re u entl u s ed on UK tramwa and li ht rail s s tems
Benefits for fu tu re' Tram-Train' links
C S49 / UI C 54 (1:20) F req u ently u s ed on Eu ropean tramway s ys tems
Table 4.4 Selec ted Non-Grooved R ail Sec tions
s r
u ., ~ , ,
CONF I DENTI AL
5. Vehicle Dynamic Simulation Overview
A key ob j ective of this profile optimis ation s tudy is to as s es s the performance of the
s hort-lis ted profiles (s ee Section 4 for details on the profile s election) with res pect to
derailment protection and minimis ation of wheel and rail profile wear. This analys is was
performed us ing the VAMP IR E railway vehicle dynamics s imulation s oftware.
5.1 The VAMP IR E P ackag e
VAMP IR E has b een developed and validated over a numb er of years , formerly b y BR
res earch and latterly b y AEA Technolog y R ail (now DeltaR ail L td) . VAMP IR E us es a
multi-b ody modelling me#hod, where a dynamic s ys tem, s uch as a rail vehicle, can b e
cons idered as a collection of rig id b odies , with mas s es acting at their g eometric centres .
Thes e b odies are then interconnected b y mas s les s s pring and damper elements to b uild up a
complete s ys tem repres enting the vehicle. Equations of motion are g enerated automatically,
along with contact tab les to repres ent wheel and rail g eometry. As VAMP IR E is s pecifically
written fir the s iml~ lation of rail vehicles , s mall ang le approximations are us ed, to increas e
s olver s peed, res ulting in low s imulation times , without s acrificing accuracy over the s mall
ang les involved.
VAMP IR E`~ 's pos t-proces s ing capab ilities include s tatis tical, time and frequency domain
analys is , peak counting , s tab ility and wheel-rail wear alg orithms , along with the ab ility to
model s us pens ions with coulomb friction and dynamic normal forces .
5.2 Generic Vehicle Model
The adopted philos ophy in terms of vehicle dynamic modelling is to us e a `g eneric' vehicle
model which is capab le of repres enting the maj ority of vehicle config urations identified in
Section 2.5. N eg lecting Blackpool tramway and the N ational Tramway Mus eum, s ix out of the
remaining s even vehicles us e conventional axles , with two of thes e b eing equipped with
independently rotating wheel (IR W) equipped trailer-trucks . N ET tramway is the only operator
to us e all independent wheels b ut thes e are powered. At the time of writing the deg ree to
which powered IR W's b ehave as conventional wheels ets was un-confirmed.
Given the ab ove, the chos en approach was to us e a vehicle model which comb ined a
powered b og ie with conventional axles , b ut als o incorporated an IR W equipped trailer-truck
s ection. This would cover the g eneral performance of s ix of the s even UK tram vehicles .
its
CON F IDEN TIAL
I t c ou ld be s u gges ted that this approac h does not ac c ommodate other as pec ts whic h may
effec t the s teering, derailment res is tanc e and lik ely wheel- rail wear of a tramway v ehic le,
howev er, limits mu s t be impos ed to be able to prac tic ally analys e a large c ombination of
wheel and rail profiles .
The fu ndamental bas is of the generic approac h is to prov ide a relativ e indic ation of the
performanc e adv antages of a partic u lar wheel and rail profile c ombination. This c an be
ac hiev ed u s ing a s ingle generic v ehic le model as the improv ements will in the mos t c as e
c arry ov er to v ehic les with differing artic u lation lengths , wheel bas es etc , of c ou rs e the
magnitu de of potential gains my differ bu t the general improv ement trend will remain.
5 . 2 . 1 Generic Vehic le Ov erv iew
The v ehic le s elec ted as a generic v ehic le model is a bi- direc tional tram, with driv ing c abs at
both ends , c ons is ting of two c ar bodies c onnec ted by an intermediate artic u lated s ec tion.
The two main body s ec tions are c ou pled to the c entre s ec tion by s pheric al j oints and
therefore the c entre bogie s u pports the end mas s of thes e two s ec tions , together with the
c entre s ec tion.
The two ou ter bogies are powered u s ing amono- motor s etu p i. e all fou r wheels driv en
together, the primary s u s pens ion arrangement u s es a twin ru bber c onic al s pring arrangement
ac ting on an ou tboard axlebox, a lev el of damping is prov ided by the s pring' s ru bber
c ons tru c tion ( hys teres is ) . Sec ondary s u s pens ion is prov ided v ia a bols ter, to the c ar body,
u s ing airs prings , and is equ ipped with lateral and v ertic al hydrau lic dampers . The primary
arrangement is s hown in F igu re 5 . 1 below.
~ , , :
~ _~
# _
. - Vii' 3 " t
_ _ . - ~ ' :
s . _ _
T' a~ ~ : . _
__. . ~ __
- - ~
_~
F igu re 5 . 1 Motor Bogie P rimary Su s pens ion Arrangement
rww~ 3 6 ~ , . ~- h~ , . r
Y i' ' I ~ ~ y~ ~ lll ~ ~
CONF I DENTI AL
A s tewing ring provides bogie rotation relative to the body, this is a low fric tion interfac e and
does not provide s ignific ant rotational damping or s tiffnes s to the bogie.
The c entral trailer bogie is q u ite different to the two motor bogies , operating with
independently rotating wheels , this permits a low floor between the two ou ter bogies . P rimary
s u s pens ion on the trailer bogie is a les s c onventional bell-c rank type arrangement, a radial
arm extends behind the wheel, as s hown in F igu re 5 . 2 , and s u pports the wheel within the
bogie frame, via a large radial bearing. The s tiffnes s is provided by a metalas tic type c hevron
s pring s hown direc tly above the radial bearing. This als o provides a level of damping throu gh
its ru bber elements .
~ is L; K
~F
F igu re 5 . 2 Trailer (c entre vehic le) P rimary Su s pens ion
Sec ondary s u s pens ion is provided by airs prings mou nted direc tly between bogie and body.
Longitu dinal (trac tion and braking) res traint between the body and bogie is provided by
trac tion s tops , rather than the rods u s ed in the motor bogies , lateral bu mps tops als o ac t
between body and bogie. Sec ondary damping is provided both laterally and vertic ally by
hydrau lic dampers .
5 . 2 . 2 The VAMP IR E Model
The generic vehic le model was s etu p and validated u s ing the c alc u lated inter-body s tatic
loads and freq u enc y res pons e data from the manu fac tu rers report. The vehic le model is
s hown in F igu re 5 . 3 and c ons is ts of the following elements :
16 Mas s es
6 Linear s prings
16 Shear s prings
7 Vis c ou s dampers
4 Bu mps tops (non-linear s prings )
11 P in links
CONF IDENTIAL
it
20 Bus h elements
The model c ons is ts of 102 Degrees of Freedom (DoF).
_ VAM#~IR E~ : VEHICLE ~L~IT
... _ ~~; _
.. .- _
- ~ ;~:~ _ ~
Figure 5.3 - VAMP IR E Generic Vehic le Model
5.2.3 Motor Bogie Model
+1A~~'~~E~ : VEHICLE F'l~O T
:'.
'`_ ~. ;~ -
fi
~:.
~.
MT^'"' X
S
~~~ ~~_ t
Final R eport
3
Vt\MiP TR F P 1 c at.
Figure 5.4 VAMP IR E Motor Bogie Model
With referenc e to Figure 5.4 above, the primary s us pens ion was s et up us ing two s hear
s prings per ax lebox , providing longitudinal, lateral and vertic al res traint to the wheels et. The
~~ .J +s Y
,~.,~ ;:
CO NFIDENTIAL
b ols ter and s tewing ring were inc orporated as a s eparate mas s , c onnec ted to the b ody v ia a
b us h element and to the b ogie v ia the s ec ondary s us pens ion. P in link elements allowed
ac c urate repres entation of the inc lined s ec ondary lateral hydraulic dampers and trac tion rods .
V ertic al s ec ondary damping us es a s tandard damper element. A irs pring elements were us ed
to repres ent the ins talled airs prings .
5 .2 .4 Trailer Bogie Model
With referenc e to F igure 5 .5 , the trailer b ogie model was dev eloped around the b ell- c rank
s us pens ion arrangement, thes e c an b e s een as rec tangular b oxes around the wheels , with a
piv ot b us h us ed at their lower inner c orner to repres ent the radial b earing.
V ~iNiP IR E~ 1I~HTCLE F 'L~T
.~ ...,.
,nom, i'~r~tW' ~ ^ ~
e 1 yr ~c r~.g I.~r
_ ..._
._
~- ~ r
,,_ _ _
_
_ _ _ _
- - _
f
~ - t
_ .._ _ _ _
j
~ _- _ ~
~_ ~._ _ F _
_ _ _
V A MP TR F P I c ~k .
F igure 5 .5 V A MP IR ES Trailer Bogie Model
The primary s prings were attac hed to the b ell- c rank arrangement. This es s entially giv es the
wheels et only a v ertic al degree of freedom ( ac tually a rotation ab out the b ell-c rank ) , as the
radial b earing prov ides a relativ ely high s tiffnes s in the other direc tions . There was no
s eparate b ols ter mas s in the trailer b ogie model, the airs prings are c onnec ted direc tly to the
b ody and no s tewing ring is req uired, the c entre b ogie ac ting as a 'truc k ' and as s uming a
tangential pos ition in c urv es .
The s ec ondary s us pens ion inc orporates a s ingle lateral damper, repres ented b y an inc lined
pinlink . The two v ertic al dampers are modelled us ing s tandard linear elements . A irs pring
elements are us ed to repres ent the airs prings . Trac tion s tops and lateral b umps tops are
inc luded b etween the b ogie and b ody.
~i
,..,~ .
CONF IDENTIA L
6. Dynamic Simulation Derailment Study
The following s ections (6, 7 and 8) detail the res ults of the s imulation s tudies . I n total, fiv e
wheel and s ix rail profile combinations (30 s ets ) were analys ed. L ater in the work additional
s imulations were performed with modified wheel and rail profile config urations , thes e will be
documented later in the report.
Table 6.1 below des cribes the fiv e wheel and s ix rail profile combinations s elected as a res ult
of the initial contact s tudy des cribed in Section 4.
Cas e Wheel P rofiles R eas ons ) for Selection
Al NET P -31 02639
Medium-hig h conicity on g roov ed /non-g roov ed rail
Conformal contact conditions
Dual cone ang le of tread, benefits conicity and conformality
A2 SST Mod DI N25
R eplaces Midland Metro as almos t identical
Different flang e-root radius &cone ang le to NET DI N-type
Good conicity on SEI 35G
B Metrolink MML 2
Us e of flang eback cut-out for potential tram -train application
R ound tip flang e des ig n
C Croydon CR 4000
Cons is tent contact conditions acros s a rang e of rail profiles
1 5mm flang e root radius
D CT3
Tramway optimis ed profile
Good contact and conicity conditions , 1 5mm flang e root
radius
Cas e
Groov ed R ail
P rofiles R eas on s for Selection
A R i 59-R 1 0
I dentical rail head profile to other g roov ed rail s ections
1 0mm g aug e corner radius
B R i 59-R 1 3 (R i 59N)
I nclined railhead profile, 1 :40
1 3mm g aug e corner radius
Non-g roov ed R ail
Cas e P rofiles R eas on s for Selection
A BS 80A F re uentl us ed on UK tramwa and li ht rail s s tems
Generall res ults in low conicit unles s inclined at 1 :40 (1 :20 and 1 :40)
B BR 1 1 3A F re uentl us ed on UK tramwa and li ht rail s s tems
Benefits for future tram -train links
(1 :20}
C S49 / UI C 54
F req uently us ed on European tramway s ys tems
(1 :20)
Table 6.1 Wheel and rail profiles s elected for s imulation
I n order to as s es s the performance of the s elected wheel and rail profile combinations , two
dynamic s imulation. cas e s tudies were dev eloped. The firs t s tudy was us ed to as s es s the
derailment protection offered by each of the profile combinations , by s ubj ecting the v ehicle to
~~
4L t~.rVulU n
-....-~ ~1 ~. .
CONF I DENTI AL
s evere trac k twis t and c u rvatu re, the s ec ond s tu dy employs a c u rvatu re as s au lt c ou rs e and
was u s ed to as s es s s teering ab ility, predic ted wear levels and c ontac t s tres s . An additional
s tab ility s tu dy was c arried ou t to inves tig ate applic ation of hig her c onic ity profiles to
independent wheel equ ipped b og ies , this is detailed in Sec tion 8.
All profiles were s imu lated in their new c ondition, with 1435mm trac k g au g e and a flang eb ac k
s pac ing relevant to the s ys tem from where the wheel profile orig inated. The g eneric vehic le
model in a tare loading c ondition was u s ed for all s imu lations . The tare c ondition was s elec ted
b as ed on it b eing the wors t c as e in terms of derailment propens ity. A fric tion c oeffic ient of
0. 30 for wheel and rail was u s ed in the c u rving s tu dy and 0. 4 for the derailment s tu dy.
following analys is of the s imu lation work , the c omb inations of profiles lis ted ab ove will b e
redu c ed in nu mb er, to form profile s ets . The profile s ets will c ontain wheel and rail profiles
whic h have throu g h the s imu lation work , b een demons trated to b e c ompatib le with eac h
other, in terms of maximis ing derailment protec tion and minimis ing wheel and rail wear. The
final s elec tion of profiles s ets will b e explained in Sec tion 9.
, As there are a larg e nu mb er of s imu lation c as es , the entire s et of s imu lation res u lts will not b e
pres ented in the b ody of the report, only thos e whic h are mos t relevant to the final profile
s elec tion. A c omplete s et of s imu lation res u lts is provided in the Appendix.
6 . 1 Derailment Stu dy Overview
As des c rib ed ab ove, a method was requ ired to as s es s the derailment protec tion offered b y
eac h of the s elec ted wheel and rail profile c omb inations . The proximity of a wheel and rail
profile to derailment is u s u ally as s es s ed throu g h analys is of the Y/Q ratio, this is a ratio of the
lateral forc e (Y) to the vertic al forc e (Q), ac ting on a wheel attac k ing the g au g e c orner of the
rail. The s implified forc es ac ting on the wheel and rail c an b e s een in F ig u re 6 . 1 b elow.
, i' ;
~ ` tu n er - , u
~ } j~~~~~~~~~~
F ig u re 6 . 1 Y/Q relations hip
. ~ ~
. . , , ~ .
CONF IDENTIAL
if the component ofi T (a fu nction of N and N) exceeds the combined component of Y and Q
then flang e climb will initiate. The Nadal formu la dictates that a Y/Q ratio of 12 is the limiting
v alu e for a flang e ang le of 68 deg rees with a typical friction coefficient.
The Y/Q ratio is an approximation of the proximity of the wheels et to derailment, therefore
s u pplementary data is extracted from the s imu lation, this inclu des the amou nt of wheel lift or
flang e climb and als o the chang e in v ertical load on a wheel, in relation to the s tatic load,
denoted as the ~Q/Q ratio. Typically in a dynamic s itu ation a limit v alu e of 0. 8 (80%
u nloading ) of the wheel is permitted.
I n g eneral it is not s traig htforward to derail a railway wheels et, often cu rv atu re alone is not
s u fficient as cant deficiency tends to load the wheel attack ing the flang e, thereby increas ing
the `Q' v alu e and maintaining res is tance to derailment. Typically, to cau s e derailment it is
neces s ary to hav e an increas ed lateral force `Y', combined with a redu ction in v ertical load
'Q'. I n practice this u s u ally occu rs du e to either one or a combination of track twis t and cant
exces s (u nloads flang ing wheel), cant exces s often occu rring du e to low s peed ru nning ,
pos s ibly throu g h s peed res trictions etc. in order to fu lly u nders tand the derailment protection
offered by the s elected wheel- rail combinations , the derailment s imu lation s cenario mu s t
promote flang e climb derailment, this will rev eal the extent of derailment protection at the
extremes of operating conditions . F ig u re 6. 2 s hows the VAMP I R E track file dev eloped to
promote flang e climb derailment within the s imu lations .
~A1~fi~'iF t~ ~~~~~ _.
F ig u re 6. 2 Derailment s tu dy track file
::
~r xw. , . , ~, u a 42
, , ~ r ,
CONF I DENTI AL
With referenc e to the F igure 6 . 2 , the y- ax is is a c ombined ax is repres enting c urv ature in
rad/km, v ertic al trac k pos ition and c ros s lev el (c ant) both in mm.
A c ons tant 2 0m c urv e radius (50km- ' c urv ature) is c ombined with atwo- part twis t profile of
1:2 00 ac ros s the bogie c entres and 1:75 ac ros s the bogie wheelbas e. The trac k twis t is
ac hiev ed through applic ation of the v ertic al and c ros s lev el profiles s hown in the abov e figure
and pic torially in F igure 6 . 3 . The v ehic le trav ers es the trac k s ec tion at approx imately 10kph,
res ulting in a 50mm c ant defic ienc y in the s teady- s tate s ec tion of the profile. This is deemed
to be a typic al v alue bas ed on light rail s ys tem s peed/c anUc urv ature profiles .
To further promote flange c limb, the fric tion c oeffic ient between wheel and rail was inc reas ed
from a typic al v alue of 0. 3 2 up to 0. 4 .
Bogie C entre Bogie C entre
1:2 00
1:75
F igure 6 . 3 - Trac k twis t profile applied in derailment s tudy
F igure 6 . 3 abov e illus trates the trac k twis t profile applied in the derailment c as e s tudy, this is
in addition to a c ons tant c urv ature of 2 0m radius .
The applic ation of a s mall radius c urv e promotes derailment through generation of a high
angle of attac k between the wheel and rail, this in turn mov es the c ontac t patc h forward of the
ax le c entre line and reduc es the effec tiv e flangeway c learanc e, all thes e fac tors inc reas ing
the ris k of derailment. The lateral forc e is als o inc reas ed (an inc reas e in ` Y ' ) through c ant
defic ienc y and attac k angle.
The applic ation of a twis t profile unloads the leading wheel whic h is attac king the high rail, this
reduc es ` Q ' .
An inc reas e in the flange fric tion c oeffic ient inc reas es the c limb- out forc e, ' T' , whic h is
oppos ed by the ` Q ' forc e, therefore this is effec tiv ely a reduc tion in ' Q ' (s ee F igure 6 . 1).
s *
. , . ~ a. , . t, , , t
- . 4 3
. . . . , ~ .
C ONF IDENTIAL
Determination of Tramway W heel and R ail P rofiles to Minimis e Derailment F inal R eport
6.2 Derailment Study R es ults
The derailment s tudy inv olv ed s ub j ec ting the v ehic le and wheel- rail c omb inations to the
VAMP IR E trac k file and s imulation s c enario des c rib ed ab ov e.
R es ults and c ommentary are pres ented for Y/Q ratio, wheel lift, c ontac t ang le and ang le of
attac k , for typic al c as es . Before pres enting the detailed res ults , a s ummary of the res ults is
des c rib ed.
6.2.1 Derailment Study R es ults Summary
The firs t point to note from the derailment s imulations is that the Manc hes ter Metrolink (MML)
profile in the new c ondition derailed when c omb ined with any of the s ix rail s ec tions . In
c ontras t there were no s imulated derailments for any of the other wheel profiles .
The c aus e of the prob lem with the MML profile is b eliev ed to b e as s oc iated with the flang e
s hape, as s hown in F ig ure 6.4 b elow.
With referenc e to the ab ov e fig ure whic h s hows the MML and the
Notting ham Ex pres s Trans it (NET) wheel profiles , it c an b e s een that MML profile has a
s ig nific antly g reater flang e heig ht, whic h would perhaps s ug g es t g reater derailment
protec tion. Howev er, the flang e ang le c an b e s een to drop off rapidly at the b lend of the
flang e root radius and flang e fac e (- 25 on y- ax is ), when c ompared to the NET profile.
It will b e s een from F ig ure 6.5 and 6.6 b elow that this flang e s hape enc ourag es wheel lift and
it is this lift whic h further enc ourag es derailment.
CONF IDENTIAL
~i111~ ~~F 1TA~
.._ . t~~ 1;54 ~ 5' s a;e~s :R q~ - w:m; ...
~' .~io' 1~D~4ae ~u wrc ::t 1~r.~
~~~~~~iP ` 116 !
F igu re 6 .5 MML and NET c ontac t angle v wheel lift
F igu re 6 .5 s hows the relations hip between c ontac t angle, the angle of the plane of the c ontac t
patc h to the horiz ontal, and the vertic al lift ( flange c limb) of the wheel plotted agains t lateral
s hift of the left wheel when moving towards the rail gau ge c orner.
The c ontac t angle between wheel and rail is c ritic al to protec ting agains t derailment, a c ertain
amou nt of wheel lift c an be tolerated, providing that the c ontac t angle remains high enou gh s o
as to c au s e the wheel to drop bac k down into the gau ge. The above plot wou ld s u gges t that
the MML wheel profiles is s u perior in this s ens e to the NET profile as the c ontac t angle does
not fall off as rapidly and als o has a dis tinc t s u s tained c ontac t angle of 6 8 , j u s t u nder 5mm
on the x- axis . Als o the wheel lift is les s for a given lateral s hift. So in s u mmary for any lateral
s hift the MML, wheel lift is lower and the c ontac t angle is higher, s u gges ting higher derailment
protec tion than the NET profile. I t is only when looking c los er at the wheel lift to lateral s hift
relation that the reas on for the MML wheels inferior derailment performanc e bec omes
apparent.
F igu re 6 .6 below s hows the relations hip between wheel lift, as pres ented above bu t withou t
the c ontac t angle, plotted agains t the lateral s hift of the wheels et, bu t for all five wheel profiles
s tu died, on R i 59- R 13 embedded rail.
:i
' J .:~..y..lif ~~
..~ ri1~.
CONF I DENTI AL
Va(Mr~4~ Cd1~ITACT DATA P L07
- . ' S 6*~M9~F ,2 Nlwiiii LNnAnu
L~AMP IR E P lot
F ig ure 6 . 6 Wheel lift c harac teris tic all wheel profiles
With referenc e to the above fig ure, there is a c lear differenc e between the wheel lift
c harac teris tic of the MML profile when c ompared to the other wheels s tudied. Wheel lift
beg ins early at around 4. 8mm lateral s hift and has an initially lower g radient, partic ularly when
c ompared to the DIN- type tramway profiles s uc h as NET and S S T. At the point where the
flang e root radius meets the flang e fac e, around 5. 6 mm lateral s hift the rounding off of the
flang e fac e (s ee F ig ure 6 . 4) , c aus es a s ig nific ant further reduc tion in the s lope of the wheel lift
relation.
Des pite a hig her c ontac t ang le, the s lope of the wheel lift relation s eems to g overn the lik ely
c limb- out behaviour of the wheel, the c harac teris tic relates well to the s imulated derailment
performanc e and the obs erved s hape of the profile s een in F ig ure 6 . 4. Ail profiles with s teeper
wheel lift g radients res is ted derailment, even with hig her abs olute lift values and lower c ontac t
ang les .
Therefore it c an be c onc luded that the relation between wheel lift and c ontac t ang le,
partic ularly the g radient of the lift c harac teris tic has a s ig nific ant effec t on derailment
propens ity.
Due to the s imulated derailments us ing the new c ondition MML wheel profiles , additional
s imulations were performed c ombining new wheels with worn rails and worn wheels with new
and worn rails .
CONF IDENTIAL
I t was found that the new wheel als o failed the derailment as s ault cours e when running on a
worn rail profile, however when a worn wheel profile was s imulated, this res is ted derailment
when running on b oth new and worn rail profiles .
I t s hould b e noted that the derailment as s ault cours e us ed in this s tudy is an extreme
derailment as s es s ment and the track g eometry us ed would fall outs ide any s ys tem twis t
limits , therefore the fact that the MML profile exhib its a tendency to derail when s ub j ect to this
track g eometry does not indicate, in ab s olute terms , that the wheel profile is uns afe, only that
it has a lower protection ag ains t flang e climb derailment relative to the other profiles s tudied.
With reg ard to the g eneral derailment performance of the s elected wheel profiles , the s tudy
s hows that the SST and NET profiles , b oth DI N- type profiles , offer the g reates t res is tance to
flang e climb . This is evidenced throug h a g ene~ ~ ally lower Y/Q level acros s the s ix rail profiiles
s imulated and a cons is tently hig h contact ang le. The hig h contact ang le is a function of the
relatively s teep flang e ang le (70- 75)and als o the flang e tip g eometry, where the flang e ang le
is s us tained up to the point where the flang e face meets the tip radius .
The Croydon Tramlink CR 4000 and CT3 prafiles s how g reater levels of wheel lift acros s all
rails s ections , this is a function of their g reater flang e root radius , which promotes tread lift
throug h the flang e root running up the g aug e corner. This does not neces s arily res ult in
g reater derailment ris k as a hig h contact ang le is maintained throug hout the wheel lift s tag e.
The SST profile s hows g enerally low levels of wheel lift, with the exception of when comb ined
with S49 rail at 1:20 inclination and BS 113A at 1:20, where wheel lift b ecome s ig nificantly
g reater, typically 7- 8mm. This is caus ed b y the interaction of the flang e root and g aug e corner
radii at 1:20 inclination. Effectively, the g radient of the wheel lift vers us lateral s hift
characteris tic b ecomes g reater for thes e tow profiles , res ulting in hig her lift values . A hig h
contact ang le s us tains g ood derailment protection.
The NET profile s hows a s imilar trend to the SST wheel b ut s hows increas ed lift when
comb ined with BS 80A rail inclined at 1:40. Similarly, the amount of wheel lift is a function of
the comb ination of flang e root and g aug e corner radius at a g iven rail inclination. As is the
cas e with all profiles other than MML, res is tance to derailment is maintained throug h a
cons is tently hig h contact ang le and hig h g radient of the wheel lift, lateral s hift relation.
As would b e expected, the res pons e of the vehicle to the s evere twis t and curvature profile of
this s tudy res ults in hig h Y/Q values , typically lying in the rang e of 1 to 1. 2 for all profiles , with
the exception of the MML profile which exceeds thes e levels at derailment. I t is interes ting to
note that des pite thes e hig h Y/Q values derailment is not predicted for any profile other than
MML, hig hlig hting the cons ervatis m of the Nadal Y/Q q uotient.
f~ M
w L. , . . b . . ~ ,
:~ . - 47 . , :~ , s , . ,
. . ~ , ~ .
CONF I DENTI AL
The angle of attac k indic ates how far a wheels et has moved firom a radial s teering pos ition. I f
a wheels et is s teering perfec tly then there will be a zero angle of attac k . I n very low radiu s
c u rves , s teering is c ompletely brok en down and the angle of attac k bec omes high. The higher
the angle of attac k , the greater the ris k of derailment (See F igu re 6 . 7 ) .
Du e to the fac t that at very low radiu s s teering has brok en down for all the wheel profiles , the
angle of attac k es s entially bec omes the s ame for all profiles . This is bec au s e at thes e radii,
the angle of attac k bec omes a s implified fu nc tion of the bogie wheelbas e, c u rve radiu s and
flangeway c learanc e, s ee F igu re 6 . 8 .
o
a
j
a h, i~ ; ; ~ n
~ l' ~ . tl,
Q ' ~ ~ ; dltx
~ ~ ~
9
1k" hv~ : 1; : 4 i . ~ ni~ . ~ w S* 1' ; u t: ~ c k [ inc : c t~ i
F igu re 6 . 7 I nflu enc e of angle of attac k on Y/Q ratio
F igu re 6 . 7 above demons trates that inc reas ing angle of attac k rapidly redu c es the limit valu e
of the ratio of Y/Q. Above angles of approx imately 15mrad angle of attac k , the redu c tion
reac hes a s teady s tate valu e. I t c an be s een that the Nadal limit valu e of 1. 2 repres ents a
c ons ervative, wors t c as e valu e, ac c ommodating large angles of attac k .
~ iM
. . . ~ s ~ ; ~ . ,~ ,. . : ; ~ , ~ ,> ,,.
4 8 ~ ,. . ; nee: ~ . .
. y. i~ b ,
CONF I DENTI AL
F igure 6.8 Approximation of angle of attac k ( ineffec tiv e s teering)
V ~ ith referenc e to F igure 6.8, when s teering break s down within a s mall radius c urv e, the
angle of attac k 4 ~ +, efF ec 4 iv ely bec omes equal to a func tion of the abov e parameters . F or this
reas on all the profiles s how a v ery s imilar angle of attac k in the derailment as s es s ment whic h
us es a v ery s mall c urv e radius of 20m.
The following s ec tion pres ents a number of plots illus trating the res ults des c ribed abov e in the
s ummary of the derailment s tudy.
6.2.2 Derailment Study Wheel L ift
F igure 6.9 below s hows the wheel lift c harac teris tic for all the s tudied wheel profiles when
c ombined with S4 9 rail inc lined at 120.
+ir
~ - , ~ s -.^r~ - w.
u- ` 4 9 fir; ..: h~ 5 ~ .r
CONF IDENTIAL
i TR AN9iL/Jfi J~ f~ f1 4 LY ' ' Jti;
. . .
~ ; ~ ~ , ' + ~ am t+
. ~ ~ ~
. . . . . _ , . . . . g~ f " _ .
. . . . . .
Y NIN. 1 1 Y ~ i~ k~ :4 J. F
_ . . . . . _ . . . . . . . . . . . . . . . . . . _ . . . . . . . - - . . ~ - , s ac~ mwa+ . . _
. . _ ~ . . ~ as aa-. ws aaaa~
F igu re 6. 9 Wheel lift for all profiles on S4 9 1 :20 R ail
It can be s een that the lift characteris tics vary cons iderably acros s the wheel profiles . The
typically higher lift valu es of the CT3 and CR 4 000 profiles with a larger 1 5mm flange root
radii, is evident. In this cas e the SST profile is s howing s ignificant lift compared with the
s imilar NET DIN- type profile, the difference between the two being attribu ted to s mall
differences in flange root radiu s . As in all rail cas es s tu died, the MML profile derails and this is
reflected in the large wheel lift.
F igu re 6. 1 0 below s hows the wheel lift relation for the s elected wheels when ru nning on S4 9
rail inclined at 1 :4 0, an inclination which matches that of the R i 59- R 1 3 (R i 59N) embedded
rail s ection.
a' is
o. ~ ,
~ . ; , , ~ . ~ .
4 . . . ,
50
, . . ~
CONF IDENTIAL
V 'f~AM~IlNT AI~ti1 4 L Y 3 1 3
~~i..~~r.~~.~.ex.~+~.~ , . v z- - , .~wr~nv ra~aapaqwae~~~n.w s ev ~m+?t~
~.. _.... ~ms ~ :. :. ~ .-.ffr_..v u c. +aoce..ma~acev .~a
'SL R
_ __
"':V 3 ~e
.....- o, ., s ..., , . Y il6''~3 iNF ~iB'!JlIIP A~l(.:i
~:~~t~zaa.~.. i, w~s s
ru ... ~_ - x r. .+
F igu re 6.1 0 Wheel lift for all profiles on S4 9 1 :4 0 R ail
With reference to the abov e plot, the MML profile extends to ov er 20mm lift du e to derailment,
therefore the y- axis has been rs - s caled to allow a clear v iew of the profiles which do not
derail.
It can be s een that the DIN- type profiles s how almos t zero lift, with the CT3 profile als o
s howing a low lev el of wheel lift with a v alu e of approximately 3 mm. This compares to 5mm
for the CR 4 000 profile. When compared to F igu re 6.9, the abov e plot illu s trates the influ ence
of rail inclination on the flange climb (wheel lift) , behav iou r. As identified prev iou s ly, a high
contact angle prev ents wheel lift continu ing and promoting derailment.
When ins talled inclined at 1 :4 0, it is lik ely that the S4 9 profile wou ld be combined with
R i 59- R 1 3 (effectiv e 1 :4 0 rail head inclination) , s treet ru nning groov ed rail. F igu re 6.1 1 below
illu s trates the wheel lift experienced by the wheel profiles when ru nning on this rail s ection.
r
- M ~~..~, :.:.
51 E , , ~, n~,
.., , ~ .
CONF IDE NTIAL
~@TR /1P 1SI~t~T AtJALY~I~
,m.~.
e+s ~~s v
..-~c*s rava~+
Sg7 P rtgjg'}8
..--~--~. 55t ~dC.R td'SVtw~
4Vheai L" aft Wheei~~t i - R i5~9 X13
F igure 6.11 Wheel lift for all profiles on R i 59-R 13 Grooved R ail
A s imilar trend as s een with S49 rail ex is ts when running R i 59-R 13, in that the DIN type
profiles of SST and NET, s how low lift values , whils t the 13mm flange root radius profiles of
the CT3 and CR 4000 s how higher values . Again the MML profile res ults in a s imulated
derailment.
The wheel lift plots pres ented in this s ection are typical of the res ults acros s the s ix rail
profiles s elected. Generally, other than where highlighted, the DIN-type profiles s how lower
wheel lift values . The CT3 and CR 4000 profiles s how generally higher lift, with the CT3 being
lower than the CR 4000. In ail cas es high contact angles were maintained, with Y/Q ratios of
between 1 and 1.2 .
6.2 .3 Derailment Study Contact Angle
The following s ection pres ents the contact angles for the cas es of S49 rail inclined at 1:40
and R i 59-R 13 grooved rail. Thes e corres pond to the wheel lift plots pres ented in
Section 6.2 .2 above.
~r~~
+r
~. ~,,:~, -.u,.
52
. ,.,v,.~~~,.
CONF IDENTIAL
Vr711A?~i TR AMSIEM7 ANALYSIS
roar.~:ial?al4t~ ~ :.t eare~.+ Amy.
- -- - ~etntrt. F R ~a3Y+ Tw- ldtSO'k.~a4o
Contact ~~~~~~~- 518 1;40
F igure 6.12 Contact angle for all profiles an S~9 R ail at 1:40
figure 6.12 above s hows a typical contact angle plot, NET and SST s how a cons is tently high
contact angle, with the s mall wheel lift as s is ting in maintaining this cons tant relation. The high
contact angle is provided by the us e of a 76 flange face angle.
The CT3 profiles can be s een to have a more non-linear contact angle relations hip than the
s imilar CR 4000 wheel, the CT3 profile s howing an increas e in contact angle as the wheel
begins to lift. It is this effect which maintains derailment protection.
Similarly the contact angle is plotted in F igure 6.13 below for all wheels when running on
R i 59- R 13 grooved rail s ection.
~~
+ r
- ~ ~-- -- M -- = 53 ., , ., a.,
, , .:~ v., i~~ , :
CONF IDENTIAL
#l~~ii1' ,,.~
r
F tanr~~s G c rna+ G t Angie Ms pt 1~= ~ft! ! ~4 rf 3
F igu re 6.13 Contac t angle for all profiles on R i 59-R 13 grooved rail
The above figu re demons trated the c ons is tent c ontac t angle s een for all profiles other than
MML,when ru nning on R i 59-R 13 grooved rail. Again,the higher flange fac e angle of the
DIN-type profiles of SST and NET is evident.
The c ontac t angle plots pres ented in this s ec tion are typic al of thos e fou nd ac ros s the
s elec ted rail profiles ,where a high c ontac t angle offers good derailment protec tion even at
higher valu es of wheel lift.
6.2.4 Derailment Stu dy Angle of Attac k
As dis c u s s ed in Sec tion 6.2.1,at the very low c u rve radiu s u s ed in the derailment s tu dy
( --20m),the angle of attac k generated between the wheel and rail, effec tively bec omes a
fu nc tion of the vehic le c onfigu ration and trac k geometry,this c an be s een in the example
angle of attac k plot of F igu re 6.14 below,for the s elec ted wheels ru nning on R i 59-R 13
grooved rail.
CONF IDENTIAL
r
is ~
b`~~,~~e5~~.ry
..:.: r'.1: .
TR ANSI~Mi' AMALY~i~S
a
~, ~
w~: ww
. ., .._ .._ ., .. i....arar..; iP ih~t+ .rp.ww
._ _ ...... *lww+ ~YMdixnee
~[li}F i S}~ I~~OCk ~tYf'l1f~S+ 3l ~ ~l~~ f
F igure 6.14 Angle of attac k for all profiles on R i 59-R 13 grooved rail
It c an be s een from the above figure that the angle of attac k (AoA) is almos t identic al ac ros s
the s elec ted wheal profiles , at 65mrads . Cons idering this res ult, no further plots will be
pres ented for AoA.
6.2 .5 Derailment Study Y/Q R atio
As des c ribed in Sec tion 6.2 .1, the ratio of lateral to vertic al forc e (Y/Q)> provides an es timate
of the proximity of a wheel and rail profile pair to derailment. The derailment c as e c reated is
des igned to promote high Y/Q values but whils t Y/Q is a good indic ator of derailment ris k the
q uotient does not take direc t ac c ount of the true c ontac t angle generated between a wheel
and rail. It is believed that wheel lift values and c ontac t angle provide a c learer view of the
proximity to derailment. This opinion is evidenc ed by the fac t that the MML profile c an be
s een to derail at a s ignific antly lower Y/Q value than s ome of the other profiles s tudied, even
when taking ac c ount of its lower flange fac e angle of 68 .
F igure 6.15 s hows a typic al Y/Q relations hip for the s elec ted profiles running on R i 59-R 13
grooved rail.
~~M
CONF IDENTIAL
V~i1A~4~ii TF iANS1ENT A#VAIYSIS
. ~. . . . . . . . , . . CSd ~tti. "iYWW~ f
Y~Q R atto Whee3~et 1 - itiSS r13
F ig u re 6. 15 Ang le of attac k for all profiles on R i 59-R 13 g rooved rail
The Y/Q ratio is bas ed on the s u m of the vertic al and lateral loads on the wheel and will
therefore be effec ted by fac tors inc lu ding the effec tive c onic ity of the wheel, c ontac t ang le and
flang eway c learanc e. The ability of a wheel to res is t hig h Y/Q levels depends larg ely on
flang e g eometry e. g . the flang e fac e ang le bu t als o the res u ltant c ontac t ang les and wheel lift
c harac teris tic . In the c as e pres ented above 1. 1 to 1. 3 is a typic al Y/Q valu e bu t it c an be s een
that the MML profile derails when operating at a Y/Q valu e of 0. 95.
With the above Y/Q plot being typic al of the wheel profile behaviou r, no fu rther plots will be
pres ented bas ed on the Y/Q ratio.
It c an be c onc lu ded that with the exc eption of the MML profiles , the flang e g eometries of the
other fou r profiles were effec tive in res is ting derailment for Y/Q valu es of in exc es s of 1. 3.
The derailment s tu dy has demons trated that, with the exc eption of the MML profile, the
remaining fou r profiles of SST, NET, CR 4000 and CT3 profiles all provide g ood res is tanc e to
flang e c limb derailment, with any of thes e profiles being s u itable for s elec tion beyond the
s c ope of derailment performanc e.
The SST and NET DIN-type profiles offered the mos t c ons is tent performanc e bu t not of
s ig nific ant marg in to warrant s elec ting thes e profiles on the bas is of derailment protec tion
alone.
~~~ ~
+i+
St. : ~ . eyiil{ Ir~~
CONF IDENTIAL
The final s elec tion of wheel and rail c ombinations will be bas ed on the res u lts pres ented for
both derailment protec tion and thos e des c ribed in the following s ec tion deriv ed from the
c u rv ing s tu dy.
~ n~ ~ ~ , ~ ,~ ~ ~ ~ :i
ate, ~ . . ~ . , . ~ w, ; ~ .
57 ~ ~ . . ~ , ~ ~ r, ~ .
. , ~ .
CONF IDENTIAL
7, Dynamic Simulation Curving Study
Us ing a s imilar philos ophy to the derailment s tudy, a VAMP IR E track file was developed in
order to as s es s the lik ely level of wheel and rail profile wear, together with other k ey
wheel- rail performance indicators .
7,1 Curving Study Overview
The primary factors analys ed included Tgamma (Ty), a wheel- rail wear index
(s ee Section 7. 2 . 1 ), contact s tres s and angle of attack .
F igure 7. 1 below s hows the VAMP IR E track file which was developed for the purpos e of the
curving s tudy.
~J TitA~~ ~C~TA,~1 ~6
P 4 tb' C
' ~r4 Adt~l~ I~N u4
F igure 7. 1 Curving s tudy track file
As in the derailment s tudy, the y- axis of the above figure is combined to indicate vehicle
s peed in m/s and track curvature in k m" ' . The x- axis repres ents the track dis tance, in m, and
is arranged s tarting with the larges t radius of 1 000m, reducing to the s malles t of 2 0m at a
track dis tance of 1 2 50m. There is no applied cant in the curves (typical of embedded s treet
rail), therefore the vehicle s peed is varied to maintain a cons tant 50mm cant deficiency. A
s ummary of the above plot is tabulated in Table 7. 1 below to aid interpretation of the res ults .
it
. ; ~. ~. ,ww,. ~ ~.
58 . ,,. ; a~, ~.
CON F IDEN TIAL
Track Dis tance
/m
Curve R adius
/m
Vehicle Speed
/ms ' (kph)
50-150 1000 18 . 1 (65)
200-300 8 00 16. 2 (58 )
350-450 600 14. 0 (50)
500-600 400 11. 4(41)
650-750 200 8 . 1 (29)
8 00-900 100 5. 7 (21)
950-1050 75 4. 9 (18 )
1100-1200 50 4. 0 (15)
1250-1450 20 2. 7 (10)
Table 7. 1 Curving Study Tabulated Track Data
As des cribed above the curving s tudy aims to as s es s the likely wear between wheel and rail
combinations in their new condition. P rofile combinations will of cours e wear from their new
s tate to an 'average' of the wheel and rail profiles . The purpos e of the s tudy is to s elect a
number of profile combinations that in the new condition will be s table in terms of wear rate
and als o offer good derailment protection.
I t is difficult to predict the worn profile s hape but by ens uring compatibility in the new
condition, prevents problems with high initial wear rates of wheels and rails and will generally
provide more s table profiles in terms of s hape change. This is often es pecially true in the cas e
of light rail s ys tems where wheel turning is carried out on a regular bas is . G enerally, wheel
and rail profiles wear to become more conformal res ulting in a larger contact area for a given
force and therefore reducing Ty and contact s tres s .
The curving s imulations include output meas ures which will provide the information to s elect
compatible profile pairs , the us e of a curvature as s ault cours e allows for a cons is tent platform
to do this , is olating features s uch as track irregularities and varying cant deficiency which can
complicate the analys is .
Simulation output is calculated in the s teady-s tate s ections for each curvature, the average
value of this 100m s ection is recorded and us ed in the analys is .
A full des cription of the output meas ures and their relevance to profile s election is included in
the s imulation res ults s ection.
~s
-~ ~ 59
. 4 , , ~. : ~; , , .
CONF I DENTI AL
7.2 Curving Study R es ults
The c urving s tudy involved s ub j ec ting the vehic le and wheel- rail c omb inations to the
VAMP IR E trac k file and s imulation s c enario des c rib ed ab ove. R es ults and c ommentary are
pres ented for Tgamma (Ty ) , c ontac t s tres s and angle of attac k , for typic al c as es .
7.2.1 Curving Study Tgamma
The wear of a wheel and rail c an b e expres s ed as a func tion of the energy dis s ipated (or work
done) in the c ontac t patc h, therefore a wheel whic h is in c ontac t with the rail at a high level of
lateral forc e (work ) and at high levels of c reepage (rate of work ) will experienc e high levels of
wear. Experiments have b een c arried out, originally b y B ritis h R ail (B R ) R es earc h ~ 6 ~ and later
b y Lewis et al ~ ' ~ , whic h have s hown that the rate of wear b etween finro b odies moving relative
to eac h other is related to the energy dis s ipation in the c ontac t patc h. The energy dis s ipation
is expres s ed as the produc t of the c reep forc e and c reepage Tgamma (Ty) and has units of
J/m (or N ) , energy per unit dis tanc e rolled. Ty c an b e expres s ed mathematic ally as the
vec tor s um of the produc t of the c reepage and c reep forc e in the lateral and longitudinal
direc tions , s ee Equation 7.7.
Ty = ~ Tyyy
+TXyX~
(7.1 )
The wear rate of rails c an b e des c rib ed as a los s in c ros s- s ec tional area due to the pas s age
of a k nown numb er of vehic le axles . Lab oratory tes ts undertak en b y B R R es earc h res ulted in
an empiric al relations hip b etween wear rate (mm2/1 000 axles ) and Ty , as s hown in
F igure 7.1 . This empiric al relations hip does not tak e in to ac c ount the effec ts of c ontainments
(lub ric ation, environmental) that might b e pres ent on the rail b ut repres ents a ` wors t wear'
s ituation.
r
wir
- ~ 6 0
ry ~ ~ lti
CON F IDEN TIAL
0.50
0.45
..., 0.40
N
m
~ ~..3.~
O
O 0.3~
N
~ 0.25
m 0.20
R
~ 0.15
~a
~ 0.10
0.05
0.00
0 50 100 150 200 250 300
Wear Number, Tgamma (J/m)
F igure 7.2 Ty and wear relations hip
As pan be s een from F igure 7.2 above, that the relations hip between the gate of metal
removal (wear) and the energy dis s ipated in the contact patch is non-linear. I t has been found
that the relations hip is a two s tage proces s , at iow values of Ty the rate of wear increas e up
to a Ty value of 100 J/m, in ` mild' wear regime. After a region of cons tant ' mild' wear, the
` s evere' wear regime initiates at a Ty value of 200 J/m and continues as a linear relations hip.
Bas ed on this relations hip it would clearly be an advantage to control the curving forces and
remain in the mild wear region of the curve. Us ing this relations hip it is pos s ible to as s es s the
lik ely wear occurring at the wheel-rail interface.
I ncluded below is a s ummary of the Ty res ults for grooved and non-grooved rail s ections .
The plots of Ty included in this s ection illus trate the total Ty (total work done) by both the
tread and flange contacts .
All tyre profiles generate moderate Ty values on the tread (<100J/m) when combined with
BS 80A rail s ection inclined at 1:40. The NET tyre profile produces s ignificantly higher values
of Ty (approximately +30%) at 20m radius curve than the other tyre profiles . I nclining the
BS 80A rail s ection at 1:20 generally produces lower Ty values acros s all wheels compared
to BS 80A inclined at 1:40, with the exception of the DI N-type profiles at curve radii below
200m. The DI N-type profiles move directly into flange contact at a curve radius of 1000m with
a Ty value of 70J/m with the SST tyre profile cons is tently higher (approximately +20%)
*~
-, .~, ~ .
CONF I DENTI AL
acros s the entire cu rv atu re range. MML and CT3 generate the lowes t Ty on BS 80A inclined
at 1 : 2 0.
When combined with BS 1 1 3A, inclined at both 1 : 2 0 and 1 : 4 0, all tyre profiles generate
s ignificantly higher Ty v alu es at high- to mid-radiu s cu rv es (1 000m to 2 00m). The highes t
v alu es of Ty were s imu lated u s ing the DIN- type profiles (SST and NET), whereas the CT3
tyre profile generated the lowes t v alu es of Ty acros s the whole cu rv atu re range.
F igu re 7.3 below s hows a comparis on of the av erage Ty res u lts for CT3 and SST tyre
profiles on BS 80A and BS 1 1 3A rail s ection inclined at 1 : 2 0. It can be clearly s een from
F igu re 7.3 that the CT3 tyre profile remains in the ` mild' wear regime u ntil a cu rv e radiu s of
approximately 1 00m where the Ty v alu e s teadily increas e to a maximu m of approximately
1 760J/m at 2 0m radiu s cu rv e for both rail s ections . The SST DIN- type profile on the other
hand remains in the ` s ev ere' wear regime for the entire cu rv atu re range when combined with
the BS 1 1 3A rail s ection.
aVV
4 50
4 00
350
300
2 50
2 00
1 50
1 00
50
T 4
1 000 800 600 4 00 2 00 2 00 1 50 1 00 50 0
Cu rv e R adiu s (m)
F igu re 7.3 Comparis on of Ty generated u s ing CT3 and SST (DIN- type) tyre profile on
BS 80A and BS 1 1 3A rail s ections (inclined at 1 : 2 0)
v wu
3500
3000
2 500 E
R
2 00 ~
R
1 500 ~
1 000
500
0
When u s ing S4 9 rail s ection, inclined at 1 : 2 0 and 1 : 4 0, s imilar v alu es of Ty to thos e
s imu lated for the BS 1 1 3A rail s ection are generated for all tyre profiles . G enerally thes e
v alu es are higher than the v alu es s imu lated when u s ing BS 80A rail s ection with the
exception of the CT3 tyre profile which produ ces the ~o+ n! ect Ty v alu es on 54 9. G enera! ! ;
~,,.,.,~,t ,
CONF IDENTIAL
lower values of Ty were ob tained on S49 rail s ec tion with a 1:40 rail inc lination rather than
1:2 0.
figure 7. 4 b elow s hows a c omparis on of the average Ty res ults for CT3 and SST tyre
profiles on BS 80A and S49 rail s ec tion inc lined at 1:40. Thes e two profiles s how the
extremes of the res ults ob tained, with CT3 produc ing the lowes t and SST the highes t Ty
values . Again, as in the previous figure, the CT3 profile remains in the ' mild' wear regime
down to a c urve radius of 100m. As the c urve radius reduc es from 100m the CT3 profile
moves in to flange c ontac t and the Ty values s teadily inc reas e to a maximum of 1500)/m on
S49 and 1712 J/m on BS 80A rail. The SST profile produc es s ignific antly higher Ty on S49
rail s ec tion through the entire c urvature range with a maximum of approximately 3500)/m at a
c urve radius of 2 0m. Thes e res ults are c ons is tent writh the philos ophy that higher c onic ity
profiles generate higher values ~f Ty .
-- ~uu
-~ CT3 - BS 80A (1:40)
~- CT3 - S49 {1:40)
450
~- SST - BS SOA (1:40) 4~
~ SST - S49 (1:40)
350
300
J 2 50
Severe Wear v~
----------- /---- 2 00
150
Mild Wear
---- ------------
100
50
0
4WU
3500
3000
2 500 ~
a
2 000 ~
E
R
1500 ~
1000
500
0
1000 800 600 400 2 00
2 00 150 100 50 0
Curve R adius (m)
F igure 7. 4 Comparis on of Ty generated us ing CT3 and SST (DIN-type) tyre profile on
BS 80A and S49 rail s ec tions (inc lined at 1:40)
Generally on R i 59-R 10 all profiles s how low (<100J/m) to moderate Ty values (<500J/m)
down to a c urve radius of 50m. As the c urve radii tightens Ty inc reas e further into the
' s evere' wear regime, with the DIN-type profiles produc ing 2 0% higher Ty values due to
b eing in two-point flange c ontac t.
: '
. . . , ~. . :. . ~. . _ r. . ~
63
CONF IDENTIAL
in comparis on, on R i 59-R 13 all tyre profiles generate higher values of Ty on high- to
mid-radius curves than R i 59-R 10, with the ex ception of CT3 which is generally the s ame. At
s maller radius curves the DIN-type profiles generate 25% higher Ty values than the other
tyre profiles . Although the Ty values generated us ing R i 59-R 13 are high compared to
R i 59-R 10 the R 13 grooved rail s ection produces lower contact s tres s es as des cribed in
S ection 7 .2.2.
F igure 7 .5 below s hows a comparis on of the average Ty res ults for CT3 and S S T tyre
profiles on R i 59-R 10 and R i 59-R 13 grooved rail s ections . It can be s een that the DIN-type
profile (S S T) produces s ignificantly higher Ty values on R i 59-R 13, moving into the ` s evere'
wear regime at a curve radius of 400m, compared to a radius of approx imately 100m on
R i 59-R 10. This demons trates the incompatibility of the S S T flange root radius (13mm) with
the 13mm gauge corner radius of the R i 59~R 13 rail s ection. it can als o be s een that the CT3
profile generates s imilar levels of Ty for both grooved rail s ections down to a curve radii of
20m.
S oo
-f CT3 - R i 59-R 10
-;- CT3 - R i 59-R 13
450
-~ S S T - R i 59-R 10
~ 400
~ S S T - R i 59-R i 3
i 350
300
r~
250
S evere Wear
------ ----- - ----- 200
150
Mild Wear
~--- --- -- 100
/; 50
_ I~ ~
4u~~
3500
3000
2500 ~
R
Z~a4 ~
1500 ~
1000
500
0
1(104 800 600 400 200 200 150 100 50 0
Curve R adius (m)
F igure 7 .5 Comparis on of Ty generated us ing CT3 and S S T (DIN-type) tyre profile on
R i 59-R 10 and R i 59-R 13 rail s ections
7 02.2 Curving S tudy Contact S tres s
The contact s tres s is calculated from the normal load (~ and the area of the contact patch,
derived from the elliptical dimens ions of the wheel-rail geometry (a, b), as s hown in
Equation 7 .2. The s iz e of the contact patch and therefore the intens ity of the s tres s in the
4y
a,~ ~.. ry .
64 ,` ,.
CONF IDENTIAL
wheel and rail depend on a number of fac tors inc luding wheel diameter, wheel load and the
trans vers e profile of the wheel and rail.
6m~ =1. 5 ~
( 7 . 2 )
Contac t s tres s is an important fac tor when c ons idering the lik ely oc c urrenc e of rolling c ontac t
fatig ue ( R CF ) and other forms of wheel- rail s urfac e damag e. H is toric ally, in the railway
indus try the c onc ept of material s hak edown ~8 ~ has been us ed to des c ribe the c onditions lik ely
to res ult in the development of s urfac e damag e and R CF .
F or a g iven material, a s hak edown diag ram c an be plotted whic h defines the relations hip
between the wheel- rail c ontac t s tres s and the trac tion c oeffic ient ( or s hear forc e c oeffic ient) .
Tes ts on typic al rail s teels has res ulted in the development of a s hak edown limit whic h
defines the c ombinations of c ontac t s tres s and s hear forc e c oeffic ient whic h are lik ely to
ac c umulate plas tic s train and ultimately rail s urfac e damag e. F ig ure 7 . 6 s hows a typic al
s hak edown diag ram previous ly us ed for rail s teels . At lo~nr s hear forc e c oeffic ients and low
c ontac t s tres s es (<- 1000MN/ r? iz ) the rail s teel material behaves elas tic ally, as the c ontac t
s tres s inc reas es , up to approximately ~2 100MN/ m2 , the elas tic s hak edown limit is reac hed.
Below whic h plas tic filow c an tak e plac e during the firs t few c yc les , but plas tic deformation,
res idual s tres s es and s train hardening may enable the s truc ture to reac h ac yc lic - elas tic
res pons e, k nown as ` elas tic s hak edown'. Above this limit, plas tic deformation tak es plac e with
eac h loading c yc le. As the s hear forc e c oeffic ient inc reas es the thres hold limits dec reas e as
s hown in F ig ure 7 . 6.
T
2 500 - i Subs urfac e flow
N
E 2 000
R epeated plas tic flow
z
~ Elas tic s hak edown
1500
Surfac e flow
in
- -
V
C ~~~
O
U
$~
Elas tic
0
0 0. 05 0. 1 0. 15 0. 2 0. 2 5 0. 3 0. 3 5 0. 4
Trac tion Coeffic ient
F ig ure 7 . 6 Theoretic al s hak edown diag ram for typic al rail s teels
ii
. ~ . . . . ~~. . . ~. ~ 65 ~~N,- ~~. , ,.
~: ry,rlit i~
~. ~ii~ - ~
CONF IDENTIAL
Typic ally, as the wheels et moves laterally into flange c ontac t the c ontac t s tres s will inc reas e
as c ontac t oc c urs on parts of the trans vers e wheel and rail profiles whic h have s maller radii
and therefore s maller c ontac t areas . A dditionally flange c ontac t res ults in higher c ontac t
angles , s ignific antly inc reas ing the normal load (l~ and henc e c ontac t s tres s . I f two point
c ontac t oc c urs then the peak c ontac t s tres s tends to be lower than s ingle point c ontac t due to
the tread reac ting a proportion of the vertic al load, but this depends on the c onformality
between the flange root and gauge c orner radii. However Ty values tend to be greater when
in two- point c ontac t.
Due to the differenc es at tread and flange c ontac ts des c ribed above, the plots pres ented in
the following s ec tions will s eparate the tread and flange c ontac t s tres s res ults .
On BS 113A rail s ec tion all tyre profiles move into flange c ontac t at 1000m radius c urve. This
generates two- point c ontac t res ulting in both tread and flange c ontac t s tres s es , with the
flange s howing higher values . The highes t flange c ontac t s tres s values were obs erved with
the rail inc lined at 120 when c ombined with the DI N- type profiles , as s hown in F igure 7 .7 .
CT3 generated the lowes t c ontac t s tres s on both the tread and the flange with values
remaining below 1500MN/m2 (elas tic region of s hakedown diagram, s ee F igure 7 .6) for the
entire c urvature range.
- i-- - CT3 - Tread
- ~ CT3 - F lange
~ SST - Tread
- ~SST - F lange
~~o
6000
5000 ~
E
z
ago ~
d
3000 y
w
V
A
2000 C
0
V
1~0
1000 800 600 400 200 0
Curve R adius (m)
F igure 7 .7 Comparis on of average tread c ontac t s tres s for CT3 and SST tyre profiles on
BS 113A rail s ec tion inc lined at 1:20
When running on S49 rail s ec tion the CT3 tyre profile generates the lowes t c ontac t s tres s in
flange c ontac t with the rail inc lined at both 1:20 and 1:40, whereas the NET tyre profile
generates the lowes t tread c ontac t s tres s es with the rail inc lined at 120. On BS 80A rail
s ec tion the CT3 profile generates the highes t c ontac t s tres s es on the tread with the
DI N- type profiles produc ing s ignific antly lower s tres s values . With a rail inc lination of 1:40
~~~ ;
~iy
~., ~~~
CONF I DENTI A L
two- point contact does not occur until a curve radius of approximately 100m for SST and
CR 4000 and 20m for the other s elected profiles .
F igure 7. 8 below s hows a comparis on of the tread contract s tres s es generated when
combining the CT3 and SST tyre profiles with the BS 80A and S49 rail s ection, inclined at
1: 40. I t can be s een that the contact s tres s predicted when us ing both the CT3 and SST tyre
profiles on S49 rail s ection remain below 1500MN/m2 on the tread for the entire curvature
range, within the elas tic region of the s hakedown diagram at low traction coefficients (s ee
F igure 7. 6 ) . On BS 80A rail the CT3 profile generates s ignificantly higher s tres s values
(-- max. 3200MN/mz) whereas the DI N- type SST profile remains below 1500MN/mz, s imilar to
that predicted for the S49 rail s ection.
- - s - CT3 - BS 80A (1: 40)
- F - CT3 - S49 (1: 40)
- CI E- SST- BS 80A (1: 40)
~I ~ SST - S49 (1: 40)
1000 800 6 00 400 200 0
Curve R adius (m)
as oo
4000
3500
N
t
3000 8
a
2500 ~
w
N
m
2~~~ ~
h
1500 w
c
0
1000 V
500
F igure 7. 8 Comparis on of average tread contact s tres s for CT3 and SST tyre profiles on
BS 80A and S49 rail s ection inclined at 1: 40
F igure 7. 9 below s hows a comparis on of the flange contact s tres s es generated when
combining the CT3 and SST tyre profiles with the BS 80A and S49 rail s ection, inclined at
1: 40. I t can be s een that the CT3 tyre profile is not in iwo- point until a curve radius of
approximately 75m on both BS 80A and S49 rail, therefore generating zero contact s tres s in
the flange contact. When two- point contact does occurs (at a curve radius of 75m) the value
of the contact s tres s reaches a maximum of 2000 MN/m2. I n comparis on, the SST profile
moves into two- point contact immediately, at a curve radius of 1000m, res ulting in a gradual
increas e in flange contact s tres s , as the curve radii reduces , up to a maximum of 2400MN/m2.
~~y
4i r iy I ils ~
CONF I DENTI AL
-a-- Cl3 - BS SOA ( 1 : 4 0 )
-a- CT3 - S4 9 ( 1 : 4 0 ) 4 ~
~-SST - BS 80 A ( 1 : 4 0 ) 350 0
~~ SST - S4 9 ( 1 : 4 0 ) r
30 0 0 E
z
250 0 ~
~~--=iE-~: ~iE~~ a
20 0 0 N
u
1 50 0 ~
0
1 0 0 0 v
/ 50 0
R ~ ~- a- 0
1 0 0 0 80 0 60 0 4 0 0 20 0 0
Curve R adius ( m)
F ig ure 7.9 Comparis on of averag e flang e c ontac t s tres s for CT3 and SST tyre profiles on
BS 80 A and S4 9 rail s ec tion inc lined at 1 : 4 0
F ig ure 7.1 0 below s hows a c omparis on of the averag e c ontac t s tres s for the CT3 and NET
tyre profiles on R i 59-R 1 0 g rooved rail. This is typic al of the res ults obtained for the R i 59-R 1 0
rail s ec tion, with all profiles ( with the exc eption of CT3) g enerating a c ontac t s tres s in the
reg ion of 1 70 0 -21 0 0 MN/m2. The CT3 profile g enerates s ig nific antly hig her s tres s es at hig h- to
mid-radius c urves due to the differenc es in the c ontac t area g enerated when matc hing the
1 5mm flang e root radius of the CT3 profile with the 1 0 mm g aug e c orner radius of the
R i 59-R 1 0 rail s ec tion. I t c an be s een that both NET and CT3 tyre profiles only c ontac t the
flang e at a c urve radius of 20 m, res ulting in two-point c ontac t.
s
.~ y., : .... " .,
68
, N iu
CONF I DENTI AL
4VW
-a- CT3 -Tread
'- -~ CT3 -F lange
3500
--~_
~~-------------ice\ ~ NET -Tread
3000
--t- NET -F lange ~: ,
\, ~ ~
2500 Z
2000 N
N
1500 ~
A
c
1000 V
500
~-- 0
1000 800 600 400 200 0
Curve R adius (m)
F igurs 7.10 Comparis on ofi contact s tres s generated us ing CT3 and NET tyre profiles on
R i 59-R 10 grooved rail s ection
F igure 7.11 below s hows a comparis on. of the average contact s tres s for the CT3 and NET
tyre profiles on R i 59-i~13 grooved rail. On R i- 59-R 13 all the DIN-type profiles move into
flange contact at 1000m radius curve, s ignificantly earlier than on R i 59-R 10 rail. This
generates two-point contact res ulting in lower contact s tres s es on the tread but higher
s tres s es on the flange. The CT3 profile remains in one-point contact until 20m radius curve
where it can be s een that the s tres s es on the tread reduce and increas e on the flange as the
load is trans ferred from the tread to the flange as two-point occurs . The CT3 profile produces
s ignificantly lower contact s tres s es on R i 59-R 13 than R i 59-R 10 with the s tres s remaining
below 1500MN/m2 until two-point contact occurs at a curve radius of approximately 20m.
~~~ ~
t}M
, , .~ ~; .
CONF IDENTIAL
s CT3 -Tread
FCT3 - Flange
t NET -Tread
iNET - Flange
~~~
~~-----f
//~
_- ~_
__
Final R eport
4000
3500
3000
N
2500 Z
200 H
H
~~`p
R
C
~000 v
500
0
1000 800 600 400 200 0
Curv e R adius {m)
Figure 7 . 11 Comparis on of c ontac t s tres s generated us ing CT3 and NET tyre profiles on
R i 59-R ' ! 3 groov ed rail s ec tion
7 . 2. 3 Curv ing Study Angle of Attac k
Us ing the c urv ature as s ault c ours e the angle of attac k of the wheels et was als o output and
plotted agains t c urv e radius . As des c rib ed in Sec tion 6. 2. 1, at low radius c urv es the angle of
attac k es s entially b ec omes the s ame for all tyre profiles due to the fac t that at v ery low radii
s teering b reaks down. This effec t has als o b een s een in the res ults of the c urv ing s tudy,
where all profiles generate the s ame angle of attac k on c urv es with a radius of les s than
200m. Therefore the res ults pres ented in this s ec tion will only c ons ider c urv es with radii of
greater than 200m.
When c omb ined with BS 113A, BS 80A ( 1: 40), S49 ( 1: 20) and R i 59-R 10 all tyre profiles
s how no s ignific ant differenc e in the s imulated angle of attac k. On BS 80A ( 1: 20), S49 ( 1: 40)
and R i 59-R 13 rail s ec tions the CT3 profile generates an angle of attac k c los e to zero,
indic ating nearly radial s teering, an c urv es with a radius greater than 600m. On c urv es with
radius in the range of 600m-200m the angle of attac k for the CT3 profile remains
approximately 2mrad b elow that produc ed b y the other profiles .
Thes e differenc es in angle of attac k c an b e ac c ounted for b y the differenc e in the tread and
flange-root radius of the profiles , with the CT3 b eing b etter matc hed to the rail s ec tions s hown
in Figure 7 . 12. This res ults in the CT3 profile generating a greater rolling radius differenc e,
and c onic ity, for a giv en wheels et lateral dis plac ement improv ing the s teering performanc e in
high- to mid-radius c urv es .
iir
. . ~. . ~. . . . ~. , :--~
7 0
CONFIDENTIAL
0
- 1
- 2
- 3
- 4 a
R
- 5 E
Q
- 6
a
- 7
- 8
- 9
- 1 0
Curve R adius (m)
F ig ure 7.1 2 Comparis on of averag e ang le of attac k for CT3 and SST tyre profile on various
rail s ec tions
7.~ Curving Stuc ly s ummary
I n s ummary, s ome hig h Ty values have been obs erved, but thes e are g enerally c ombined
with low c ontac t s tres s es res ulting in the unlik ely g eneration of F 2CF . I V on-g rooved rail
s ec tions s how hig her Ty values due to the head radius , produc ing hig her c onic ity.
Differenc es in flang e Ty values are attributed to the c onformality of the flang e root and
g aug e c orner and tread Ty to head radii and inc lination/c one ang le of the wheel- rail
c ombinations .
O verall, the CT3 tyre profile g enerated the lowes t Ty values on the maj ority of the rail
s ec tions , es pec ially S49 (inc lined at 1 :40) non-g rooved and R i 5 9- R 1 3 g rooved rail s ec tions .
Lower values of Ty were als o obs erved when us ing SST DI N- type profile on R i 5 9- R 1 0 and
BS 80A rail s ec tions .
The CT3 profile g enerated lower c ontac t s tres s es on S49 and R i 5 9- R 1 3 rail s ec tions with the
DI N- type profiles produc ing lower c ontac t s tres s es on the BS 80A and R i 5 9- R 1 0 rail
s ec tions .
Lower values of ang le of attac k were obs erved when us ing the CT3 tyre profile on
BS 80A (1 :20), S49 (1 :40) and R i 5 9- R 1 3 down to a c urve radius of 200m, due to the inc reas e
in rolling radius differenc e that is g enerated with thes e c ombinations .
**
,,,..~~~
CO NF I DENTI AL
8. Dynamic Simulation L ateral Stab ility
The eq uiv alent conicity can defined from rolling radius difference g enerated b y a wheel and
rail profile pair. Hig h lev els of conicity enab le a v ehicle with conv entional ax les to s teer around
s maller radius curv es without flang e contact b ut can lead to v ehicle ins tab ility caus ed b y the
wheels et k inematic hunting motion b ecoming uns tab le.
The point at which a v ehicle b ecomes uns tab le is v ery much dependant on the v ehicle
parameters , s uch as plan v iew s us pens ion s tiffnes s , damping lev els and b ody/ b og ie inertias .
As the modelling work us es a g eneric v ehicle model, the lik ely s tab ility b ehav iour of the
profiles would only apply to the v ehicle b eing modelled and therefore dynamic s imulation
res ults cov ering v ehicle lateral s tab ility will not b e pres ented in this report, a des criptiv e
s ummary of the s imulation res ults is prov ided b elow.
8.1 Conv entional Ax les
All the wheel and rail profiles s tudied were s ub j ected to s imulations to inv es tig ate lateral
s tab ility (hunting ) b ehav iour when running under the conv entional ax les of the g eneric v ehicle
model. All comb inations prov ided s tab le running at up to 50mph and this is deemed to b e
acceptab le. As dis cus s ed ab ov e, due to the influence of a particular v ehicle config uration, the
relativ e s tab ility performance of a particular profile/ v ehicle comb ination s hould b e inferred
from the eq uiv alent conicity v alues pres ented in Section 3 .1.2 . V ehicles with relativ ely low
plan v iew s us pens ion s tiffnes s or s ys tems with hig her lines peeds s hould pay particular
attention to thes e v alues and cons ider the us e of lower conicity wheel and rail comb inations if
it is lik ely that lateral s tab ility may b ecome an is s ue.
8.2 Independently R otating Wheels (IR Ws )
The g eneric v ehicle model us ed in the s imulation work includes IR Ws on the centre
trailer- truck s ection. Throug h the remov al of the rotational coupling (the ax le) b etween the two
wheels , the res toring forces g enerated b y wheel- rail conicity are no long er pres ent, therefore
IR Ws ex hib it different running b ehav iour to conv entional ax les . In curv ed track the wheels et
will tend to mov e rapidly towards flang e contact, on s traig ht track there tends to b e a g eneral
lack of g uidance, leading to either the wheels et `wandering ' from flang e to flang e or running in
an offs et pos ition. A b enefit howev er, of IR Ws , is that the remov al of the ax le eliminates the
prob lem of k inematic hunting , as s ociated with hig h conicity and conv entional ax les .
The lack of g uidance in the running b ehav iour of IR Ws can caus e wheel wear is s ues ,
particularly increas ed filang e wear and can als o mak e the wheel more s us ceptib le io
~~~ t
+r
.. ....: ~.; w
.,: : a
72
..ch~t~.r
CONF IDENTIAL
F inal R eport
impacting (and derailing) at partly clos ed s witch blades , as the wheels et is more lik ely to be
running offs et from the track centre, clos e to the s tock rail, than a conv entional ax le.
I t is us ual practice to apply the s ame wheel tyre profile s hape to all wheels of a v ehicle but
pos s ible gains may be achiev ed by us ing an increas ed conicity profile on I R W trailer- truck
wheels . S ignificantly increas ed conicity through the incorporation of a large flange root radius
can begin to produce as elf- cen# ring force on the wheels et induced by the grav itational forces
acting (k nown as grav itational s tiffnes s ) .
S imulations hav e been performed to as s es s any potential adv antages that could be gained by
introducing higher conicity wheels on I R W ax les .
x . 2 . 1 I R V 1 ~ ' S imulation Ov erv iew
The potential benefit of us ing higher conicity wheel profiles on I ftW ax les has been
demons trated through as s es s ment of the running behav iour of the v ehicle on s traight track .
As s es s ment through curv es has not been performed as the s mall grav itational s elf-centring
affect of increas ed conicity is not s ufficient to improv e the s teering behav iour of an I R V V but
can improv e offs et running and general poor guidance on s traight track .
The s imulation us es a s tability as s es s ment method whereby the v ehicle is s ubj ected to 1 00m
of v ertical and lateral track irregularity, after this dis tance the irregularities are held cons tant.
The lateral dis placement of the wheels et is output, to as s es s the lev el of offs et running and
lik ely lev els of res ulting flange wear.
I n order to promote and offs et running attitude (as reported by Midland Metro) a 1 mrad of
wheeis et yaw mis alignment is introduced to the v ehicle model.
Two wheel profiles hav e been s elected for comparis on agains t the s tandard DI N- type profile
currently us ed at S heffield S upertram and Midland Metro. The firs t is the current
Dock lands L ight R ailway (DL R ) , DL R 5 profile, which incorporates a 2 2 mm flange root radius ,
the s econd tes t profile being the Croydon Tramlink CT3 prototype profile, which us es a
s maller 1 5mm flange root. The DI N- type wheel profiles are typical in hav ing twapoint contact
in the flange contact zone, whils t the CT3 and particularly the DL R 5 are s ingle point contact
profiles .
The s imulations combine the abov e wheel profiles with S 49 rail s ection inclined at 1 : 40, the
S 49 generates s imilar contact conditions to the R i 59- R 1 3 groov ed rail s ection and for the
purpos e of this s tudy the res ults can be cons idered to apply to this rail s ection als o.
~ s ~
~
i is
. ~ . w. ~ . . z~ ,. r. .
73 : s ~ . . ; h~ ~ 5~ r
r ry. d0,.
CONF I DENTI AL
The s imu lations have been u s ed to as s es s any improvement in gu idance of the wheels et
which wou ld be lik ely to lead to a redu ction in flange contact and wear. This is as s es s ed
throu gh the wheels et lateral pos ition and the res u ltant Ty valu es .
8.2.2 IR W P rofiles -Offs et R u nning
F igu re 8.1 below s hows the lateral dis placement of the independent third ax le, u s ing the
DLR 5, CT3 and the SST tyre profiles when ru nning on S49 rail s ection, inclined at 1 : 40 .
._ _ ...._ _ _ _ ...._ _ _ ~ ~ _ _ _ ._ _ .. _ _ ._ w~ _ ._ .._ _ _ .r.._ _ _ .._ _ ._ ..
r, rr 4~ TP fR NSl~ NT A MiA L' ~ ' SF : ~
*..~ . ' ~ Y i~ P Il lti4wX Si
~ as t ~ :
~ ~ ~ ~ ~ ~
F igu re 8.1 IR W wheels et lateral dis placement IR W ax le
With reference to the above figu re, it can be s een that the s maller flange root profiles of the
SST ( root radiu s = 1 3mm) and CT3 ( root radiu s = 1 5mm), move into flange contact as a res u lt
of the lateral creep force generated by the 1 mrad ( approx imately 1 mm longitu dinal offs et at
the ax lebox ) wheel mis alignment. The DLR 5 profile is a pu rely s ingle point contact wheel
profile and moves approx imately 8mm laterally, moving u p the large flange root radiu s .
None of the wheel profiles demons trate a s ignificant s elf centring effect bu t it mu s t be noted
that the tread profiles are conventional and therefore the gravitational s tiffnes s effect only acts
in the flange region, where a s ignificant contact angle difference is created. This can be
ex plained in F igu re 8.2 below.
CONF IDENTIA L
per[ Gi?~i'~AGi b / 4 7 t~1 i~
~i,a$ ~F If1 d W ~ WiG~ '.::ice! 1 d~l '~.
}W 7 eg&firyp4 ~?awt'.ryp. aq~
"t
MA#1 ~lR E ~F s i
F igure 8.2 Contact angle difference
F igure 8.2 ab ove s hows the contact angle difference, for the three profiles cons idered fior
evaluation for us e an IR W's , plotted agains t lateral s hift of the wheels et. By generating a
contact angle difference b etween the left and right wheels , a net s elf-centring force is
generated from the normal forces acting at the wheels .
It can b e s een b y comparing F igure 82 with that of F igure 8.1 , that for the s maller flange root
profiles , the contact angle difference is generated at around 6mm lateral s hift, the large root
radius of the DLR S profile does not have a dis tinct flange contact point b ut in the s imulation
cas e pres ented, an equilib rium point is found at 8mm lateral s hift (approximately 20 contact
angle difference) .
The ab ove s imulations demons trate the prob lem that even with larger flange root radii, an
IR W trailer truck s b ogie will run off-centre in res pons e to very low lateral forces . This can
res ult in offs et running and res ultant flange wear.
,.._ _ .. , 7 5 4 ,..~n~,~.~
u,,,,,~
~~~,.
CONF IDENTIAL
8 . 2 . 3 IR W P rofiles F lange Wear
The res u lts pres ented in the prev iou s s ec tion highlight the problem of potential offs et ru nning
with IR W trailer tru c k s . Withou t adopting radic al profile s hapes for the IR W ax les , whic h may
c au s e problematic worn rail profile s hapes , offs et ru nning is diffic u lt to av oid.
The following plot s hows the Ty wear parameter (as des c ribed in detail in S ec tion 7 ) , for the
three profiles analys ed in the prev iou s offs et ru nning s tu dy. Higher v alu es of Tgamma
indic ate greater wear, with an ac c elerated s ev ere wear regime for Ty v alu es abov e 2 00J/m.
V~tv 1? TR ANS IENT AL'YS IS
~. ~- . . . ,
, ~. . . . _ - - ~~~~~~~~~~~~~~~r~r~e~'a. . S ~+ # e> c c + r~c ~'tan~; , : . , ~. , ~.
_ . _ . . _ . , . .
~T'lir~'e+ Ais . tgdtm~i. a` m. ~a P aEn6tany"~
_ . . . ~_ . . _ . ~. . ~
. y~` 4~A5~19~~4fdk 8 "{r~fY110. S ~` 4~f 'lMG. Y: i
~{AtVlR lft~ P loi
F igu re 8 . 3 Tgamma v alu es , IR W ax le
When c ons idering F igu re 8 . 3 abov e, it is important to note a fu ndamental differenc e in the
wheel- rail c ontac t between the S S T DIN- type profile and that of the CT3 and DLR 5 profile,
that is , the S S T profile promotes two- point c ontac t.
F rom the res u lts of the offs et ru nning s tu dy, the s ignific antly larger flange root radiu s of the
DLR 5 profile res u lted in a larger lateral s hift of the wheels et. This allows the wheel to oppos e
lateral forc e wiihou t c ontac ting the flange at high c ontac t angles . Larger c ontac t angles res u lt
in greater s pin c reepage and higher wear. In c omparis on the CT3 profile s howed a lower
lateral s hift bu t remained in s ingle point c ontac t, albeit at a s lightly higher c ontac t angle than
~~~
s r
a: ~N. ilf r~
CONF IDENTIAL
F inal R eport
the DLR 5, although not very c lear from the s c ale of F igure 8 . 3 , this res ults in s lightly higher
Ty for the CT3 profile over the DLR 5.
F igure 8 . 3 s hows that a s tandard DIN- type profile s uc h as SST, promotes two- point c ontac t
and this c aus es muc h higher Tgamma values than s een in the c as es of CT3 and DLR 5.
This res ult demons trates that for IR W wheels ets a wheel profile whic h remains in s ingle point
c ontac t is b enefic ial in terms of reduc ing flange wear. It is als o c lear that the profile does not
nec es s arily req uire a large flange root radius to ac hieve this (CT3 flange root = 15mm) and
therefore a wheel profile c ould b e des igned to operate effec tively on b oth the c onventional
and IR W ax les , eliminating the c onc ern of prob lematic worn rail profiles s hapes c aus ed b y
us ing mix ed ~ ~ ~ ~ ~ ~ ~ ~ profiles .
It is pos s ib le that further improvements c ould b e made to the running b ehaviour ofi IR AN
wheels through the des ign of anon-c onventional wheel profile b ut this would b e unlik ely to b e
c ompatib le with a c onventional ax le and would therefore req uire two profile types for a typic al
tramway vehiele. F urther des ign work for anon-c onventional IR W optimis ed profile is b eyond
the s c apL of this work .
:~*
'Ji'~iy~lil ~~
CONF IDENTIAL
9. Wheel- R ail P rofile R ecommendations
The fundamental ob j ectiv e of the s tudy pres ented in this report is to determine wheel and rail
profiles which are compatib le in manner which reduces the ris k of derailment and wheel and
rail wear. The following s ection b ring s tog ether the detailed technical analys is work of the
s ub s equent s ections and formulates awheel- rail compatib ility matrix b as ed on the analys is
finding s .
I n ev aluating the profile s election options pres ented, it is important to cons ider the
practicalities of the proj ect tas k , in that, it is not pos s ib le to s ug g es t a wheel profile which is
compatib le with the complete s et of rail profiles and inclinations currently us ed in the UK. it is
als o not pos s ib le to hav e a s ing le wheel profile which will work for ev ery s ys tem type.
Howev er, b y adopting a matrix of pos s ib le wheel and rail profile comb inations , then a wheel
and rail profile s et can b e s elected which will operate s afely on many s ys tems with an
acceptab le lev el of wear.
I t was cons idered that three final profile s ets s hould b e created; Set 1 , an optimal profile s et
for new s ys tems , Set 2 for s ys tems cons trained 1 0mm g aug e corner rails and Bet 3, a tram
' tram - train compatib le profile s et. Thes e s ets will b e fully ex plained in Section 9.2.
9.1 Selection Methodolog y
A methodolog y was required to further reduce the wheel profile comb inations from the fiv e
wheels and s ix rails s tudied in the s imulation work to form an optimal s et of profiles which will
offer compatib ility in terms of derailment protection and res is tance to hig h wear rates . The
philos ophy was to s elect a s mall numb er of comb inations of the new wheel and rail profiles
which throug h b eing inherently compatib le in the new condition would not s uffer from hig h
initial wear rates .
With the derailment analys is indicating that all profiles , with the ex ception of the MML profile
in the new condition, s hould offer g ood derailment protection, the res ults from the curv ing
analys is b ecame the focus of the methodolog y to reduce the profile comb inations . The
curv ing s tudy pres ented k ey wheel- rail performance indicators related to the compatib ility of
the s hapes of the profiles s uch as contact s tres s and Ty as well as curv ing indicators s uch
as rolling radius difference and ang le of attack .
A wheel- rail interface s ummary s heet was dev eloped for each of the s imulated wheel and rail
profile comb inations , an ex ample s ummary s heet is s hown in F ig ure 9.1 on the following
pag e.
c~~~ +
~s r
CONF I DENTI AL
R i59 R ail -NET P 3102639 Wheel
1:20 - 1:60 TAP ER 1:20 - 1:40 TAP ER
R =t4 _~ R =1a
R =tO R =13
R i59-R 10 R i59-R 33
Simulation
Out ut
C~rva R adii m
400 100 75 50 20
a -Tread M N m 2125.6 2146.6 2132.4 2136.1 1923.8
a -F lan e MN m 0.0 0.0 0.0 0.0 2228.3
T~ -Tread (N) 19.9 274.2 355.2 491.6 622.1
T -F la e N 0.0 0.0 0.0 0.0 2037.5
A -Tread de 25.23 43.60 44.40 44.40 22.43
A -F m elAe 00 0.0 0.0 0.0 7.03
Ao4 matl 4.02 13.99 17.04 23.05 50.03
~R (mm) 2.11 5.55 5.71 5.71 1.94
IaG Shin mm 3.77 4.64 4.68 9.68 3.62
Simulation
Out ut
Carve R adii m
400 100 75 50 20
a -Tread MN/m 1289.5 1213.2 1223.1 1217.5 1120.5
a - F lan e MN/m 1922.3 2254.2 2266.8 2265.1 2300.6
Ty -Tread (N) 26.7 148.6 182.5 241.7 526.6
T~ -F lan e N 188.6 648.5 795.0 1047.5 2735.6
A -Tread de 4.30 4.37 4.48 4.48 4.10
A -F lan e de 76.4 76.24 7 .2 76.4 7624
AM mra0 4.13 15.12 18.12 24.07 50.35
eR (mm) 1.01 1.01 1.01 1.01 0.86
lat. Shin mm 5.42 5.30 5.23 5.16 3.66
~1a'Ji 1'MS!' 11G+ ~1 ~~
F igure 9.1 -Wheel-rail interfac e s ummary s heet
With referenc e to the above figure, the wheel-rail interfac e s ummary s heets inc lude interfac e
information relevant to s elec ting c ompatible wheel and rail profiles , c ons is ting of, Miniprof
information to vis ually as s es s the c ontac t c onditions , the s imulation output for key parameters
at a range of c urve radii and a rolling radius differenc e plot to indic ate the eq uivalent c onic ity
generated ac ros s the full range of lateral s hift.
CC.~! ~
~
~s
.r K ., ~~:~..~, .
79 ., ..kh~,
'.41 i:~y~, tiir~~
CONF IDENTIAL
8y comparis on ofi the s ummary s heets and through cons ideration of other fiactors relev ant to
rail s ection s election, s uch as av ailab ility in terms of phys ical production lev els of a s ection,
the cros s-s ectional area from a s tray current pros pectiv e, then the profile comb inations could
b e reduced to a s mall numb er and grouped within the three profile s ets des crib ed prev ious ly.
The wheel-rail interface s ummary s heets for each of the s imulated wheel and rail profile
comb inations can b e found in Appendix A2.
9. 2 Definition of ` P rofile Sets '
The requirement to form a group of compatib le wheel and rail profile comb inations b as ed on
commonly adopted wheel and rail profiles led to the dev elopment of three profile s ets . Thes e
are des crib ed in further detail in the flowing s ection. A matrix will then b e pres ented and
explained, which s hows the cros s-compatib ility of the s elected wheel and rail profiles .
9. 2. 1 P rofile Set 1 (New Sys tem)
F rom the analys is work carried out, it b ecame apparent that there was a comb ination of wheel
and rail profile which produced b oth good derailment protection and the mos t compatib le
wheel-rail interface conditions . This was found to b e a CT3 b as ed wheel profile, as s hown in
F igure 9. 2 b elow.
Modified CT3 P rofile
F lange root radius = 15mm
Taper angle = 1:26
R un-off angle = 1:10 '
F lange angle = 70
F lange tip radius = 5mm
. .e. . . . . , ,i . . . . . . . . . . . . . . . . . . . . . . . , . . . . j . . . . o. , . , . . . . . . . . . , . . . , . . . . , , . . . . , . . . . , . . . . . . , . . . . , j . . , . . . . . . . . . , . . , . . v . . . . . .
304 . Q )K '0. . _ . 't~ lP -0. . 3u1R A. . 3189 ~~~~ 'XAD ?@~ 9
F igure 9. 2 ` Set 1' wheel profile
The modified CT3 profile, as s hown ab ov e with the flat flange tip option, was s elected as the
op# imai profile for new s ys tems when comb ined with G4 9 and U9r54 non-groev P d rai! inclined
~ ~ ~
*r~
CONF IDENTIAL
at 1:40 and R i 59/60-R 13 grooved rail with a 1:40 head inc lination. The key featu res of the
profile are s u mmaris ed below:
15mm flange root radiu s redu c es flange c ontac t s tres s
1:26 taper angle as s is ts c onic ity and lower tread c ontac t s tres s
1:10 tread ru n-off as s is ts in preventing c ontac t between the ou ter s ec tion of the wheel
and the road s u rfac e, partic u larly when embedded rail bec omes head-worn or s u ffers
from drop within the trac kform
70 flange angle offers good derailment protec tion withou t being u nnec es s arily s teep,
whic h c an promote higher c ontac t s tres s throu gh inc reas ed normal forc e and als o
redu c es effec tive flangeway c learanc e at high angles of attac k
5mm flange tip radiu s offers a good c ompromis e between effec# ive derailment protec tion
and redu c ing the tendenc y for high c ontac t s tres s , whic h c an be c au s ed by very s mall tip
radii c ontac ting a s witc h blade (s mall c ontac t area) . A very s mall radiu s c an res u lt in high
initial wear, as evidenc ed with s mall radiu s flange tips at S&G in the UK
food derailment protec tion on all rail profiles s tu died
R elatively low wear and c ontac t s tres s when c ombined with s elec ted rail profiles
9. 2. 2 P rofile Set 2 (Exis ting Sys tems )
As oppos ed to produ c ing a s ingle profile s et to addres s only the requ irements for newly bu ilt
s ys tems or thos e happening to u s e the s u gges ted rail profiles , it was c ons idered es s ential to
provide an alternative profile s et whic h c ou ld be c ons idered for u s e by s ome of the exis ting
s ys tems , already operating in the UK.
Throu gh the analys is of the P has e 1 work, pres ented in Sec tion 2, it is evident that a
s ignific ant nu mber of s ys tems u tilis e grooved and non-grooved rail s ec tions with a 10mm
gau ge c orner radiu s . The s imu lation work demons trated that the 15mm gau ge c orner radiu s
of the ` Set 1' profile is not partic u larly well matc hed to thes e rail profiles and therefore a
s ec ond, ' Set 2' profile s et was generated.
A wheel profile bas ed on a modified vers ion of the SST profile was s elec ted from the analys is
work as being the mos t c ompatible profile for rail s ec tions with a 10mm gau ge c orner radiu s ,
and worked well with the head profiles of thes e profiles . The SST profile was modified in that
the gau ge c orner radiu s was redu c ed to 12mm, to redu c e c ontac t s tres s in flange c ontac t and
the flange geometry was modified to be the s ame as that of the ' Set 1' wheel, 70 flange
angle and 5mm flange tip radiu s . This provides benefits in c ons is tent flange geometry ac ros s
the profile s ets and c ou ld lead to the development of a s tandard s witc h blade des ign.
The ` Set 2' wheel profile is s hown in F igu re 9. 3 on the following page.
CONF IDENTIAL
. x i
~ . - ~ ~ ~ . m
i~
_
~ ~ _
. .
. _ _ . _ . _ _ .
z~ ae
Modified DIN P roF le
rmra F lange root radiu s = 12mm
Taper angle = 1:40
,: _
R u n- off angle = 1:10
; ~ ,,; ; F lange angle = 70
F lange tip radiu s = 5mm
; ts o
~ c ,
. i
. . . .
_ _ _
_ _ _ . _ . _ _ _ _ _ _ _
_ . . _ . . .
- - -
- . ~ _ . . . ' . _ . . _ ! - . . j . . _
F igu re 9. 3 ` S et 2' wheel profile
The modified S S T ` S et 2' wheel profile was fou nd to be the mos t c ompatible option of the
profiles s tu died when ru nning on BS 80A non- grooved rail profile, inc lined at 1:40 and
R i 59/60- R 10, S EI 35/41 grooved rail s ec tions .
The key featu res of the profile are s u mmaris ed below:
12mm flange root radiu s redu c es flange c ontac t s tres s on 10mm gau ge c orner radiu s
Tread profile provides a good c ompromis e when ru nning on BS 80A and grooved rail
s ec tions
1:10 tread ru n- off as s is ts in preventing c ontac t between the ou ter s ec tion of the wheel
and the road s u rfac e, partic u larly when embedded rail bec omes head- worn or s u ffers
from drop within the trac kform
70 flange angle offers good derailment protec tion withou t being u nnec es s arily s teep,
whic h c an promote higher c ontac t s tres s throu gh inc reas ed normal forc e and als o
redu c es effec tive flangeway c learanc e at high angles of attac k
5mm flange tip radiu s offers a good c ompromis e between effec tive derailment protec tion
and redu c ing the tendenc y for high c ontac t s tres s , whic h c an be c au s ed by very s mall tip
radii c ontac ting a s witc h blade (s mall c ontac t area) . A very s mall radiu s c an res u lt in high
initial wear, as evidenc ed with s mall radiu s flange tips at S&C in the UK
Good derailment protec tion on all rail profiles s tu died
R elatively low wear and c ontac t s tres s when c ombined with s elec ted rail profiles
s M
i' t'
CONF IDENTIAL
Determination of Tramway Wheel and R aii P rofiles to Minimis e Derailment F inai R eport
9.2.3 P rofile Set 3 (Tram-Train)
P rofile Sets 1 and 2, prov ide a number of c ompatible wheel and rail profile options for
tramway and lig ht rail s ys tems , howev er it is lik ely in the future that there will be an inc reas ed
adoption of tram-train s olutions .
Tram-train s olutions allow dual mode c ompatibility between Network R ail (NR ) heav y-rail
infras truc ture and lig ht rail s ys tems , inc luding thos e whic h employ embedded (g roov ed) rail
s ec tions and as s oc iated c ros s ing s work .
There are a number of s trateg ic hig h lev el dec is ions to be made, or perhaps derog ations by
Network R ail, before any propos ed new tram-train profile c ould be widely adopted. Howev er
the work c arried out within this report has hig hlig hted profiles whic h may be s uitable for
tram-train applic ations .
The fundamental differenc es between heav y and lig ht rail profiles are; a heav y rail profile will
hav e a treater flang e heig ht, s hallower flang e ang le and mos t s ig nific antly a s maller wheels et
bec k to bac k (fiang ebac k ) dimens ion than a lig ht rail wheels et. Thes e points are dis c us s ed
below:
The differenc e in flang e heig ht c an be addres s ed by running a lig ht rail type wheel
profile with an extended flang e heig ht to meet the NR s tandards , howev er, this would
reduc e the c learanc e to the bottom of the rail g roov e whic h may or may not be a
problem depending on the g roov ed s ec tion us ed. (See Table 2.3 Sec tion 2.2 pp.6 ). I t
may als o c aus e is s ues at flang e running c ros s ing s .
A lternativ ely NR c ould c ons ider allowing a derog ation to run with a wheel profile with
a reduc ed flang e heig ht, this would need c ons ideration of is s ues at c ros s ing s uc h as
s witc h blade ov erlap protec tion.
The differenc e in flang e ang le, 6 8 typic ally for heav y rail and 70 to 75 for lig ht rail
s ys tems , c ould be ov erc ome in the flang e des ig n s tag e, with it being more benefic ial
to mov e towards the 70 to 75 rang e to inc reas e derailment protec tion within the lig ht
rail s ys tem than mov ing towards the heav y rail ang le.
The differenc e in the wheels et flang ebac k dimens ion c an be ov erc ome by the s imple
adoption of a s tepped or c ut-out flang ebac k , as adopted at Manc hes ter Metrolink
(MML).
+s ~
.~. ~...,.~ ,,,.~ 83 .~.-..ht,~~.
..,~ ,
CONF I DENTI A L
Determination of Tramway VVheel and R ail P rofiles to Minimis e Derailment F inal R eport
There are a number of ather prac tic al is s ues to c ons ider in the development of a tram- train
profile whic h are bes t addres s ed through light rail indus try/NR c ommittee, thes e as ide the
res ults of this s tudy s ugges t that a modified vers ion of the CT3 profile would be a good
s olution on the bas is of performanc e and c ompatibility with both BS 113A and the propos ed
` Set 1' rail profile groups .
At the time of writing it was known that indus try dis c us s ions regarding tram- train wheel
profiles have c ons idered the MML and BR P 8 wheel profiles as pos s ible options , however the
P S would req uire s ignific ant flange modific ations to allow grooved rail running and the MML
profile has been s hown to be les s tolerant to flange c limb when in the new c ondition, than
other light rail profile des igns .
The figures pres ented below c ompare the c ontac t s tres s es on tread and flange between a
new BR P 8 profile, MML and khe CT3, running on new BS 113A rail.
s ea i~s ~s ~~wy
A~ ~ttat~
Ganta~c t St#~s z - Tr~s ~t
F igure 9.4 Tread c ontac t s tres s tram- train
I t c an be s een from F igure 9.4 above that the MML and CT3 profiles are s imilar in terms of
tread c ontac t s tres s , with the P 8 s howing s lightly higher values at radii of 600m and above,
this is s hown at between 0 and 450m s imulation dis tanc e on the x- axis of the above plot.
I n terms of flange c ontac t s tres s the CT3 profile s hows a c lear advantage, as s hown in F igure
9.5 below.
~s
iK5!
1~.. l
~~y.illt~,.
CONF I DENTI AL
Canka~d a3r~s s - ~a~ng~
F igu re J.5 F lange contact s tres s tram- train
The flange geometry of the CT3 profile res u lts in a s ignificantly lower contact s tres s to the
MML profile and the P 8 profile, " v hich can 4e s een to mov e into two- point contact at 600m on
the x- axis , which relates to cu rv es of 400m and below. The CT3 and MML profiles are in
two- point contact throu ghou t.
To conclu de, it may be that practicalities and/ or politics will dictate the s hape of the tram - train
( ' S et 3' ) profile, howev er the work pres ented here wou ld s u gges t that a s u itable profile cou ld
be bas ed on a modified CT3 type wheel which wou ld prov ide cros s compatibility with ' S et 1 '
s ys tems .
As has been dis cu s s ed, one of the is s u es relating to a tram- train profile is that it is not
pos s ible to dev elop a wheel profile which is entirely compatible with the s tandard BS 1 1 3A
heav y rail profile and the complete s et of rail profiles ( groov ed and non-groov ed) u s ed in the
UK light rail s ector. Therefore any tram- train profile will be a compromis e or pos s ibly des igned
to s pecifically s u it the light rail s ys tem onto which the exchange will take place.
i
CONF IDENTIAL
9.2.4 IR W Wheei P rofiles
As part of this profile s tu dy work , the feas ib ility of adopting larg e flang e root wheel profiles for
u s e on IR W ax les was inv es tig ated as means to redu c ing the tendenc y for IR W ax les to ru n
off-c entre and thereb y c au s e inc reas ed flang e wear prob lems .
Analys is work pres ented in S ec tion 8 demons trated that ev en with a s ig nific antly inc reas ed
flang e root radiu s of 22mm, ov er that of a more c onv entional 13mm root radiu s wheel profile,
there was ins u ffic ient lateral g rav itational s tiffnes s forc e produ c ed to affec t the g u idanc e of an
~~~~wheels et.
Withou t des ig ning a radic al IR W wheel profile to produ c e inc reas ed g rav itational
s elf-c entering , an approac h to improv ing the performanc e of IR W ax les , c ou ld b e to ac c ept
the offs et ru nning is s u e and to redu c e flang e wear throu g h improv ed c ontac t c onditions .
As the CT3 wheel profile has b een s hown to maintain s ing le point c ontac t at hig h lateral
s hifts , it is this profile whic h has b een adopted to demons trate the princ iple of redu c ing flang e
wear throu g h lowering c ontac t s tres s and Ty wear v alu es . Therefore profile c omb inations
identified in ` het 1' wou ld b e s u itab le for IR W equ ipped v ehic les and wou ld b e ex pec ted to
produ c e lower flang e wear rates than c onv entional DIN type profiles , whic h promote two- point
c ontac t and henc e hig her Ty' s .
The b enefit of redu c ed wear throu g h s ing le- point flang e c ontac t c an b e g reatly inc reas ed
throu g h the adoption of effec tiv e b og ie mou nted wheel flang e lu b ric ation. Trac k b as ed
lu b ric ation c an b e ineffec tiv e with IR W wheels du e to their often u npredic tab le ru nning
b ehav iou r.
9.2.5 Wheel- R ail P rofile S elec tion Matrix
The rec ommended wheel and rail profiles des c rib ed ab ov e c an b e repres ented b y a matrix ,
as s hown in Tab le 9.1 b elow.
~~
, , ., , ~; ;
CONF IDENTIAL
F inal R eport
R ail P rofile
S49 UIC 54 BS 113A BS 80A R i 59/60- R i 59/60- SEI 35/41
R 13 R 10 ( V ert. )
1:40 V ert.
Set 1 `~ ~ ~~~
x x x
~ ~:w tao~ ~:ao i:ao
N
= _ ' s et2
x x x x
p
~~:ao~
a
Set 3
~~~ ~~~ x x x
( 720) ( 720) ( 1:Y0)
( 7:40)
x/~~~~~~~~~~~profiles will run acceptably but are not the optimal choice
Table 9. 1 Wheel-rail profile s election matrix
The above matrix s ummaris es the profile s ets des cribed in the previous s ection and hig hlig hts
the fact that it is not pos s ible to produce a s ing le wheel profile to s uit all rails .
Thy `Set 3' , profile s et is bas ed on a CT3 type wheel profile, but it may be that an alternative
may be us ed bayed on indus try opinion. Additionally, the CT3 profile is not well s uited to
10mm g aug e corner rail, hig hlig hting the conflict between thes e rail s ections and the 13rr~rn
g aug e corner radius of BS 113A.
Table 9. 2 below s hows how the current tramway s ys tems would fit into the above s election
matrix.
Tramway Network
CTL MML NET SST MM
Sett x x x x
t.
d
~' Set 2 x ~~~
0
a
Set 3 ~~~ ~~~ x x x
X /~~~~~~~~~~~profiles will run acceptably but are not the optimal choice
Table 9. 2 Wheel-rail profile current s ys tems
It can be s een from Table 92 above that CTL is the only current s ys tem compatible with the
`Set 1' wheel profile due to its us ag e of R i 59-R 13 g rooved rail and S49 non-g rooved rail.
r
iry
~ny. rl~. l. ~
CONF IDENTIAL
The ` S et 2' wheel is lik ely to be compatible with the rail profiles us ed at NET, S S T and MM, all
employing 10mm gauge corner radius rails and partly compatible with the MML s ys tem which
has a wide v ariety of rails s ections in us e.
The 'S et 3' wheel is partly compatible with the CTL s ys tem due to the CT3 profile work ing well
with BS 113A but being compromis ed on S 4J at 1:20 inclination. Again, MML due to its
relativ ely heav y us age of BS 113A rail can be cons idered to fit relativ ely well within ` S et 3'.
The profile s ets recommended abov e repres ent fundamentally compatible wheel and rail
profiles , howev er, the characteris tic of any one particular s ys tem may dictate detail changes
to the s ugges ted profile forms . F or example a s ys tem with a number of mid-range curv e radii
may benefit from an increas ed lev el of conicity.
p ~ *
s i-f iyn "-
8V t51~ 5! .ry
~ rV= ! I I i.~ ~
CONF I DENTI AL
10. Conclus ions
A wide ranging s tudy has been carried out to review current tramway s ys tems and their wheel
and rail profiles within the UK. This was reported under the P has e 1 OR R s tudy, entitled
'A s urvey of UK tram and light railway s ys tems relating to the wheel/ rail interface' ~ '~ . P has e 2
of the work, pres ented within this report, analys es this initial s tudy and extends the work
through the application of wheel- rail contact analys is techniques and railway vehicle dynamics
modelling to determine optimis ed wheel and rail profile combinations which minimis e
derailment ris k and wear.
The following s ections s ummaris e the conclus ions which can be drawn from both the review
and analys is s ections of the P has e 2 work.
14.1 P has e 1 R eport R eview
There are eight different des igns of non- grooved rail and four different grooved rail
profiles currently us ed in operation an UK tram and light railway s ys tems
The eight UK tramway s ys tems each us e a different wheel tread profile des ign
Many of the UK s ys tems us e s everal different rail profile ( with one particular s ys tem us ing
s even different rail profiles )
The modern 2" d generation s ys tems us e nine des igns of vehicles from four manufacturers
Up to 15 vehicle derailments have been cited by a s ingle tramway operator
50% of all derailments , reported within the P has e 1 work, can be attributed to incidents at
turnouts , the maj ority of thes e derailments were due to detection and clos ure failures
rather than wheel- rail interface is s ue.
22% of cited derailments were caus ed by flange climb, thes e may be related to a
combination of poor wheel rail interface conditions and contributory factors s uch as track
twis t, poor bogie s etup, high friction condition or poor wheel finis h following turning.
7% of derailments were attributed to keeper wear. 4 / a and 2% res pectively were caus ed
by obs tructed groove and un- s pecified caus es at diamond cros s ings
CONF IDENTIAL
A wide v ariation in eq u iv alent c onic ity was fou nd ac ros s the s ys tems operating in the UK,
lev els between 0.06 and 0.58
C onic ity s hou ld be c ons idered when s elec ting wheel profiles for new and ex is ting
s ys tems , es pec ially where a s ig nific ant nu mber of mid rang e c u rv es ex is t whic h c an
benefit from improv ed wheels et g u idanc e
10.2 P rofile Analys is and Selec tion
Three profile ' Sets ' hav e been identified; for new s ys tems , for ex is ting s ys tems with
10mm g au g e c orner rails and for ` tram-train' s c hemes
The rec ommended profile s ets offer relativ ely low c ontac t s tres s es and v alu es of 7 ' q
(wheel-rai! wear index ) when ~; ompared to many of the profile c ombinations in c u rrent UK
operation.
The rec ommended profile s ets repres ent fu ndamentally c ompatible wheel and rail
profiles , howev er, the c harac teris tic of any one partic u lar s ys tem may dic tate detail
c hang es to the s u g g es ted profile forms . F or ex ample a s ys tem with a nu mber of mid-
rang e c u rv e radii may benefit from an inc reas ed lev el of c onic ity
New wheel and rail profiles hav e been rec ommended whic h are g eometric ally c ompatible
in the new c ondition, this prev ents problems with hig h initial wear rates of wheels and rails
and will g enerally prov ide profiles whic h are more s table in terms of s hape c hang e
I t has been s hown that a c onv entional wheel profile with a s ig nific antly inc reas ed flang e
root radiu s will not prov ide s u ffic ient g rav itational s tiffnes s forc e to improv e I R W wheeis et
g u idanc e
The c onc ept of adopting profiles whic h g enerate s ing le point flang e c ontac t, partic u larly
for I R W non-powered ax les , has been s hown to offer benefits in terms of dec reas ed
flang e wear rates , thes e benefits als o apply to c onv entional ax les
The Manc hes ter Metrolink (MML) wheel profile in its new c ondition was s hown to offer a
lower lev el of protec tion ag ains t flang e c limb relativ e to the other profiles s tu died
The MML wheel profile in the new c ondition is not c ons idered as a derailment ris k and
relativ ely low lev els of wheel flang e wear has been s hown to s ig nific antly improv e the
wheel' s flang e c limb protec tion
4M
s b ~r: .y..~it r
C ONF I DENTI AL
it was demons trated that the flange des ign methods s hould inc lude analys is of wheel- rail
c ontac t angle and wheel lift as an indic ator to flange c limb protec tion
With the exc eption of the MML wheel profile, all other wheel profiles s tudied s howed good
derailment protec tion
Wheels et fit analys is has demons trated that c ons ideration needs to b e giv en to the
groov e width, es pec ially in s mall radius c urv es , if premature keeper rail wear and
s ub s equent replac ement is to b e av oided
Vehic le c onfiguration and highway/ pedes trian s afety s hould b e a c ons ideration in the
s elec tion of groov e width, that is , a b alanc e needs to b e ac hiev ed b etween s uffic ient wear
allowanc e anr~ c learanc e for the wheels et in tight c urv es and maintaining ac c eptab le
lev els of s urfac e adhes ion/ entrapment hazard for other road/ s urfac e us ers
Alight rail and tramway wheel- rail 'Bes t P rac tic e' guide has b een dev eloped whic h further
expands on the prac tic al is s ues related to the findings of the s tudy and is pres ented in the
b es t prac tic e guide, titled 'A Good P rac tic e Guide for Managing the Wheel- R ail I nterfac e
of Light R ail and Tramway Sys tems ', R TU referenc e 90/ 3/ B ~ 9~
~ +
, ~ . , ~
CONF I DENTI AL
R eferences
1 . E . J H ollis , 'A s urvey of UK tram and lig ht railway s ys tems relating to the wheel- rail
interface', H ealth and Safety L ab oratory, F E /04/1 3, 2006,
www.rail- reg . g ov. uk /upload/pdf/R eport_F E -04- 1 4b . pdf.
2. BS E N 1 481 1 - 1 , R ailway applications Track Special purpos e rail Grooved and
as s ociated cons truction.
3. BS E N 1 3674- 1 , `R ailway applications Track R ail P art 1 : Vig nole railway rails
46k g/m and ab ove'.
4. BS E N 1 3674- 4. 'R ailway applications Track R ail P art 4: Vig nole railway rails from
27k g /m to, b ut ex cluding 46k g /m'.
5 . 'L ocal and reg ional railway track s in Germany', ( in German and E ng lis h) , VDV, Colog ne,
2007, ISBN 978- 3- 87094- 674- 6
6. I. McE wen and R . H arvey, 'Interpretation of Wheel/R ail Wear Numb ers ', Britis h R ail
R es earch, R eport R ef. TM VDY 004, J uly 1 986.
7. R . L ewis et al, 'Integ rating Dynamics and Wear Modelling to P redict R ailway Wheel
P rofile E volution', 6th International Conference on Contact Mechanics and Wear of
8. A. F Bower and K. L J ohns on, 'P las tic flow and s hak edown of the rail s urface in repeated
wheel- rail contact, Wear, vol. 1 44, pp. 1 - 1 8, 1 991 .
9. J . Stow and P . Allen, À Good P ractice Guide for Manag ing the Wheel- R ail Interface of
L ig ht R ail and Tramway Sys tems ', Manches ter Metropolitan Univers ity ( R TU) document
90/3/6, F eb ruary 2008
~~~
: : w
, ~. , ~; ,
CONF IDE NTIAL
Appendix Al Wheel- R ail Interface P lots
A1.1 Contact P os ition P lots
Croydon CR 4000 Wh~~l P rofii~ and S49 ( ' Iin2 0) R ail P rofile Croydon CT9 Wh~~l P roiil~ and 549 ~' Iin2 0) R ai! P rotfl~
80 80
i i i i i i
i i i i ~ i
60
_ 1 _ _ _ _ ~ _ _ _ _ L _ _ _ ! _ _ _ _ ~_ ' _ ' 1 _ _ _ 5p _ _ r _ _ _ y _ _ _ ' i_ ' _ _ T _ _ _ y _ _ _ _ i _ _ _ __
~ i i i i i
i i
i i i
i i i i i
i ~ i
40
_ _~__ _~' _' _f 1 __~____~____ 40
_ _1__ _I____I__ __J____I____.
~
\ I I I i
1 I I I
I 1 I I 1 1
1
z o
- - ; ;-- '- - - - ;- - ; - - 2 0 - - - r r~-- ~ - - -,- - - - ~- - - - -
0 _ - - Y _ _ _ _ _ - 1_ _ _ _ T
_ _
- - 0
_ _ - i _ - _ - _ _ ~ _ _ _ _I_ _ _ __
~ I I
I 1 Q
1 I I 1 I 1
- - '- - ' - - - -' - - - -'- - - - ' ' 2 0 ~
~- - - -' - - - - ' - - - - '- - - - ' -
i i i i i i
i i i ~ i i
' q p - - - r - - - ~- - - - r- - - r- - - ~- - - - r - - - -
~0 ___~_______' r___,____i____r____
i i i i i i
~0 ~BO 700 72 0 740 760 780 800 8:
680 700 72 0 740 760 780 800 B.
F igure A1.1 Wheel- rail contact pos ition, Croydon CR 4000 ( left) and CT3 ( right) wheel on
S49 ( 1: 2 0) rail
Dooltlands DLR 6 Wh~~l P refil~ and BS BOA R ail P roR l~
i i i i i i
40
___1___~_' _ ~I ____i___
i i ~ ~ i i
i i ~' i i
30
_ _ _ 7 _ ;. ' _ i_ _ ~ _ _ _ _i_ _ _
~,
2 0
__ _ ~1 _ _ _~__ _J_ __ _~_ __
~ i i i
P '
i I i
10 -
~__~ __i- __a___' ~_' _
i i i i
p _ _ _ _ ~ ~ ~ _7 _ 1__ _ _~__
1 0
~a
___I_ - _ I- - __~_- ~
_______~_
1
' ~ _ _ _ ~ _ _ - - - - _ F _ _ - } _ _ _ 1 - - _ _~- -
1
i 1 1
_ _ ~ - _ -
~ _ _ ~
- - I - - - I - - - - I - I I - - I - -
1
680 700 72 0 740 760 780 800 e:
Doelcl~nds DLR 6 WMN P rofile and BR 113A R ail P rofile
~ ~~I
~/ ;
1 ~~ r/
i
~/
~~
onu iw ~~u iau iou rau euu aw
F igure A1.2 Wheel- rail contact pos ition, Docklands DLR 5 wheel and BS 80A ( left) and
BR 113A ( right) rail
.:
93 ~,a~-~ F ,,,
t,. ,, ,~,.~~~
..
CONF IDENTIAL
Midlantl Metro WMSi P rofile and SE136G R ail P rofii~
~ ~ i ~ I ~ ~
1 I I 1
6p _ _ _I _ _ _ 7_ _ _ T _ _ _ Y _ _ _I _ _ _ y _ _ _ t _
1
1 1
40
___~ ___J___ _'__ ___1__
1 I I 1
1 i
ZO
I I ~ I '{ ~ ~ I 1
__.i. __I __I --_7_
1 1 I 1
i I i I 1
1 ~ I i
1 1 I I I 1
1 I I I 1
1 1
1 1
__.I ._ __T___r__.I ._~ --_Y__
1 I I 1
5p 1 I 1 1
fi60 680 700 720 740 7fi0 780 800
F inal R eport
F ig u re A1.3 Wheel-rail c ontac t pos ition, Midland T69 R ev.A wheel and SEI 35G ( left) and
BS 80A { rig ht) rail
SMfiNd Su p~ rWm WI N P roR b and SE13bG R ail P roR H
i i i i i i i
i i ~ i i i i
---r --7---t---r---r---~ ---r --
i i i i ~ i i
i i i i i
_ _ i _ _ J_ '_ ~ _ '_ J '_ _ 1 _ _
i
i
1
___i' l_ __i _'_~ ___T__
i 1 i ~ i
~ i i i i
i 0 i i i i
t i i i i i i
i i i i i i i
i i i i i i i
i i i i i i i
___~ __ __r___r___~ ___y___r__
1 I I l 1
1
S80 700 720 740 760 789 800
SMfiNd Su p~ R ram W hNl P rofile and BS BOA ( 1i M0) R ail P rafib
i ~ i i i ~
s i i i i i
---t----~ ----r---r---7----r --
~ i i i i i
i i i i
_ 1 _ _ _ _ ~ I 'I _ J_ _ _ _I '_ _
1 ~ ''
_ _ _ T _ _ _ _ ~ '_ ', _ _ ''~ _ _
i i i i l
i t i i i
i i 0 i i
i ~ i i i ~
i i i i i i
---------------------------
i i i i i i
i i i i i i
i i i i i i
---z----~ ----r---r---7----r--
i i i i i ~
1 1 1
700 720 740 760 780 8l) 0
F ig u re A1.4 Wheel-rail c ontac t pos ition, Sheffield Su pertram wheel and SEI 35G ( left) and
BS 8UA ( rig ht) rail
F ig u re A1.5 Wheel-rail c ontac t pos ition, Tyne and Wear Metro wheel BR 113A rail
~ 94 + i, ~ .e~ ,
~ r .~ a, rnrn-
.es ~ wmk~
CONF I DENTI AL
A1.2 R olling R adius Difference P lots
F igure A1.6 R oiling radius difference, Dockland L ight R ailway
F igure A1.7 R olling radius difference, Midland Metro
<tti
, ~ ~ ~ . , .
CONF IDENTIAL
~~~
F ig u re A1.8 R olling radiu s differenc e, S heffield S u pertram
v v av ~rint riot
F ig u re A1.9 R olling radiu s differenc e, Tyne and Wear Metro
CONF IDENTIAL
F inal R eport
A1.3 Contact Angle v Wheel L ift P lots
F igure , 4 1.10 Contact angle v wheel lift, Croydon Tramlink ~R 4 0 0 0
F igure A1.11 Contact angle v wheel lift, Croydon Tramlink CT3
Iiy
i, ~y.il~t ~~
r.... i' ~I~ .
CONF IDENTIAL
Determination of Tramway Wheei and R ail P rofil es to Minimis e Derail ment F inai R eport
F ig u re A1.12 Contact ang l e v wheel l ift, Dockl ands L ig ht R ail way
F ig u re A1.13 Contact ang l e v wheel l ift, Manches ter Metrol ink g rooved rail s ections
, ~ ` I ~ i"
CONF I DENTI AL
F igure A1.14 Contact angle v wheel lift, Manches ter Metrolink non- groov ed rail s ections
F igure A1.15 Contact angle v wheel lift, Midland Metro
~.
~*r
CONF IDENTIAL
F ig u re A1.16 Contact ang l e v wheel l ift, Notting ham Expres s Trans it
F ig u re A1.17 Contact ang l e v wheel l ift, S heffiel d S u pertram
aiy
e, : ~ ~ .., ~ ~ , , ~ , ~
100 ~ , ~ < ~ ~ ~ , , .~
r: rJ , .l iti
..... r' .i~ ' .
CONF IDENTIAL
VAMP IR E P lot
F igure A1.18 Contact angle v wheel lift, Tyne and Wear Metro
~i
~, .
*~..., ~- 101 ~, ~h.,
h4 .~~.iyi~lif i~
CONF IDENTIAL
Appendix A2 Wheel- R ail I nterfac e Summary Sheets
A2.1 BS 113A R ail Sec tion
BS 113A R ail - CT3 Wheel
t:t0 - 1:28 TAP ER 1:10 - 1:28 TAP ER
~ ' - R =tS - R =15
R =12.7 " .. R =12.7
t in 20
tin 40
Simulation
Output
Curve R adii m
400 100 75 50 20
a - TreaO MN m~ 835.7 711.1 690.6 675.8 630.1
a - F lan e MN m~ 1087.6 1231.9 1252.2 1258.9 1285.9
T - Tread N 25.8 98.0 112.9 140.3 259.9
T - F lange N 156.9 327.9 424.7 575.0 1499.4
A - Tread Oe 221 2.21 2.21 2.21 2.21
A - F lan e de 56.79 58.79 58.79 58.79 58.79
AoA mraA 2.66 16.68 21.49 28.66 68.44
~ R mm 0.60 0.67 0.67 0.65 0.53
lat. 5hik mm 6.23 6.26 6.25 6.18 5.60
Simulatlon
Output
Curve R adii m
400 100 75 50 20
s - Treatl MN m 621.6 338.7 190.7 174.2 0.0
a - F lan e MN m 1188.0 1420.4 1450.3 1453.8 1445.6
T - Tread N 20.3 18.7 9.6 4.6 0.0
T - F la e N 100.9 275.9 384.2 514.9 1357.0
A - Treatl tl 213 2.23 2.23 2.23 0.00
A - F lan e de 44.97 44.97 44.97 94.97 46.58
AoA mrad 218 16.44 21.11 28.46 68.22
~ R mm 0.54 0.59 0.62 0.62 0.30
L at. Shin mm 5.95 5.92 6.00 6.00 5.57
102 t
44- ~ i:ryi.11l l~
' v.-.' il.' .
CONF I DENTI AL
Determination of Tramway Wheel and R aii P rofiles to Minimis e Derailment
BS 113A R ail -NET P 3102639 Wheel
1:20 ^ 1:40 P AP ER
R =14
R -12.7
7 in 10
Simulation
Output
Curve R adii m
400 100 75 50 20
n -Tread MN m' 1285.6 1204.5 1204.5 1243.7 1482.6
a -Fla e MN/m) 3192.9 3448.2 3477.6 3475.1 3552.6
T -Tread N) 37.2 164.1 196.6 259.5 545.5
T -Fla e N 359.3 721.5 836.0 1056.1 2765.7
A -Tread tle 1.07 1.10 1.12 1.09 1.10
A -Fla e tle 75.68 75.68 75.68 75.68 75.68
AoA mrad 4.10 15.08 18.06 23.97 50.22
~R mm 0.38 0.38 0.37 0.35 0.01
~FT7~q~rim~E3:[~E~1~Eirt~FE~~Ed~:~~ I
1:20 - 7:4o TAP ER
R =14
R = 12.7
1 in 40
Final R eport
Simulation Curve R atliff m
Ou ut 400 100 75 50 20
a -Tread MN m~ 1244.9 1196.9 1217.3 1175.7 1113.4
a - Flan e MN/m) 2172.0 2479.8 2496.5 2495.1 2548.0
T -Tread N 31.0 171.5 207.6 275.0 597.5
T. -Fla e N 303.9 684.8 800.3 997.9 2579.0
A -Tread d 4.15 4.16 4.12 4.10 3.99
A -Fla e de 75.52 75.52 75.52 75.52 75.52
AoA mrad 2.36 9.15 12.15 18.15 45.19
~R mm 0.89 0.88 0.88 0.86 0.76
L at. Shift mm 5.36 5.30 5.22 5.09 3.84
BS 113A R ail -NET Wheel
C~~
~fw
., .
~.._ ~._ ~, ~
103
., , ., n.~, , .
, v. ~~i~ .
CONFI DENTI AL
Simulation
Output
curve R adii m
900 100 75 50 20
c -Tread MN/m 1057. 0 1005. 3 998. 1 980. 6 945. 5
- F lan e MN m~ 5140. 6 5556. 5 5597. 0 5600. 08 5759. 7
T - TreaA (N) 37. 3 169. 0 206. 1 270. 6 553. 9
T - F lan e (N 398. 8 630. 0 751. 3 914. 0 2301. 5
A -Tread de 1. 07 1. 06 1. 04 1. 05 7. 11
A - R an e de 73. 06 73. 06 73. 06 73. 06 73. 06
A oA mrad 3. 17 16. 48 20. 75 27. 74 66. 87
aR mm 0. 42 0. 91 0. 41 0. 40 0. 23
L at. Shin mm 5. 76 5. 60 5. 53 5. 4fi 4. 1?
. . . . . .
r{~. ~L wll ~itl%~F YR 6 R fa i0
5imutaHon
Outpu[
Curve R adii m
400 100 75 50 20
c -Tread MN/m 871. 0 818. 9 815. 1 812. 9 802. 4
a - F lan e MN m~ 2082. 9 2254. 8 2275. 6 2275. 4 2315. 7
T~ -Tread (N) 36. 9 158. 9 192. 7 253. 8 534. 1
T- - F lang e N) 300. 0 675. 7 809. 1 1040. 5 2695. 7
A -Tread de 1. 21 1. 24 1. 24 1. 26 1. 48
A - F lan e de 75. 52 75. 52 75. 52 75. 52 75. 52
A oA mrad 310 17. 29 21. 34 28. 33 67. 08
, ~R mm 0. 43 0. 42 0. 92 0. 42 014
L at. Shin mm 5. 72 5. 59 5. 52 5. 45 4. 06
BS 113A R ail -SST Wheel
t
+r
104 , ~~ , , , s , , ~
CONF IDENTIA L
A2.2 BS 80A R aii Section
BS 80A R ail - CT3 Wheel
w. ' .
1:10-.1:26 TAP ER 1:to -728 TAP ER
-j ~R = 11AY... R =71.H ` ~
~~..:.,~. ,.. ~.. ...
.r . -. .. ~ - .. ..~ .. . .. ....
.....
..... ...
... ... .... - .~
tin 20 tin 40
d::.
N3i
.. ~
'
i t/ Gf31~1TACT dATd P LOT
. .~
---
~r~,
BS SOA R ail - CT3 Wheel
Simulation
Output
Curve R adii m
q00 100 75 50 20
a -Tread MN m 3164.4 19 36.4 19 35.5 19 39 .4 1789 .2
a - F lan e (MN/m 0.0 0.0 0.0 0.0 209 6.10
Tv- TreaA (N 12.4 259 .6 355.2 489 .2 150.7
T. -F la e N 0.0 0.0 0.0 0.0 1030.7
A -Tread de 12.08 4238 43.77 43.84 33.12
A - F lan e de 0.00 0.00 0.00 0.00 70.38
AoA mraA 0.60 15.22 19 .74 27.4 67.44
~R mm 2.40 6.35 6.62 6.63 4.83
lat Shin m 2.9 7 5.28 5.33 5.34 4.9 0
SimWaGOn
Uu u[
Curve R adii m
400 100 75 50 20
rs -Tread MN m 3164.4 19 36.44 19 35.5 19 39 .4 1789 .2
c - F lan e MN/m 0.0 0.0 0.0 0.0 2046.1
T~ -Tread N 21.8 250.9 352,9 486.5 79 2.9
T.-F lan e N 0.0 0.0 0.0 0.0 9 19 .1
A -Tread de 16.02 9 2.25 13.38 43.44 34.48
A - F lan e de 0.00 0.00 0.00 0.00 70.03
AoA mratl 1.19 14.68 19 .R 27.30 67.40
nR mm 2.58 6.08 618 6.29 9 .75
La[. Shin mm 3.89 5.20 5.24 5.24 4.87
1*M
~~ ~- ~ 105 u,~,h~~,.~.
v~~~~~~~
,..,,~~,
CONF IDENTIAL
BS 80A R ail -NET P 3102639 Wheei
120 - 1:40 TAP ER 120 ^ 1:40 TAP ER
. . . _ . . .
_ _ -
R =74 . . . . . . . . . . . _ . R =14
R =71. 11 R =11. 11
1 in 20 9 in 40
Simulation
ou~uc
Curve R adii m
aoo ioo ~s s o z o
s -Tread MN m~ 631. 3 0. 0 0. 0 0. 0 0. 0
a - F lan e MN/m 1594. 3 1752. 1 1751. 7 1753. 2 1794. 8
T~ -Tread N 16. 23 0. 0 0. 0 0. 0 0. 0
T~ - F lan e N 45. 11 274. 0 354. 9 487. 8 1335. 3
A -Tread De 1. 47 0. 0 0. 0 0. 0 0. 0
A-F lan e de 29. 94 43. 42 44. 59 44. 63 4fi. 49
AoA mrad 4. 07 13. 90 16. 97 23. 01 50. 09
oR mm) 0. 10 0. 24 0. 24 0. 24 0. 05
Lat Shift mm 4. 07 4. 57 4. 60 4. 67 3. 99
Simulation
oup,u[
Curve R adii m
aoo ioo ~s s o z o
o -Tread MN/m~ 1773. 3 1755. 5 1754. 7 1756. 6 1402. 8
a - F lan e MN/m 0. 0 0. 0 0. 0 0. 0 2409. 9
T -Tread N 18. 5 265. 0 343. 2 477. 5 585. 6
T, - F lan e N 0. 0 0. 0 0. 0 0. 0 2262. 5
A -Tread d 17. 58 43. 04 43. 78 43. 86 12. 73
A - F lan e de 0. 0 0. 0 0. 0 0. 0 76. 02
AoA mrad 4. 17 13. 81 16. 84 22. 85 49. 72
dR (mm 2. 38 5. 57 5. 70 5. 71 2. 06
LaL Shin mm 3. 53 4. 46 4. 49 4. 49 3. 30
w ,;~
106 ti~:~~:~,~, ~,
Sv. ~ ~ytir,.
CONF IDENTIAL
BS 80A R ail -SST Modified DIN25 Wheel
f ~.4. - 1. 40 TAP ER `y. .. .. 1:40 TAP ER '~..
R =13 R =73
Simul ation
~+~P ~
Curva R adii m Simul ation
Output
Curve R adii m
400 100 75 50 20 400 100 75 50 20
c -Tread MN/m 814.1 627.8 609.1 594.2 521.9 a - Treaa MN/m 1523.8 1435.8 1470.3 1438.0 579.4
a - F l an e MN/m 1168.9 1341.0 1356.4 1358.9 1383.1 a - F l an e MN/m 0.0 826.7 714.3 720.7 1235.7
T - TreaA N 25.4 64.2 69.5 88.0 149.59 T -Tread (N 12.7 203.0 293.1 397.7 322.4
T~. -F l a e N 106.8 295.7 378.3 519.9 1368.0 T -F l a e N 0.0 110.1 86.0 123.0 1541.7
A -Tread de 1.46 1.47 1.46 1.48 138 A -Tread Oe 19.58 37.52 40.69 40.69 1.29
A - F l an e tl e 53.05 53.05 53.05 53.05 53.05 A -F l a e d 0.0 62.34 62.34 62.34 6234
AoA mrad 2.67 16.79 20.50 27.61 66.60 AoA m20 0.67 16.33 20.17 27.47 66.36
oR mm 0.94 0.44 0.46 0.44 0.21 dR mm 2.17 9.29 4.76 4.76 0.19
LaG Shik mm 5.19 5.11 5.17 5.10 3.71 La[ . Shin mm 4.38 4.84 9.91 4.91 3.61
OJ 8UN R al l ~ JJ 1 rvnee~
/*
+Ic, rr{ nAit, , .
CONF IDENTIAL
A2.3 S49 R ail Section
S49R ail -CT3 Wheel
c~~
as ~.!
1
.rn.s ;
1 : 1 0 ^1 26 TAP ER 1 : 1 0 -1 : 26TAP ER
'~ -
R =1 3 R =1 3
J
1 in 20 1 in 40
1 a - . _
y'~ _ ~ONTAGT DATA P 6C~T
-_ _
1
~:
_
~
, .
i
--^-._ CTJ bd+s d R iS~}Y: F 1 NwlY 1 r1 ~
YrTV
F .
Simulation
Output
Curva R adii m
40 0 1 0 0 75 50 20
a -Tread MN m' 832.2 1 450 .8 1 487.1 573.8 347.0
v -F lan e MN/ 1 244.5 0 .0 0 .0 1 483.1 1 527.5
T~~ -Tread N 21 .9 272.8 383.7 51 .7 38.5
T -F lan e N 1 1 5.7 0 .0 0 .0 50 0 .4 1 355.5
A -Tread de 2.0 1 42.23 44.27 1 5.57 2.0 1
A -F lan e de 47.93 0 .0 0 0 .0 0 47.93 47.93
AoA mrad 2.40 1 6.25 21 .27 28.52 68.46
4R m 0 .54 6.69 7.0 1 2.65 0 .53
Lat Shik mm 6.1 4 6.1 9 6.1 9 6.22 5.81
Simulation
Output
Curve R adii m
40 0 1 0 0 75 50 20
rz -Tread MN/m' 1 457.8 1 386.1 1 292.8 1 344.8 1 0 1 4.5
a -F lan e MN m 0 .0 0 .0 1 693.0 1 0 23.0 20 96.1
T~~ -Tread N 1 2.9 279. i 1 97.2 454.2 260 .4
T. -F lan e N 0 .0 0 .0 1 91 .8 58.9 1 240 .1
A -Tread de 1 5.86 42.85 36.26 43.29 1 9.61
A -F lan e de 0 .0 0 0 .0 0 54.28 541 8 54.28
AoA mrad 0 .56 1 6.24 20 .83 28.29 68.0 6
~R mm 2.37 6.0 4 5.0 1 6.1 2 2.81
Lat Shin mm 5.24 5.98 5.94 5.99 5.49
*~
.i4e s ., ~"
`~'" 1 0 8 k~, -, : n~s ~.r=
.. ~, ~y. diE r
CONF IDENTIAL
y ~ ~ ~ t ~ - ~
_ . ~ t r a n s f o r m i n g p u b l i c t r a r ~ s p Q r t
Met r o Del i ver y Pr o g r a m m e
Tr a m Si d Revi ew Meet i n g
16 ~ l o vem b er 2011 La r g e Br ea k- o u t r o o m 2n d Fl o o r Rea r '1300 - '1600
AGENDA
1. Su m m a r y o f Fi n a n c i a l Eva l u a t i o n GM
2. ~ u r r ~ m ~ r y o f Tec hn i c a l eva l u a t i o n ?
3. ~ L~ r r ~ m ~ r y r ef ~ c n t r ~ c t u ~ l Eva l u a t i o n UG
4. CJver vi ~ w ~ f ~ s ~ ~ s ~ r r ~ en t PN
~ . ~ u ks t ~ : ~ ~ di ~ g I~ ~ u es/~ o n c e~ r ~ ~ P~
~ . ~ ~ s # a n d ~ ~ n a l f Jf f ~ r ~ r o ~ es s P, A
7. Do c u m en t a t i o n Re~ ~ i r ed PA
8. Ti m es c a l es PA
9. An n o u n c em en t s/Bi dder Ma n a g em en t PA
S49R ail -NET P 3102639 Wheel
120 - 1:40 TAP ER 1:20 - 1:40TAP ER
R =14 R =76
R =13 R =13
1 in 20 tin 40
~ i/ ~ 3?N1'A~ T L i~ R TA F ~ . $ rY
a
_ _
. _ .
_ _ . . . _ _ ---
----_ . . _ .
pt rr
. _ . . . .~ . . .v. . . ,.~ . =.r.~ ~ . . . . em. . vus . . . ww. ~ ~ . . e. . ~ . . vw - n3.rrn,nwwwm~ . vm. . . . m. . w~ ~ . 0-:. :as--
Simulation
Output
Curve R adii m
400 100 75 50 20
a -Tread MN/ m 5773 538. 5 5343 812. 9 1230. 3
a - F lan e MN/ 2845. 5 3135. 8 3163. 1 3178. 3 3248. 9
T - TreaO N 29. 9 151. 5 182. 9 241. 0 504. 2
T~ - F lan e N 26J. 2 636. 6 756. 5 993. 1 2631. 1
A -Tread de 2. 66 2. 86 2. 86 S. R 1. 12
A - F lan e de 74. 70 74. 70 74. 70 74. 70 79. 70
AoA mrad 4. 09 15. 15 18. 18 24. 13 50. 38
~ R mm) 0. 76 0. 77 0. 77 0. 57 0. 03
L at Shin mm 5. 42 5. 34 5. 32 5. 18 3. 90
Simulation
Output
Curve R adii m
400 100 75 50 20
a -Tread MN/ m 1328. 2 1243. 6 1240. 1 1238. 1 1208. 0
c - F lan e MN/ m 1926. 6 2292. 2 2312. 7 2317. 9 2352. 5
T-i -Tread N 21. 3 151. 0 184. 3 245. 7 535. 9
T -F la e N SJ0. 4 725. 0 888. 6 1187. 4 3098. 2
A -Tread de 4. 89 4. 89 4. 70 4. 13 9. 13
A - F lan e de 77. 85 77. 85 77. 85 77. 85 77. 85
AoA mrad 4. 10 15. 02 18. 04 23. 96 5030
oR mm 1. 18 1. 18 1. 16 1. 08 0. 71
L at. Shin mm 5. 31 5. 23 511 5. 07 3. 76
+
~ . ~ ; ,. . . . ~ :. ~ .
109 ~ +,~ n. ,~
_ ,,. ,~ ,
CONF IDENTIAL
Simulation
Output
Curve R adii m Simulation
Output
Curve R adii m
400 100 75 50 20 qW 100 75 50 20
a -Tread MN/m~ 3083.0 1029.0 1032.5 1176.6 1074.6 a -Tread MN m~ 1297.9 1215.5 1201.8 1194.0 1158.8
c -F la e MN/m~ 6547.2 7091.5 7150.3 7162.7 7376.3 a - F a MN m' 3373.0 3555.1 3557.9 3548.0 3544.9
T - Treatl N 37.1 i6J.9 205.0 268.8 546.1 T - Treatl N 28.0 151.3 184.3 242.5 525.1
T - F lang e N 390.8 603.83 717.4 875.5 2202.4 T - F lan e (N) 197.1 708.2 873.6 1154.9 2967.9
A -Tread tle 0.96 7.00 0.98 0.74 0.85 A - Tread tl 3.41 3.41 3.41 3.41 3.41
A - fla e tle 7214 72.24 72.24 72.24 72.24 A -F la e tle 7710 T7.20 77.20 77.20 77.20
AoA mratl 3.18 16.38 20.70 27.68 66.86 AoA mrad 2.76 17.56 21.44 28.37 67.18
dR mm 0.41 0.41 0.40 0.38 0.24 dR mm 0.75 0.75 0.74 0.73 0.48
L at. Shin mm 5.76 5.61 5.53 5.46 4.17 L at. Shin mm 5.62 5.53 5.42 5.35 3.98
tir~nrr~ ~ari~F ~Y xa P ~4}T
ris
!~~ 110
,
CONF IDENTIAL
A2.1 R i 59 Grooved R ail Sections
Simulation
Output
Curve R adii m
400 100 75 50 20
c -Tread MN m' 2396.7 2271.3 2258.2 2262.6 2155.0
a - F lan e MN m' 0.0 0.0 0.0 0.0 1724.4
T- -Tread N 16.2 266.1 352.9 467.5 905.4
T. - F lan e N 0.0 0.0 0.0 0.0 703.8
A -Tread tl 15.69 42.10 43.13 43.13 37.71
A - F la e de 0.0 0.0 0.0 0.0 70.42
AoA mraA 1.06 15.56 19.77 27.39 67.53
~R mm 2.26 6.25 6.56 6.59 0.81
L at. Shik mm 3.62 5.33 5.41 5.41 5.09
Simulation
Output
Curve R adii m
400 100 75 50 20
c -Tread MN m 956.5 1360.0 1354.7 1355.8 583.2
a - F lan e MN/m~ 0.0 0.0 0.0 0.0 1869.4
T. - Tread N 11.8 281.0 381.1 511.9 447.3
T - F lan e N 0.0 0.0 0.0 0.0 1986.9
A -Tread de 15.66 42.67 44.36 44.53 4.21
A -F la e Oe 0.0 0.0 0.0 0.0 69.88
AoA mrad 0.45 16.46 21.06 28.45 68.17
~R mm 2.29 4.73 5.95 5.95 4.95
L ot Shin mm 5.65 6.06 6.08 6.09 5.59
K~aa ex a~s - ~: i s kaner~
iii
~, . a, ~.~, t~
111 .~a.< n., ,
~: , N., ~; r, -
CONF IDENTIAL
R i59 R ail -NET P 3102639 7Nheel
- 120^1:40 TAP ER 720-7;40 TAP ER
. . _ , i R =1q , . . .
~ s ~ ~ _
R =14
R i59-R 10 R i59-R 13
Simulation Curve R adii m Simulation Curve R adii m
Output 400 100 75 50 20 Output 400 100 75 50 20
a -Tread MN/m~ 2125. 6 2146. 6 2132. 4 2136. 1 1923. 6 rs -Tread MN/m~ 128 9. 5 1213. 2 1223. 1 1217. 5 1120. 5
a - F lan e MN/ 0. 0 0. 0 0. 0 0. 0 2228 . 3 a - F lan e MN/m 1922. 3 2254. 2 2266. 8 2265. 1 2300. 6
7 -Tread N 19. 9 274. 2 355. 2 491. 6 622. 1 T -Tread N 26. 7 148 . 6 18 2. 5 241. 7 526. 6
T- - F lan e N 0. 0 0. 0 0. 0 0. 0 2037. 5 Tv - F lan e N 18 8 . 6 648 . 5 795. 0 1047. 5 2735. 6
A-Tread de 25. 23 43. 60 44. 40 44. 90 22. 93 A-Tread de 4. 30 4. 37 4. 48 4. 48 4. 10
A - F lan e de 0. 0 0. 0 0. 0 0. 0 76. 03 A -F la e de 76. 24 76. 29 76. 24 76. 24 76. 24
AoA mrad 4. 02 13. 99 17. 04 23. 05 50. 03 AoA mrad 9. 13 35. 12 18 . 12 24. 07 50. 35
oR mm) 2. 11 5. 55 5. 71 5. 71 1. 94 oR mm 1. 01 1. 01 1. 01 1. 01 0. 8 6
L at Shin mm 3. 77 4. 64 4. 68 4. 68 3. 62 lat Shin mm 5. 42 5. 30 5. 23 5. 16 3. 8 6
~ . , . . . ~ , . . _ _ _ . _ . . e. . . _ . , . , _ . . _~ ~ ~ ~ ~ "_ ~ _ . . _ _ ~ . . . . . . . . . _ _ . _ . _ .~ . . .~ _ . _ _ . . ~ . _ _ _ ~ _ . . . _ _ . _ . _ _ . . . . . . _
~ +~ fTA~ t GATd I'4~ T
_ .
i
:, . v
-~ >. .~ ~ . . .~ . . . . . _ >. . . :a. ~ , ~ . ~ -~ a~ ~ . . . ~ , . .
~ .
F T'~
**
~ ~ ~ 112
CONF IDENTIAL
R i59 R ail -SST Modied DIN25Whee1
1:40 TAP ER 1:40 TAP ER
R =13 R =13
R =10 R =73
R i59-R 10 R i59-R 13
r
. , . . p . .
SlmulaHon
Outp ut
Curve R adii m
900 100 75 50 20
a -Tread MN/m 1940. 5 1929. 1 1933. 4 1934. 9 1617. 1
a -Fla e MN rr~ 0. 0 0. 0 0. 0 0. 0 2311. 0
T -Tread N 22. 8 283. 4 356. 2 489. 7 5603
T~ -Fla e N 0. 0 0. 0 0. 0 0. 0 2248. 7
A - Tread d 23. 06 43. 33 93. 65 43. 65 12. 66
A - Flan e d 0. 0 0. 0 0. 0 0. 0 75. 63
AoA mrad 1. 66 16. 05 20. 15 27. 55 66. 59
~ R mm 2. 26 4. 92 9. 97 4. 97 1. 54
L at. Shin mm 4. 32 5. 07 5. 08 5. 08 3. 87
Final R ep ort
Simulation
Outp ut
Curve R adii m
900 100 75 50 20
c -Tread MN/ 858. 4 644. 9 622. 7 578. 1 774. 7
a - Flan e MN rr~ 1985. 4 2219. 3 2237. 0 2291. 9 2289. 4
Tv -Tread N 28. 6 150. 3 181. 7 2371 502. 5
T~ + - Flan e N 200. 6 646. 2 791. 3 1050. 7 2751. a
A -Tread d 332 2. 81 2. 72 2. 17 1. 22
A - Flan e de 75. 70 75. 70 75. 70 75. 70 75. 70
Ao~ mrdd 2. 77 17. 65 21. 48 28. 41 6739
~ R ~ mm 0. 78 0. 67 0. 65 0. 55 0. 22
L at. Shin mm 5. 72 5. 58 5. 51 5. 37 4. 15
p ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~b~ Td ~ L ~ ~
GL ti. ~ i~ /
ii
+ . ~ . o. ~ . . y. , . rt
113 L , , . n~ ~ ; . ,
v. . ~ i. iv. dirr
CONFIDENTIAL

Determination of Tramway Wheel and Rail Profiles To Minimise Derailment

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Determination of Tramway Wheel and Rail Profiles To Minimise Derailment

Uploaded by

Copyright:

Available Formats

Rail Te~hl~ l~l~~t

You might also like