Professional Documents
Culture Documents
CHAPTER 7
7.1 INTRODUCTION
diesel engine using jatropha biodiesel as fuel. Jinlin Xue et al. (2011)
predicted the performance and exhaust emissions of blended fuels using
artificial neural network (ANN). Celik et al. (2005) observed that many
research about optimization and modification on engine and methodology for
measurements should be performed when petroleum diesel is substituted
completely by biodiesel. Many researchers indicated the need of research in
the areas of engine modifications so as to suit to higher blends without severe
drop in performance so that the renewability advantages along with emission
reduction can be harnessed to a greater extent.
The first part of the chapter deals with the optimal selection of
these parameters based on sixteen test runs conducted using S/N ratio of
taguchi method. Next part of this study using multi object optimization, the
optimal combination was found out to minimize the BSFC, emissions and to
maximize the BTE. The second part deals with optimization of various
parameters using RSM. Third part of chapter deals with multiple regressions
to find out the equation to predict the emission and performance.From the
empirical model developed to find out the responses a C++ PROGRAM was
written to find an optimal solution
152
7.2 OBJECTIVE
1 r 2
S / N = −10 log ∑ y i (7.1)
r i=1
1 r
S / N = −10 log ∑1 / y i2 (7.2)
r i=1
The negative sign is used to ensure that the largest value gives an
optimum value for the response variable and therefore robust design.
x
OPT = T + ∑
i =1
(X i − T ) (7.3)
Where T is the overall mean value of the output response variable for the test
runs conducted. Xi is the design and control parameter value for the I level of
the parameter X.
The first step was the selection of the parameters that influence the
performance and emission characteristics. In this study, the compression ratio,
injection pressure, injection timing, fuel blends and power were considered as
the input parameters.
As per the run order, the experiments were conducted on the engine
and the responses were fed on the responses column.
BTE BSFC
Factors CO HC NOx
Run Number S/N ratio
1 30.63 12 -27.95 -29.54 -63.90
2 30.37 11.37 -12.04 -35.56 -62.27
3 28.94 10.17 -6.19 -37.38 -60.48
4 28.63 9.37 -4.4 -46.02 -56.77
5 27.60 8.17 -1.93 -33.06 -64.11
6 29.25 10.17 -5.68 -33.25 -62.27
7 30.63 11.70 -9.89 -39.27 -60.99
8 30.37 11.37 -33.97 -37.26 -55.95
9 29.54 10.17 -15.91 -32.04 -61.99
10 28.30 8.63 -27.95 -33.97 -60.83
11 30.63 11.37 -2.85 -41.80 -61.08
12 30.88 11.70 -17.72 -42.54 -56.41
13 29.54 10.45 -19.17 -30.37 -64.71
14 29.83 10.45 -4.4 -40.00 -62.27
15 29.25 9.89 -10.45 -32.25 -58.06
16 28.63 9.11 -4.58 -48.30 -55.91
156
31.5
31
30.5
30
S/N ratio
29.5
29
28.5
28
27.5
27
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Experimental runs
9
8
7
6
5
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Experimental runs
Experimental runs
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
0
-5
-10
-15
S/N ratio
-20
-25
-30
Experimental runs
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
-25
-30
-35
S/N ratio
-40
-45
S/N ratio of HC
-50
Ex perimental runs
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
-55
-56
-57
-58
-59
S/N ratio
-60
-61
-62
-63
-64
-65
S/N ratio of NOx
-66
For the engine performance, the response variable BTE was higher-
the –better. The criteria for optimization of the response parameters was based
on the higher the better S/N ratio.
where Y bar is the average value of BTE from the experiments and CR, IP, IT,
B and P are the parameters chosen for the experiment.
30.0
29.5
Mean of SN ratios
29.0
28.5
30.0
29.5
29.0
28.5
Table 7.2 The Response table for BTE in % from the experimental
results
Static Biodiesel
Average Compression Injection
Injection fuel Power
S.No. Value ratio Timing
fraction
Pressure (P)
Y bar (CR) (IP) (IT)
(B)
1 30.5 32 31 33.5 30.5
2 30 33 30.25 32.5 33
30.25
3 31.25 31 30.5 26.25 28.5
4 29.25 30.5 29.25 28.25 29
10.8
10.2
Mean of SN ratios
9.6
9.0
17.5 17.7 17.9 18.1 190 210 220 230 23 25 27 29
D E
11.4
10.8
10.2
9.6
9.0
10 20 30 50 3.64 4.16 4.68 5.20
Table 7.3 The Response table for BSFC in kg/kWh from the
experimental results
Biodiesel
Compression Static Injection
Average Injection Timing fuel Power
S.No. ratio
Y bar Pressure fraction (P)
(CR) (IP) (IT)
(B)
1 0.295 0.3137 0.29 0.267 0.3025
2 0.31 0.315 0.31 0.278 0.276
0.3053
3 0.29 0.29 0.29 0.355 0.317
4 0.3175 0.30 0.317 0.32 0.325
Figure 7.8. From the value the optimization of engine parameters obtained
were,CR-17.9, IP-190 bar, IT-23° Btdc, Blend –B30, Power 5.2 kW that
predicts minimum HC.
The value of HC is calculated from the response table shown in Table 7.4
-35
Mean of SN ratios
-40
-45
17.5 17.7 17.9 18.1 190 210 220 230 23 25 27 29
D E
-30
-35
-40
-45
10 20 30 50 3.64 4.16 4.68 5.20
Table 7.4 The Response table for HC in PPM from the experimental
results
Biodiesel
Compression Static Injection
Average Injection Timing fuel Power
S.No. ratio
Y bar Pressure fraction (P)
(CR) (IP) (IT)
(B)
1 91 37 114 89 49
2 64 64 70 73 81.5
123
3 87 83 72 107 103
4 109 167 94 87 117
The value of NOx is calculated from the response table shown in Table 7.5
-58
-60
Mean of SN ratios
-62
-64
17.5 17.7 17.9 18.1 190 210 220 230 23 25 27 29
C4 C5
-56
-58
-60
-62
-64
10 20 30 50 3.64 4.16 4.68 5.20
Table 7.5 The Response table for NOX in PPM from the experimental
results
Biodiesel
Compression Static Injection
Average Injection Timing fuel Power
S.No ratio
Y bar Pressure fraction (P)
(CR) (IP) (IT)
(B)
1 1154 1539 1157 1164 1024
2 1163 1250 1092 1195 1085
1117
3 1038 1190 1060 1097 1175
4 1111 651 1158 1012 1182
7.4.2.5 CO Emissions
Table 7.6 The Response table for CO in % from the experimental results
(7.11)
Response Weighing
S.No.
variable Factor
1 BTE 0.2
2 BSFC 0.2
3 NOx 0.2
4 HC 0.2
5. CO 0.2
168
The sum of the weighing factors for all the response variables
should be unity and the weighting factor assigned to each particular response
variable is determined on the basis of its relative importance. In this present
experimental work the weighting factor for each response variable assumed as
given in Table 7.8 .The weighting factor plays a very important role in this
type of analysis. The experimental results were used to Calculate S/N ratios
for all the response variables.
S / N = W 1SN 1 + W 2 SN 2 + W 3 SN 3 + W 4 SN 4 + W 5 SN 5 (7.12)
Where, W1, W2, W3, W4 and W5 are the weighing factor of the response and
SN1,SN2, SN3, SN4 and SN5 are the signal to noise ratio of the response
variable. Using equation (7.12) S/N ratio for the combined objective is found
out and it is given in Table 7.9 and shown in Figure 7.10.
For the combined objective with equal weightage for all the
responses, the optimum setting is found out, CR-17.9, IP-200 bar, IT-23°
bTDC, B20. P -3.64kW is the optimal combination to achieve multiple–
performance characteristics of the engine.
Response
S.No. Values
variable
1 BTE 34.75 %
2 BSFC 0.2393 kg/kW h
3 NOx 598 PPM
4 HC 22 PPM
5 CO 0.015 %
170
In this study, RSM was applied to optimize the input parameter that
gives maximum efficiency, minimum fuel consumption and emissions.
Fuel
Run Compression Pressure Injection Blends Power BTE BSFC CO HC NOx
Order Ratio ο
bar Btdc % kW % kg/kWh % ppm ppm
1 17.5 190 23 10 3.64 34.84 0.24 0.22 69 1258
2 17.5 210 25 10 4.16 35.42 0.23 0.42 69 1319
3 17.5 220 27 10 4.68 34.94 0.24 1.11 70 1335
4 17.5 230 29 10 5.2 33.41 0.26 2.28 71 1306
5 17.5 190 23 20 3.64 31.31 0.28 0.14 69 1239
6 17.5 210 25 20 4.16 31.63 0.27 0.3 69 1282
7 17.5 220 27 20 4.68 30.89 0.28 0.84 74 1279
8 17.5 230 29 20 5.2 29.10 0.31 1.76 84 1232
9 17.5 190 23 30 3.64 29.45 0.30 0.11 69 1206
10 17.5 210 25 30 4.16 29.50 0.29 0.25 67 1230
11 17.5 220 27 30 4.68 28.50 0.31 0.72 81 1209
12 17.5 230 29 30 5.2 26.45 0.34 1.53 112 1143
13 17.5 190 23 50 3.64 30.68 0.27 0.17 67 1097
14 17.5 210 25 50 4.16 30.22 0.28 0.34 61 1084
15 17.5 220 27 50 4.68 28.69 0.30 0.93 110 1027
16 17.5 230 29 50 5.2 26.12 0.34 1.95 214 924
17 17.7 190 23 10 3.64 34.38 0.26 0.12 69 1061
18 17.7 210 25 10 4.16 35.28 0.24 0.26 70 1173
19 17.7 220 27 10 4.68 35.12 0.25 0.74 69 1241
20 17.7 230 29 10 5.2 33.91 0.27 1.58 68 1264
21 17.7 190 23 20 3.64 30.97 0.30 0.06 70 1075
22 17.7 210 25 20 4.16 31.60 0.28 0.17 70 1169
23 17.7 220 27 20 4.68 31.19 0.29 0.55 67 1218
24 17.7 230 29 20 5.2 29.72 0.32 1.20 62 1223
25 17.7 190 23 30 3.64 29.21 0.31 0.04 70 1074
26 17.7 210 25 30 4.16 29.59 0.31 0.13 71 1150
27 17.7 220 27 30 4.68 28.91 0.32 0.47 63 1181
28 17.7 230 29 30 5.2 27.18 0.35 1.04 47 1167
29 17.7 190 23 50 3.64 30.67 0.29 0.08 70 1031
30 17.7 210 25 50 4.16 30.53 0.29 0.20 74 1070
31 17.7 220 27 50 4.68 29.32 0.31 0.62 48 1064
32 17.7 230 29 50 5.2 27.07 0.34 1.34 73 1013
33 17.9 190 23 10 3.64 33.17 0.27 0.10 69 883
34 17.9 210 25 10 4.16 34.39 0.26 0.23 69 1048
35 17.9 220 27 10 4.68 34.56 0.25 0.68 72 1167
36 17.9 230 29 10 5.2 33.67 0.27 1.45 78 1242
37 17.9 190 23 20 3.64 29.87 0.31 0.05 69 929
38 17.9 210 25 20 4.16 30.83 0.30 0.15 66 1076
39 17.9 220 27 20 4.68 30.74 0.30 0.05 87 1177
40 17.9 230 29 20 5.2 29.59 0.32 1.10 133 1234
41 17.9 190 23 30 3.64 28.23 0.33 0.02 67 962
172
Fuel
Run Compression Pressure Injection Blends Power BTE BSFC CO HC NOx
Order Ratio
bar Btdc % kW % kg/kWh % ppm ppm
42 17.9 210 25 30 4.16 28.93 0.35 0.11 59 1090
43 17.9 220 27 30 4.68 28.57 0.31 0.42 122 1173
44 17.9 230 29 30 5.2 27.16 0.30 0.95 256 1211
found to be significant as the values of p were less than 0.05. The different
models for the responses were developed in terms of actual factors and are
given below as equations (7.13) to (7.17) and shown in table
It can be seen that the regression coefficients of the linear terms for
compression ratio, injection pressure, injection timing and blend the quadratic
terms in IP2 and B2, and the interaction terms in CR*IP,CR*IT and IP*IT had
significant effects on the yield ( P value <0.05).
175
Model equation for carbon monoxide (CO): The predicted model for CO is
CO=-0.078-0.092CR+0.76IP-0.078IT-0.071B+0.098CR*IP
Model equation for Nitrogen oxides (NOx): The predicted model for NOx
is
factors. The (Adjusted R2 ) value accounts for the number of predictors in the
model. Both the values indicate that the model fits the data very well.
Model BSFC
BTE % CO % HC PPM NOx PPM
kg/kWh
Mean 33.85 0.268 0.098 31 884
Std.Deviation 0.036 0.030 0.158 0.035 0.27
Model degree Quadratic Quadratic Quadratic Quadratic Quadratic
R 0.9999 0.9591 1.0000 0.9392 1.000
Adj.R2 0.9998 0.9400 1.000 0.9307 1.000
Pred.R2 1.000 0.9249 1.000 0.9147 1.000
Figure 7.11 The BTE variations against compression ratio and injection
pressure
182
Figure 7.12 The BTE variations against compression ratio and Injection
timing
Figure 7.13 The BTE variations against compression ratio and Power
183
Figure 7.14 The BTE variations against Compression ratio and fuel
Blend
Figure 7.15 The BTE variations against blend and injection timing
184
Figure 7.16 The BTE variations against fuel blend and power
Figure 7.17 The BTE variations against fuel blend and injection pressure
185
Figure 7.18 The BTE variations against fuel blend and power
Figure 7.19 The BTE variations against injection pressure and injection
timing
Figure 7.21 The BSFC variations against Biodiesel fuel fraction and
Compression ratio
Figure 7.22 The BSFC variations against Power and Compression ratio
189
Figure 7.23 The BSFC variations against injection timing and injection
pressure
Figure 7.24 The BSFC variations against Biodiesel fuel fraction and
injection pressure
190
Figure 7.25 The BSFC variations against Biodiesel fuel fraction and
injection timing
Figure 7.26 The BSFC variations against Power and Injection timing
191
Figure 7.27 The BSFC variations against power and Biodiesel fuel
fraction
emissions by 37.09% and reduced CR and higher injection pressure 230 bar
increased CO emissions by 9.67 % compared to the results of original CR
(17.5) for B 50. Cenk sayin et al. (2011) found out at lower CR ,insufficient
heat of compression delays ignition and so CO emissions increase.Raheman
et al .(2008) investigated that the possible reason for this trend could be that
the increased CR actually increases the air temperature inside the cylinder
therefore reducing the ignition lag causes better and more complete burning
of the fuel .It is seen in Figure 7.28 that the percentage of CO is less than
0.3% at CR 17.7,B20 and maximum percentage of 1.9% at CR 17.5 ,B50 .The
lower CO emissions of biodiesel blends may be due to their more complete
oxidation as compared to diesel.Some of the CO produced during combustion
of biodiesel might have been converted into CO2 by taking up extra oxygen
molecule present in the biodiesel chain and thus reduces CO formation.
fuel resulted in the smaller droplet size; faster evaporation of fuel sprays; and
improved reaction between fuel and air. These resulted in comparatively
better combustion and contributed for with lesser HC emissions at all
injection timings. Beyond 220 bar of injection pressure, faster velocity of the
fuel jets caused most fuel particles to hit the wall of combustion chamber
where the fuel particles got cooled and not participated in the combustion
process effectively thus resulting in incomplete combustion. From the
experiments is very clear that HC emissions reduces for biodiesel blends.
Figure 7.42 The HC variations against power and Biodiesel fuel fraction
Gumus (2008). found out that when the IP was decreased, NOx
emissions diminished for all the fuel mixtures. Increase in IP decreased the
203
Figure 7.44 The NOx variations against Pressure and compression ratio
204
Figure 7.45 The NOx variations against Biodiesel fuel fraction and
compression ratio
Figure 7.46 The NOx variations against Power and compression ratio
205
Figure 7.47 The NOx variations against injection timing and Pressure
206
Figure 7.48 The NOx variations against Biodiesel fuel fraction and
injection timing
Figure 7.49 The NOx variations against Biodiesel fuel fraction and
Pressure
7.4.5 Optimization
The criteria for the optimization such as the goal set for each
response lower and upper limits used, weight used and importance of the
factors are presented in Table 7.13. In desirability based approach, different
best solutions were obtained. The solution with high desirability is preferred.
lower upper
Name Goal Weight Weight Importance
Limit Limit
CR target=17.70 17.5 18.1 1 1 5
IP target = 230.00 190 230 1 1 5
IT target = 27.00 23 27 1 1 5
Blend target = 20.00 10 50 1 1 5
Power target = 3.64 3.64 5.2 1 1 5
BTE maximize 26.12 35.42 1 1 5
BSFC minimize 0.23 0.35 1 1 5
CO minimize 0.02 2.28 1 1 5
HC minimize 31 298 1 1 5
NOx minimize 726 1335 1 1 5
Compression
17.5 18.1 1 1 5 Target 1
Ratio
Injection
190 230 1 1 5 Target 1
pressure
Injection
23 27 1 1 5 Target 1
timing
Biodiesel fuel
10 50 1 1 5 Target 1
fraction
Power 3.64 5.2 1 1 5 Target 1
BTE 26.12 35.42 1 0.1 5 Maximize 0.998
BSFC 0.234 0.358 0.1 1 5 Minimize 1.000
CO 0.028 2.287 0.1 1 5 Minimize 0.983
HC 31.73 298 0.1 1 5 Minimize 1.000
NOx 725.6 1334.7 0.1 1 5 Minimize 1.000
Combined 0.995
BTE Regression
Expt. No. BTE Experimental Error
model
1 34.41 34 -0.41
2 32.03 33 0.97
3 29.65 28 -1.65
4 26.32 27 0.68
5 28.51 24 -4.51
6 27.70 29 1.3
7 32.16 34 1.84
8 32.96 33 0.04
9 26.73 30 3.27
10 29.83 26 -3.83
11 30.77 34 3.23
12 33.75 35 1.25
13 29.69 30 0.31
14 30.24 31 0.76
15 28.54 29 0.46
16 30.70 27 -3.7
213
CO CO NOx
Exp. No.
Regression Experiment Regression Experiment
1 0.075407 0.04 1541.988 1572
2 0.260186 0.25 1329.042 1300
3 0.444966 0.49 1116.096 1057
4 0.654746 0.6 634.75 690
5 0.684476 0.8 1581.69 1606
6 0.571196 0.52 1210.136 1300
7 0.167957 0.32 1138.702 1122
8 0.036677 0.02 594.148 627
9 0.293447 0.16 1361.522 1258
10 0.013688 0.04 1187.728 1100
11 0.716466 0.72 1160.21 1133
12 0.261687 0.13 666.042 662
13 0.194958 0.11 1483.802 1720
14 0.617237 0.6 1411.27 1300
15 0.146918 0.3 840.368 800
16 0.551197 0.59 594.836 625
215
After selecting the optimal levels of the engine, the final step is to
verify the results using the optimum design parameter levels in comparison
with standard engine parameters with biodiesel fuel. A confirmation test for
the combined objective is conducted by choosing the five design and control
parameters as found in multi objective optimization. An experiment was
217
conducted with the optimized parameters and the results are shown in
Table 7.29.
For the combined objective with equal weight age for all the
responses, the optimum setting is found out.
Model 0.267
3. 17.9 200 23 20 3.64 33.74 104 0.016 783
equation
Experiment
4. 17.9 200 23 20 3.64 33.5 0.26 35 0.020 600
results
35
34.8
34.6
34.4
34.2
BTE (%)
34
33.8
33.6
33.4
33.2
33
32.8
RSM Taguchi Model Experimental
BTH 33.8 34.75 33.74 33.5
From Figure 7.51 the experimental value for the brake specific fuel
consumption for RSM and empirical model values are nearer to experimental
results. Whereas Taguchi optimization gives more deviation result.
0.275
0.27
0.265
0.26
bsfc (kg/kWh)
0.255
0.25
0.245
0.24
0.235
0.23
0.225
0.22
RSM Taguchi Model Experimental
bsfc 0.268 0.239 0.267 0.26
120
100
80
HC (ppm)
60
40
20
0
RSM Taguchi Model Experimental
HC 31 22 104 35
For NOx emission, the experimental value for the optimized results
is 600 PPM. Taguchi and model equation values are closer to experimental
values whereas RSM optimization techniques result is not valid. The value of
NOx emission is 884 PPM when RSM optimization is used.
1000
900
800
700
Nox (ppm)
600
500
400
300
200
100
0
RSM Taguchi Model Experimental
NOx 884 598 783 600
7.6.3.5 CO Emission
0.12
0.1
0.08
CO ( %)
0.06
CO %
0.04
0.02
0
RSM Taguchi Model Experiment
CO % 0.098 0.015 0.016 0.02
7.7 CONCLUSION
When the blend is low (B 10),the BTE is high 33.9 % and fuel
consumption is less and it is minimum when the blend is maximum (B 50)
this is due to the fact at 50 % blend the viscosity and density is is high and
HHV is less.
optimal conditions for the maximum BTE and minimum BSFC, HC, CO,
NOx with Desirability approach of 0.995 was obtained at the optimum engine
parameters of compression ratio of 17.9,injection pressure 200 bar, injection
timing 23 o bTDC, fuel blend B20 and 3.64kW power, where the values of the
BTE,BSFC,CO,HC and NOx were found to be 33.85%, 0.268 kg/kWh, 0.09
%, 31 ppm, 884 ppm respectively.