Professional Documents
Culture Documents
ABSTRACT
Effort has been made to develop a software for tractor design. It provides flexibility to incorporate the
coefficient related to component material, system type and working environment. Basic design equations have
been used to obtain the dimensions of various systems of the tractor. Eleven different models of tractor in 18 to
50 hp range have been considered for validation of the computer software. The actual and designed values for
the engine displacement, cylinder bore and stroke, fuel tank capacity, clutch plate diameter, rear wheel diameter
and wi dth have been compared. Other design dimensions could not be compared due to lack of needed
inform ation. The comparison indicates that the design dimensions are in close agreement with the actual
existing dimensions.
(31)
using computer has been presented in this In a similar way the second part of the
paper to develop a computer software for software contains various tractor system
the design of tractors in the size range of 15 d esign modules such as clutch ,
to 150 hp. transmission, brake, hydraulic, operator's
2. TRACTOR DESIGN seat, wheel and steering.
(32)
Knowledge b a se
no
no
(33)
Table 1. Inputs and outputs of design modules
S.No. Design module Input Output
~
1. Engine a) Brake power, Engine
displacement b) Per cent overload power, displacement
c) Mass of air required for combustion
of one kg of diesel,
d) Lower heat value of diesel,
e) Engine ~overned speed,
2. Cylinder
bore and stroke
!t) Mechamcal efficiency,
g) Thermal efficiency,
h) Volumetric efficiency.
(a) Engine displacement,
(b) Stroke-bore ratio.
Cylinder bore
and stroke
3. Fuel tank (a) Specific diesel Fuel tank
capacity consumption, capacity
~
b) Brake power,
c) Refilling time,
d) Overhead allowance.
4. Clutch plate al Engine governed speed, clutch plate
5.
diameter
Rear wheel
!b Engine brake power,
c Friction coefficient.
d) pressure between the
friction surfaces.
a) Load on rear wheel,
inner and
outer dia.
wheel
diameter and
width
!b) permissible slip,
c Surface cone index ran~e,
d) Minimum tractive efficiency desired.
diameter and
width
(34)
Oy_ KT = (0.4 to 0.6) . Ka (22)
Sya = and (11)
(oa +Om )
The total safety factor of the part jointly
~=
• (12) affected by tangential and normal stresses
(<a+ Tm) could be obtained as given below.
However, in case the effects of stress S = ST.Sa
concentrations, dimensions and the quality (23)
of surface finish are considered for the
vS¥ + s&
design, the maximum stress is given as Irrespective of the method used for
computation safety factor, it should be
Oa.Ka
Omax = es . ess
+ Om (13) within permissible limits depending upon
the component. In case it lies beyond the
Ta. Ka limits, the component dimension should be
or Tmax =- --
es. ess
+To) (14)
changed as per pre-determined steps and
and the corresponding safety factors safety factor be recomputed. The process is
considering the fatigue limit would become iteratively repeated until converge nce
occurs (Fig.2).
0 -1
Sa = Oa,c + rx. a .rx.m
(15) 3.3. Optimization technique
T- 1 For optimization of tractor system
ST = Ta , c + rx.T . Tm
(16)
component design, standard procedure
when computing by yield limit, the safety was follwed as cited by Motz et al., (1990) .
factor is given as T here are three terms common to all
optimization algorithms namely, the design
Oy_
Sya = (17) vari able, the state variables and the
Oa,c +Om
objective function. A design variable is a
T
~ = Ta , c+ Tm (18) parameter whose value is allowed to
change through successive iterations during
Where the optimization process. Generally, the
Oa. Ka design variables are geometric parameters
Oa,c = and (19) used to define the part. State variables are
es . ess
Ta. KT
the response of that part to the changes in
Ta , c = (20) the design variables. State variable may
es. ess
include stresses, deflections and the natural
He re Ka is the e ffective stress frequencies. A range of acceptable values
concentration factor and is expressed by must be defined for each of the state
th e following based on approximate va ri ables which will determine the
relationship (Kolchin et al., 1984). feasibility of the design. Any parameter
Ka = 1 + q . ( rxca- 1) (21) that can be quantitatively described could
be the objective function. Commonly used
The relationship between factor Ka and
objective functions are the weight, cost or
KT can be expressed by the following
volume of the material used. The only
based on experimental dat'a available.
Fig. 2 Simplified flow chart for failure and fatigue analysis
(36)
search technique was used. This technique S. RESULTS AND DISCUSSION
uses a series of unidirectional search along Table 5 shows the results of the
each design variable within its feasible range statistical comparison made between actual
and process is iteratively repeated until a nd the designed dimensions of the
convergence or termination takes place. different tractors under consideration.
After the optimum objective function It may be seen that there is an excellent
has been found, the new design variables agreement between actual and designed
are calculated. The best design is the one in d imensions of engine displacement
which the lowest positive value of the (Fig.3a), cylinder bore (Fig.3b ), stroke
objective function is obtained. (Fig.3c) and rear wheel diameter (Fig. 4b)
4. VALIDATION a nd width (Fig.4c). The correlation
One of the requirements of the software coefficient for these dimensions are in the
was its applicability across wide range of range 0.923 to 0.976. The relative errors for
tractors, with only the change of input data. these dimensions varied from 2.90 to 9.68%
Accordingly, evidence is reported here of and the intercept of the linear regression
the software applicability to eleven models were found to be non zero. The ralative
of the tractors. The validation of the error and non zero linear regression
software involved comparison of the actual intercept could be due to the difference in
and designed data and computation of the values of coefficient and safety factors
relative errors. A linear,regression analysis used in the design calculations for different
was also conducted on the data available d imensions and actual value used by
fo r the tractors. The actual data for manufacturers. However in case of fuel
comparison as obtained from available tank capacity (Fig. 3d) and diameter of
information on these tractors were used as cl utch plate (Fig. 4a) correlation
input data, in the software and have been coefficient between designed and actual
presented in Table-3. . values are less (0.854 and 0.704) and
re lative error high (18.02 and 9.6%)
The output data from the softwar·e re spectively. The difference may be
indicated the designed dimensions of the anticipated in case of fuel tank capacity
components. However only some of the d ue to variation in specific diesel
parameters could be compared e.g. engine consumption of tractor engines and tank
displacement, cylinder bore and stroke, being designed for different period of
fuel tank capacity, clutch plate diameter operation by different manufacturers. The
and rear wheel diameter and width. The variation in clutch plate diameters,
other component dimensions could not be however could not be ascertained. A
compared due to unavailability of actual possible explanation to this may be the
data. The actual data and the designed concept adopted for the clutch plate
data as computed using the software are d esign. In the present case uniform
presented in Table 4. pressure concept has been assumed and
the friction coefficient used was 03 in each
case.
(37)
Table 3. Tractor data of eleven different models
Tractor Brake E ngine No. of Compression Stroke
model power [hp) speed [rpm) cylinder ratio bore ratio
Mitsubishi
MT 180 D 18.0 2700 3 17.0 1.114
Zetor 2511 23.80 2000 2 17.0 1.16
Swaraj 725 23.60 2000 2 17.0 1.10
Eicher 24.00 1650 1 17.7 1.30
Harsha T25 25.00 1800 2 16.0 1.14
IH B-275 30.76 1898 4 21.1 1.19
MF 1035 32.00 2200 3 16.5 1.43
Escort 335 35.00 2200 2 17.0 1.18
Kirloskar 04006 K 43.00 2300 3 17.0 1.20
Ford 3600 47.00 2000 3 16.5 1.00
Hindustan 50 50.00 1650 4 15.6 1.14
(38)
Table 5. Statistical comparison between designed and actual dimensions
Dimensions Regression coefficient Correlation Relative
a b coefficient error,%
E ngine
displacement 33.316 1.033 0.996· 9.68
Bore 16.439 0.841 0.970 2.90
Stroke 16.489 0.867 0.976 3.19
Fuel tank
capacity 4.899 0.854 0.841 18.02
Dia. of
clutch plate 12.179 0.962 0.704 9.67
Rear wheel
diameter 20.701 0.891 0.969 ' 3.02
Rear wheel
width 1.990 1.108 0.923 6.16
6. CONCLUSIONS llla mass of air required for combustion
of one diesel, kg
The software developed is applicable for
a wide range of tractors. It was tested for 11c number of cylinder in engine
t he prediction of component design q coefficient of material sensitivity
dimensions of eleven different models of rsb stroke bore ratio
Indian tractors. Out of the seven predicted rpm rev/min
d imensions, compared with actual
Q:. a coefficient of reducing an
d imensions, five (namely engine asymmetric cycle to the
displacement, cylinder bore and stroke and aquidangerous symmetric cycle
rear wheel diameter and width) were found under normal stress
to be closely correlated. The correlation
Q:.T coefficient of reducing an
coefficient for these dimension were in the
asymmetric cycle to the
range 0.923 to 0.976 while relative error aquidangerous symmetric cycle
varied from 2.90 to 9.68%. However in case under tangential stress
of fuel tank capacity and clutch plate
Q:. ca theoretical stress concentration
diameter correlation coefficient between
factor
designed and actual values were less (0.854
and 0.704) and relative error high (18.02
fl. ratio of the fatigue limit due to
torsional stress to yield limit
and 9.6%) respectively.
Oa cycle amplitude normal stress, MPa
Nomenclature
Om mean normal stress, MPa
es scale factor a- 1 fatigue limit for bending, MPa
ess surface sensitivity factor Omax cycle maximum normal stress, MPa
hu lower heat value of disel,kJ!kg r -1 fatigue limit for torsional stress,
MPa
(39)
- - - ZERO ERROR LINE - - - ZERO ERROR liNE
E 0
600 0 - - FITTED LINE E 120 - - FITTED LINE
M
E UJ
0 0::
0
0.: CD
l/)
5 400 0 ffi0 10 0
UJ
z z
t.:> ~
z >-
u
UJ
fiJ200 0 fil 80
z z
t.:> t.:>
Vi l/)
UJ
UJ
0 0
6000
ACTUAL ENGINE DISP., cm3 ACTUAL ENGINE BORE, mm
(a) (b)
;::.. 8 0
ZERO ERROR LINE 1- - - - . ZER:l ERROR LINE
18 0 -- FITTED LINE
/
/
~ u
~
- ·- FITTED LINE o
E / < 0
• E / u
ui' / X:
~
X: z
""
0
0::
1- 14 0 v ~ 0
l/) ~
0
w 0
0 :::>
UJ IJ..
z 0
t.:>
0
Vi 10 0 w
UJ
0 z 0 ""/,
\:> /.
iii 2 0
ERROR = 3.1 'lo w
0 /~RELATIVE ERROR =1 8. 08 'lo
180 20 60 4080
ACTUAL STROKE, mm ACTUAL FUEL TANK CAPACITY, I
(c l (d)
Fig.3 Comparison of actual and designed data of (a) engine displacement, (b) cylinder
bore, (c) stroke and (d) fuel tank capacity
(40)
180
E ZERO ERROR LINE ZERO ERROR LINE
E
f: 350 FITTED LINE u FITTED LINE
<(
<(
5
UJ
0 140
~
0
<(
UJ
..J
cr
Cl. >-
~
:I:
u
~ 0
::::>
..J
u 250 0
UJ
0
UJ z
C)
z
C)
0
Ill 60
iii RELATIVE ERROR= 9-67"/o UJ
0
~ 200
250 300 350 60 100 140 180
ACTUAL CLUTCH PLATE DIA.mm ACTUAL REAR TYRE DIA . em
Ia) I bl
cr
<(
UJ
a::
0
UJ
z
C)
iii ~ RELATIVE ERROR= 6.16%
UJ
0 /
0 60
ACTUAL REAR TYRE WIDTH c m
I J c
Fig.4 Comparison of actual and designed data of (a) clutch plate diameter, (b) rear tyre
diameter and (c) rear tyre width
(41)
cycle amplitude tangential stress, De Vries, R.I., Aadziwon, H.V., and
MPa Aghssa, P., 1987. The computer aided
mean tangential stress, MPa design analysis accessory components.
SAE paper No. 870872. Warrendale, PA.
nm mechanical efficiency
Kolchio, A. and Demidov, V. 1984. Design
nt thermal efficiency of Automotive Engine. Mir publishers.
nu volumetric efficiency Moscow.
(42)