Journal o f Terramechanics, Vol. 30, No. 2, pp. 85-100, 1993. 0022-4898/93 $6.00+0.

00
Printed in Great Britain, Pergamon Press Ltd
© 1993 ISTVS

T R A C T I O N P R E D I C T I O N U S I N G SOIL P A R A M E T E R S
O B T A I N E D WITH A N I N S T R U M E N T E D A N A L O G D E V I C E

S. K. UPADHYAYA*and D. WULFSOHNt

Summary--A fully instrumented device capable of measuring soil sinkage and shear para-
meters developed at the Agricultural Engineering Department, University of California, Davis
was employed to conduct in situ sinkage and shear tests in a tilled and a firm, dry, Yolo loam
soil. Similar tests were also conducted in a tilled, moist Yolo loam soil. An 18.4R38 tire was
tested at different levels of inflation pressure and axle loads in these soil conditions. Soil
parameters obtained using the instrumented device were related to the traction prediction
equation parameters using traction mechanics, principle of conservation of energy and
dimensional analysis.

NOTATION
a, a', b', c, c' traction prediction equation parameters
b tire width
C cohesion
i slip
i' proportionality constant
] horizontal soil deformation
k soil sinkage constant
contact length
lw contact width
n exponent in sinkage equation
P average contact pressure
r rolling radius
Xl a,/lo
x2 asA(. + 1)to]
Y the coordinate direction along the line of travel of tire
A minor axis of the elliptical contact area
A¢ actual contact area
B major axis of elliptic contact area, 1¢/2
K soil shear modulus
Kt tire stiffness in the tangential direction
K. tire stiffness in the normal direction
NT net traction
T input torque
W dynamic load on the axle
&t tire deformation in the vertical direction
soil deformation in the vertical direction
maximum deformation of soil in the vertical deformation including slip sinkage
¢ internal friction angle of soil
rh tire deformation in the tangential direction
Y -Ya
~wr net traction coefficient

*Professor, Agricultural Engineering Department, University of California, Davis, CA 95616, U.S.A.

tAssistant Professor, Agricultural Engineering Department, University of Saskatchewan, Saskatoon,
Saskatchewan, Canada S7N 0W0.
85
86 S.K. UPADHYAYA and D. WULFSOHN

~AGT gross traction coefficient

0 tire angular rotation
p proportionality constant
7- shear stress
Tmax maximum shear stress
zX loss of forward motion due to soil and tire deformation

INTRODUCTION AND REVIEW OF LITERATURE

SOm--eNVUMAT~C tire interaction is a very complex process. Tire tractive ability
depends on tire type (radial vs bias), tire geometry (width, overall diameter, section
height), lug design, inflation pressure, dynamic load on axle, and soil type and
conditions [1-3]. Upadhyaya et al. [3] developed traction prediction equations for
radial ply tires. They found that soil type and condition were the most important
factors influencing traction. Upadhyaya and Wulfsohn [4] have presented a review of
empirical traction prediction equations which utilize soil cone index to represent soil
strength. The soil cone index is a very convenient parameter to use in predicting
traction, since it is easily and rapidly obtained in the field. It is a composite soil
parameter which depends on compressive and shear properties of soil. Since traction
is also dependent on compressive and shear characteristics of soil, it is argued that
soil cone index can be used to predict traction. From a theoretical standpoint, for clay
soils which can be assumed to exhibit rigid-plastic behaviour, cone index can be
shown to represent soil shear properties. However, for other soil conditions, the
relationship between soil cone index and soil properties is very complex. Therefore,
except for purely clay soils, the ability of soil cone index to predict traction is
expected to be poor [4]. Upadhyaya et al. [3] conducted extensive field tests in 10
different soil conditions and found that:

/~N,r = N T / W = a(1 - e -c/) (1)

#01= T / r W = a'(1 - b'e -c'i) (2)
where /~NT = net traction coefficient:
/~T = gross traction coefficient;
NT = Net traction:
W = dynamic load on axle:
T = input torque;
r = rolling radius:
i= slip;
a, c, a', b', c' = traction prediction parameters.
Equations (1) and (2) fitted the experimental data with a very high coefficient of
multiple determination (R 2 of 0.95 or better). They claimed that the five parameters
(a, c, a', b', c') depended on tire type (radial vs bias), tire geometry (width, section
height and overall diameter), tread geometry, tire loading (inflation pressure and
dynamic axle load), and soil type and conditions. They found that soil cone index did
not correlate well with these traction prediction parameters. Based on traction
mechanics, conservation of energy and dimensional analysis they showed that sinkage
and shear properties of soil are more appropriate soil parameters to use.
Upadhyaya et al. [5, 6] developed an instrumented analogue device to measure
traction related soil properties in situ. This device is capable of measuring soil sinkage
 TRACTION PREDICTION USING SOIL P A R A M E T E R S 87

parameters using rectangular plates, soil shear parameters using rectangular grouser
plates and soil cone index using the standard cone penetrometer. Field tests were
conducted using this device near the U. C. Davis campus. The results indicated that
this device works well in all three modes [5].
The objective of this paper is to relate the traction prediction equation parameters
(a, c, a', b', c') to soil sinkage and shear parameters obtained using the instrumented
analog device.

EXPERIMENTAL T E C H N I Q U E

Traction tests in a dry Yolo loam soil

An 18.4R38 tire was tested at three different axle loads and two different inflation
pressures in a tilled and in a firm Yolo loam soil. Each of these 12 tests were
replicated twice. Plate sinkage tests were also conducted in the same two soil
conditions using our instrumented soil test device. Three different plate sizes were
used for the sinkage tests. Each sinkage test was replicated eight times. Three grouser
plates were also tested. Each plate was tested at three different vertical loads, in both
the tilled and firm Yolo loam soil. Each test was replicated three times. Table 1 lists
the size of sinkage and grouser plates used in this study. Upadhyaya et al. [5, 6] have
described the field test procedure employed to determine soil sinkage and shear
parameters using the instrumented soil test device in detail. Twelve soil cone index,
density and moisture readings were obtained in each of the test soil conditions.
Unfortunately, the cone index data for the tilled soil were lost due to computer
hard-disk problems. Table 2 lists means and standard deviations of the soil physical
parameters obtained in the test locations.

Traction tests in a moist Yolo loam soil

Only a limited number of six tests were conducted in a twice-disked, moist Yolo
loam soil. The results of the traction test in the dry condition indicated that the
replicates of traction tests do not differ from each other significantly. The traction
tests were conducted at three different vertical loads and two different inflation
pressures. Moisture, density and cone index data were obtained in the test location as
before. Moreover, grouser plate tests and plate sinkage tests were also conducted
using the same test design as before. During the grouser shear tests, an attempt was
made to collect data to get some insight into slip sinkage process. The grouser plate
was placed on a test site and the vertical load was slowly increased to the desired

T A B L E 1. SIZES OF SINKAGE AND GROUSER PLATES USED 1N THIS STUDY

Width, ram Length, mm

Sinkage pla~
1 75 200
2 85 180
3 100 150
Grouser plate
1 75 200
2 85 180
3 65 180
88 S. K. UPADHYAYA and D. WULFSOHN

TABLE 2. SOIL P H Y S I C A L P A R A M E T E R S IN A D R Y YOLO L O A M SOIL A T I'HE TRA(711ON q ES'I SITE

Parameter Location Firm soil Tilled soil

50 mm 1446/50" 1377/49
Density, kg/m 3 100 mm 1499/64 1395/62
150 mm 1486/40 1439/80
Moisture content, % dry
basis Surface 6.78/0.88 6.43/0.41
Cone index, kPa Top 150 mm 3017/1057 -
Sinkage parameterst
kl, kPa -348.4 267.6
k:, kPa/m 20.8 0.516
n 0.65/0.06 0.60/0.04
Shear parameters
Cohesion, c, kPa 15.36 i1.92
Internal friction angle, q~, degrees 40 22.5
Shear modulus, K, mm 14.1/7.8 10.0/4.6

*Mean/standard deviation.
tSinkage parameter, k = k~ + k2b; where b = plate width, kt and k: = soil sinkage parameters.

value. Once the vertical load reached the desired value, the shear load was increased
until the soil sheared. This type of test provides a complete history of vertical load
versus soil sinkage in the presence of shear load.
The soil physical parameters obtained during these tests are listed in Table 3. The
density and moisture data given in this table correspond to the average of six data
points, and the cone index values correspond to the average of eight data values.

Analysis of traction test results'

Figure 1 is a typical plot of experimental traction data. Upadhyaya et al. [3]
developed a non-linear regression technique to curve fit experimental data to
equations (1) and (2).
Table 4 lists the results of our traction tests. Upadhyaya et al. [3] developed
expressions to relate, a, c, a', b' and c' to soil and tire parameters using traction

TABLE3. S O I L P H Y S I C A L P A R A M E T E R S IN A T I L L E D , M O I S T Y O L O L O A M SOIL, A ] T H E TRA('TION "IEST SII L

Parameter Location Value

50 mm 1258/34*
Density, kg/m 3 100 mm 1285/27
150 mm 1365/'61
Moisture content, % dry basis Surface 12.61/1.53
Cone index, kPa Top 150 mm 228/'149
Sinkage parameterst
kl, kPa 73.64
k2, kPa/m 3058.5
n 0.89/0.1
Shear parameters
Cohesion, c, kPa 22.67
Internal friction angle, ~p, degrees 22
Shear modulus, K, mm l 3.0//3.4

*Mean/standard deviation.
[Sinkage parameter, k ~ k~ + k2b; where b = plate width, k~ and k2 = soil sinkage parameters.
 TRACTION PREDICTION USING SOIL P A R A M E T E R S 89

1.0"

• Expeflmenl TIrW
• Exp~im4mtalTE
/ |~z~ .,o- Fi.edD~W
o~ t ~,.I 4- FitledT/rW

o2./i, "'"
0.0"
--..,
0 20 40 60 80 1 O0
Slip, %

FIG. 1. Traction curves for an 18.4R38 tire in tilled Yolo loam soil with 85 kPa inflation pressure and
14.4 kN dynamic axle load.

mechanics, dimensional analysis and conservation of energy principles. They assumed

a 2D rectangular contact area in developing these expressions, In this work we extend
these expressions to a 2D elliptical contact area or reactangular contact area with
elliptical edges which occurs at high tire deflections [7]. Figure 2 shows the 2D
elliptical contact area (or rectangular contact area with curved edges).

Coefficient a
We have net traction N T given by
+b Yz
N T = f-b fyl (C + p t a n ( p ) ( 1 - e-i'(y-yO/K)dydx (3)
where C = soil cohesion;
q~ = soil internal angle of friction;
K = soil shear modulus;
p = average contact pressure;
i'~ = actual soil shear;
= Y - Yl;
i' = proportionality constant;
NT
-i
Ktl c'
lc = contact length;
i = slip;
K t = tire stiffness in the circumferential direction.
90 S. K. U P A D H Y A Y A and D. W U L F S O H N

TABLE 4. TRACTION TESI RESULTS

Inflation W
Soil (kPa) (kN) a c R~* a' b' c R~, TE~

1 85 14.42 0.7388 0.0770 0.954 0.8601 0.9254 0.0693 0.964 69.(/9

2 85 21.20 0.7426 0.0734 0.967 0.8439 0.9359 0.0727 0.967 68.(11
3 Moist 85 27.29 0.5939 0.1543 I).957 0.6909 0.9119 0.1442 0.961 69.90
4 tilled 125 14.97 0.5549 0.0697 0,942 0.7757 0.8498 0.0578 0.968 54.87
5 125 20.66 0.5468 0.1413 (11.965 0.7205 0.8673 0.1132 0.972 63.t9
6 125 27.ll 0.5275 0.1053 0.962 0.6973 0.8783 0.0885 0.971 6I .47

7 85 14.57 0.5760 0.1060 //.974 0.6966 0.9122 0.0946 0.970 67.26

8 85 14.87 0.5577 0.0985 0.983 0.7368 0.9201 0.0820 0.987 63.69
9 85 20.55 0.5681 0.1359 0.982 0.6675 0.9349 0.1222 0.985 70.74
10 85 21.01 0.5270 0.1893 0.995 0.6552 0.9162 0.1709 0.993 66.56
11 85 26.82 0.6068 0.1165 0.930 0.6787 0.9329 0.1160 0.919 71.03
12 Dry 85 27.06 0.5659 0.2099 0.982 0.6939 0.9257 0.1895 0.979 67.85
13 tilled 125 14.65 0.4830 0.0850 0.952 0.6505 0.9101 0.0681 0.968 62.56
14 125 14.70 (I.5386 0.0896 0.972 0.6822 0.9131 0.0780 0.972 64.34
15 125 20.76 0.5123 0.1300 0.959 0.6786 0.9382 0.1030 0.970 66.05
16 125 20.68 0.5197 0.1311 (I.981 0.7012 0.9176 0.0956 0.979 66.04
17 125 26.77 0.5046 0.1884 0.970 0.5997 0.9269 0.1695 0.972 70.00
18 125 26.76 0.5018 0.3436 (I.850 0.642l 0.9200 0.2769 (I.874 66.51

19 85 15.05 (I.534111 0.2360 t).991 0.6260 0.9493 0.2237 0.989 7(I.8(t

20 85 15.30 0.5531 0.235t 0,988 0.6253 0.9517 0.2344 0,986 72.64
21 85 21.02 0.6025 0.1698 0.986 0.6777 0.9482 0.1636 0.991 73.05
22 85 20.73 0.5498 0.2737 0.989 0.6195 0.9548 0.2677 0.991 73.41
23 Dry 85 27.09 0.5903 0.2098 0.981 0.6527 0.9591 0.2077 0.981 74.41
24 firm 85 26.70 0.6584 0.1548 0.988 0.7165 0.9666 0.1588 (I.990 74.55
25 125 14.97 0.678l (I.1242 0.963 0.7731 0.9583 0.1179 0.964 72.36
26 125 20.83 0.5654 0.2066 0.967 0.6474 0.9561 0.2048 0.967 71.74
27 t25 20.89 0.6040 0.1295 (I.972 0.6966 0.9596 0.1226 (1.974 71.73
28 125 27.21 (I.6128 (I.2586 0.962 0.6875 0.9669 0.2461 0.965 74.35
29 125 26.83 0.5550 0.1579 0.965 0.6393 0,9669 0.1476 0.966 72.65

*Net traction equation coefficient of determination.

t T o r q u e equation coefficient of determination.
SAverage T E over 0 to 30% slip range.

.'. N T = ( C + p tan -h y + i' e"°'-"')/a dx

/lYl

= (C+ ptanq0) l b (Y2 - Y i ) + - - ( e-c(~':-v'}i~ - 1) dx. (4)

- l

A l t h o u g h equation (4) cannot be integrated explicitly, w e can l o o k at a limiting

case, as i--~ 1. In this case, i'---, 1 and N T approaches saturation value. T h e r e f o r e ,
the e x p o n e n t i a l term should die out. Thus, as i - ~ 1, w e have i' ---, 1
/ 4 '+b

NT ~ ( C + ptanq~)J_h [(Y2 - Yl) - K] dx

NT = ( C + p tan q~)[Ac - Klw]

= rmaxAc- t-maxK/w (5)
where lw = contact width;
C + p tan q~ = m a x i m u m shear stress:
Tma x ~
Ac = actual contact area.
 TRACTION PREDICTION USING SOIL PARAMETERS 91

Fro. 2. 2D Soil tire contact area.

From our traction test as slip approaches 100%,

NT
W
= a or NT= aW. (6)

Equating equations (5) and (6), we get

aW = rmaxA c - l"maxK/w. (7)
The results of regression analysis of our experimental results based on equation (7)
yielded the following expression.
aW = 6.675 + 0.952Ac~rmax - 19.208rmaxKlw. (8)
The coefficient of determination for this equation was 0.82 and both the parameters
(Acl"max) and (~'maxg/w) were highly significant (Fig. 3). The presence of the intercept
indicates that most likely aW
is a non-linear function of the parameters selected. The
linearized equation (8) is good in the range in which the experimental variables were
varied. In this equation W is in kN, Ac is in m 2, Z'max is in kPa, K and lw are in m.

Coefficientc
From equation (4), if i is very small, i' will also be small and we can expand
in a Taylor series. This expansion gives us
e -i'(y2-y~)/K

NT= (C+
[
/5 tancp) Ac +
f+_b-g[
- 1
i'(y2-yl) +
i'2(y2-ya)2 + -1
]dx
b i' K 2K 2 "'"

-~ (C + /~tanq~) Ac - _ (Y2 - yl)dx + ~--~ f_b (y2 - yl)2dx

[
-- (C + /3 tan ~p) A ¢ - A ¢ + ~
i'f+ bb4B2
7 (A2-
]
x2)dx .

This is because the elliptic contact area gives

X2 y2
~+~-g=l (9)
92 S.K. UPADHYAYA and D. WULFSOHN

2o

18" •

16 " OO

14"

12 •

10"

6 • , • v • , , , • , - , •

6 8 10 12 14 16 18 20
Fitted a W

FIG. 3. V a r i a t i o n o f e x p e r i m e n t a l l y o b t a i n e d a W as a f u n c t i o n o f p r e d i c t e d aW,

or

B
Y2,1 = + - - V 'A2 - x2
A
2B
Y 2 - YJ = ~ ~v/A2 - x2

where 2B = lc = contact length

A = minor axis of the ellipse. This can be calculated from tire
deformation.

• NT= (C+ ptanq0~-i' 1-

7
2B2 [ 263]
= (C + p t a n ~ ) ~ - i' 2b - 3A2]

2B2 [ b2]
= (C+ ptan~)~lw 1- -3A
- i 2'

b 2
= + --i'

We can show that

 TRACTION PREDICTION USING SOIL PARAMETERS 93

A c oc Iclw

pXAc= 1-

[lc~
•". NT = /9(C + p tan Jp)[-KJA~i (10)
where p is proportionality constant.
But when i << 1, our experimental result gives
NT -- acWi. (11)
From equations (10) and (11)

acWi = o(C + P tan ~) -~- Ac i -

(/c)[ acW1
= p ( C + p tan tp) -~- A c 1 - ~-~t/c]t

c ac l
Kt/c]"
Solving for acW, we get

"'axAc()
ac = 1 + p~'maxAc -~- t
Inverting equation (12) and simplifying, we obtain
1 1(-)()
- a P----P--- K + _ (13)
C P "t'max lc Ktlc"
A regression based on equation (13) resulted in a coefficient of multiple determina-
tion of 0.79. The first coefficient on the right was very highly significant (better than
1% level). The second coefficient on the right marginally significant (about 13%).
The regression equation has the form:
1__ = -4.682 + 413.067[a/P-L--](--r]]
+ 41.377( a w l . (14)
c [ krmax]\l¢]] \Ktlc]
Figure 4 shows the variation of predicted (l/c) versus experimentally obtained (l/c).
In equation (14), p is in kPa and g t is in kN/m.

Coefficient c'
We expect c' to be the same as c. Figure 5 shows the plot of experimentally
determined c versus c'. This equation resulted in a coefficient of determination of
0.98. The regression equation is
c' -- 1.0213c - 0.0152. (15)
For all practical purposes, we can assume c and c' are one and the same.
94 S.K. UPADHYAYA and D. WULFSOHN

16

14-

t2

10

O0 • •

00

4 6 8 I0 12 14

Fitted 1/c

FIG. 4. Variation of experimentally determined (l/c) versus predicted (l/c).

0.30 "

y= .0.0

0.25

0.20

0.15 "

0.10

0.05
0.05 0.10 0.15 0.20 0.25 0.30

FIG. 5. Variation of experimentally determined c' versus experimentally determined c.

Coefficient a' and b'

U p a d h y a y a et al. [3] used a n e n e r g y principle to relate a ' a n d b ' to soil a n d tire
p a r a m e t e r s . F r o m their w o r k , we have
 TRACTION PREDICTION USING SOIL PARAMETERS 95

Total input energy = T 0

Tic
(16a)
r
Output energy = NT(le - A) (16b)
where 0 = tire rotation as it traverses the contact length;
A = loss of forward motion due to deformation of both soil and tire.
Strain energy related to normal and tangential deformation of tire

1
1 Ktr/2 + gn62

1 (NT) 2 1
(16c)
--2 K t + 2 w(~t
where K n = tire stiffness in the vertical direction;
r/t = tangential tire deformation;
~t = vertical tire deformation.
Energy dissipated in compressing the soil
n
= dz

n+a
k vs _ __k n6s
(b) n n + 1 n + 1

W 6s
(16d)
n+l
where k = soil sinkage constant;
n = exponent in the sinkage equation;
6s = vertical soil deformation.
Energy dissipated in shear deformation of soil
J
= fA fo['rdj]
dA

= fAl;max[j + K ( e -]/K - 1)] dA

= fArmaxjdA - K r m a x f A ( 1 - e-J/K)d A

= fA rmaxj dA - KNT
96 S. K. U P A D H Y A Y A and D. W U L F S O H N

because

N T = faTmax(1 -- e i / K ) d A .
Consider

fa
rmaxJ dA = f ~ b ( t ( ~ ) rmaxi' y d x d y
h 3O

(because j = i ' y )
where
2B
l ( x ) = Y2 - Yl = ~ ~v/ A 2 - x 2

• f4Vmaxi' [/(X)]2 dx
• 2

= s;i 'ma ir4

- 2 [~- 1
(A2 - x 2 ) d x

= 2rmaxB2i ' 1 -- dx
-b

= 2rma×B2i ' 2b - [ 3A2]

= PTmaxAclci'
where
/c = 2 B
lw = 2b
/c/w[ b2 ]
pAc - 2 1 - ~A T .

.'. E n e r g y dissipated in shear d e f o r m a t i o n of soil is:

= PTmaxAclci' -- K N T . (16e)
In addition, we should consider energy dissipated due to slip sinkage. This part was
neglected in our previous analysis. If 6smax is the total sinkage of soil, then 6~ ax - 6s
can be attributed to slip sinkage. Since this sinkage occurs at a constant axle load W,
energy dissipated due to slip sinkage is given by
= W ( 6 ~ ax - as). (16f)
F r o m equations (16a) through (16f), we obtain
1 NT 2 W [
Tlc = N T ( / c - A) + - - - + [6 t + -n -+ 1 + 2(6smaX - 6s)
r 2 Kt 2-
+ PrmaxAclci' - KNT (17)
 TRACTION PREDICTION USING SOIL PARAMETERS 97

A
because i = - -
lc
ifi=0, NT=0, i' = 0 , 6 max= 6s

T W 5t + (18)
r - 21¢ --~--1-]"
From experimental results it follows that as i ~ 0,

T
rW = a'(1 - b')

or
T
--= a'(1 - b')W (19)
r
w s]
.. a'(1 - b')W = + n + 1]" (20)

If the wheel slips 100%, then experimental results give us

T
--
r
-- a'W
NT = aW.
Therefore, from equation (17), we get

a'W = 0 + c("Ktl--
azW ) 2+X'z-'
Wztc
[ 6t +
25max 2n
n+l 6s
]

+ PVmaxAc[1- a&] KaWlc (21)

Since we do not have slip sinkage information for all our tests, we decided to ignore
the slip sinkage phenomena in this analysis. Therefore, 5smax will be assumed equal to
~s.
A multiple regression model based on equation (21) resulted in the following
equation:

a'W = 8.204 + 10.066 (aW)Z + 0.273 [ ~t + --""Y'-'T-~

26s ]| W
K tl------
~ t n -f- I ]

+ 0-062Aevmax- 19.496(~cW ). (22)

This equation resulted in a coefficient of multiple determination of 0.962. However,

the coefficient 0.273 for

5t + W
n+l
and 0.062 for Acrmax were only marginally significant [probabilities of 0.16 and 0.15
98 S.K. UPADHYAYA and D. WULFSOHN

respectively]. The other two terms were highly significant. If we ignore these terms
the following equation results.

a ' W = 9 . 0 2 1 - 18.88(a/~K)

W 2
+ 10943[ (aW)~ ] (23)
• t Kdc ]
Equation (23) resulted in a coefficient of multiple determination of 0.921 (Fig. 6).
Both the variables were highly significant. Once again, the presence of an intercept
indicates that the relationship is non-linear• The linearized equation (23) is valid in
the range tested. The unit of lc in equation (23) is in m.
A regression based on equation (20) in which we eliminated W from both sides,
and allowed for higher order terms related to
c~t 0s
- - and
lc (n + 1)lc
yielded:
a'(1 - b') = 0.522 - 7.657xl + 5.84x2
+ 29.243x~ - 20.861x~
- 31.05x~x2 (24)
where

XI ~ ~c

22

20 ¸

• O0 00
18

16

12

10"~ •

8 ' ' '

l0 12 14 16 18 20 22
Fitted a'W
F ~ c . 6. V a r i a t i o n of experimentally determined a'W versus predicted a'W.
 T R A C T I O N P R E D I C T I O N U S I N G SOIL P A R A M E T E R S 99

and

x2 - (n + 1)lc"
This equation resulted in a coefficient of multiple determination of 0.839 and all the
coefficients were highly significant (Fig. 7) except the coefficient related to the term
Xx which was significant at 5% level• In equation (24) S t and 6s are expressed in m.
The results presented in this paper indicate that it is possible to predict traction
using soil shear and sinkage parameters obtained using an instrumented analog
device. We plan to conduct additional experiments in several different soil conditions
to validate the methodology outlined in this paper.

CONCLUSIONS
We reached the following conclusions:
Using traction mechanics, dimensional analysis and energy principles, we have
shown that the following equations relate a, c, a', b' and c' to soil and tire
parameters and loading conditions:
aW = 6.675 + 0.952Acrmax - 19.208rmaxK/w

+ ,Vmax]\ic]
1 = c - 4 . 6 8 2 + 413.067[a( P----~--/(Kt]
[ ] 41.377[ a~--~t~c]

• Lgclcj
a'(1 - b') = 0.921 - 13.269x I + 4.681x2
+ 29.243x~ - 20.861x2z - 31.05XlX2

0.12

0.10" • •

0.08
. i

"~ 0.06 •

0.04 S
•

0.02
0.02 0.04 0.06 0.;8 0.'10 0.12
Fitted a' (I - b')
FIG. 7. Variation of experimentally determined a'(1 - b ' ) versus predicted a'(1 - b').
100 S. K. U P A D H Y A Y A and D. W U L F S O H N

where

t~t t~s
X1 : ~- and x~ - lc
t~ n+l
and
C t ~ C.

T h e tire tractive characteristics can be expressed using these parameters as:

Net traction coefficient -N~T- = a(1 - e_Ci )

Gross traction coefficient T = a'(1

~-~ -
b,e_c,i)

and

Tractive e f f i c i e n c y -: -(~-wT)
- (1 - i),

REFERENCES
[1] W. R. GILL and G. E. VANDEN BERG, Soil Dynamics in Tillage and Traction. Agr. Handbook No. 316.
U.S. Government Printing Office, Washington, D.C. (1968).
[2] R. N. YONG, E. A. FATTAH and N. SKIADAS, Vehicle Traction Mechanics; Developments in Agricultural
Engineering, Vol. 3. Elsevier Science, Amsterdam, Netherlands (1984).
[3] S. K. UPADHYAYA, D. WULFSOHN and G. JUBBAL, Traction prediction equations for radial ply tyres. J.
Terramechanics 26 (2), 149-175 (1989).
[4] S. K. UPADHYAYA and D. WULFSOHN, Review of traction prediction equations. A S A E Paper No.
90-1573. A S A E , St. Joseph, MI 49085 (1990).
[5] S. K. UPADHYAYA, D. WULFSOHN and J. MEHLSCHAU, An instrumented device to obtain traction
related parameters. J. Terramechanics 30 (1), 1-20 (1993).
[6] S. K. UPADHYAYA, D. WULFSOHN and J. MEHLSCHAU, An instrumented device to obtain traction
related parameters. A S A E Paper No. 90-1097, A S A E , St. Joseph, MI 49085 (1990).
[7] S. K. UPADHYAYA and D. WULFSOHN, Relationship between tire deflection characteristics and 2-D tire
contact area. Trans. ASAE 33 (1), 25-30 (1990).