INTERNATIONAL JOURNAL OF R&D IN ENGINEERING, SCIENCE AND MANAGEMENT Vol.4, Issue 2, June 2016, p.p.78-84, ISSN 2393-865X Comparison of Karanja, Mahua and Polanga Biodiesel Production through Response Surface Methodology Sunil Dhingra Assistant professor, Mechanical Engineering Department, UIET, Kurukshetra University, Kurukshetra, Haryana, India-136118 ABSTRACT The biodiesels are produced from various edible/non-edible oils through trans-esterification process. The present work deals with the enhancement of bidiesel yield of karanja, mahua and polanga oils firstly through response surface methodology using trans-esterification process. The biodiesel yields of above oils are found to be lower than 90 %. Other technique like hybrid RSM-GA applied to these oils and the results are found to be enhanced as compared to RSM approach. Keywords: Karanja biodiesel, Mahua biodiesel, Polanga biodiesel response surface methodology, transesterification 1. INTRODUCTION The non-edible oils are used for so many years for the biodiesel production. Various researchers [Taghavifar et al. 2014; Zarei et al., 2014] applied optimization tools for the production of biodiesels. Hence the current work deals with the application of RSM and hybrid RSM-GA tools in order to enhance the biodiesel production. A comparison is also done for these produced biodiesel to find the effective tool for the production of biodiesels. 2. RSM APPLIED TO KARANJA, MAHUA AND POLANGA BIODIESEL Trans-esterification process is used in the production of karanja, mahua and polanga biodiesel. Ethanol and KOH are used as ingredients to achieve the required objective. Five input process parameters namely ethanol concentration, catalyst concentration, reaction time, reaction temperature and mixing speed are evaluated from various research studies [Deb et al., 2014; Taghavifar et al. 2014; Dhingra et al., 2013a; Dhingra et al., 2013b; Dhingra et al., 2014a; Dhingra et al., 2014b; Dhingra et al., 2014c; Dhingra et al., 2014d; Dhingra et al., 2016a; Dhingra et al., 2016b]. A central composite rotatable design is used in RSM for predicting design of experiments. Various optimum solutions are obtained for these biodiesels. The confirmatory experiments/tests have been conducted to verify the accuracy of the predicted models. The various biodiesels are produced at the optimized process parameters as proposed by desirability approach. The experiments are performed thrice and the average value is taken as the actual biodiesel yield. The confirmatory experimental results of various biodiesels produced at optimized process Available at :www.rndpublications.com/journal Page 78 R&D Publications
parameters (as predicted by desirability approach) are shown in table 1. These results reveal that actual biodiesel yield of various oils is near to predicted values verifying the authenticity of the models. The biodiesel yield of karanja, mahua and polanga oils is found to be 72 %, 76 % and 79 % respectively. Hence GA technique is applied to these oils in order to enhance the biodiesel production. Table 1: Confirmatory experiments/tests of various biodiesels Type of oil Process parameters Biodiesel yield (wt. %) Error (%) EC Rt RT CC MS Predicted Actual Karanja 17 50 50 1 320 68.50 72-4.8 Mahua 20 30 45 1 260 73.24 76-3.6 Polanga 17 33 55 1.6 290 82.91 79 +4.9 3. OPTIMIZATION OF PROCESS PARAMETERS OF KARANJA, MAHUA AND POLANGA BIODIESELS USING HYBRID RSM-GA Response surface methodology combined with genetic algorithm (hybrid RSM-GA) is applied to karanja, mahua and polanga oils for predicting their optimum process parameters For each population, the fitness function is to be created and run in the MATLAB. GA sort best fitness as elite, crossover and mutation as reproduction. Crossover and mutation are the correction algorithms since their values can be varied to obtain the best yield. This process remains in continuation until the limit of stopping condition. The GA parameters and their values for predicting optimum solution (maximum biodiesel yield of 92.77 %) are shown in table 1. Table 2 shows the various solutions obtained using GA along with the optimum solution (solution no. 1). The plot depicting the number of generations and karanja biodiesel yield is shown in figure 1. The biodiesel yield of 92.77 % is predicted by the best and mean solutions obtained from GA. To validate the optimum solution, experiments are conducted thrice at optimum process parameters as predicted by GA (solution no. 1, table 2). An average yield of 90.5 % (by weight) is obtained against the predicted value of 92.77 % (by weight). The actual karanja biodiesel yield of 90.5 % (by weight) using hybrid RSM-GA is much higher as compared to 72 % using RSM. Similarly hybrid RSM-GA technique is applied to mahua and polanga oils for enhancing their biodiesel yield. Tables 4 and 5 show the proposed genetic algorithm operators along with their values for predicting the optimum biodiesel yield of mahua and polanga oils respectively. The confirmatory experiments/tests are performed for mahua and polanga biodiesels at optimized trans-esterification process parameters as suggested by hybrid RSM-GA technique. Average biodiesel yield of 93 % and 82.5 % are obtained for mahua and polanga oils respectively against the predicted values of 90.02 % and 84 % showing the authenticity of the predicted results. Table 5.14 shows the comparison of biodiesel yield of karanja, mahua and polanga using hybrid RSM-GA and RSM techniques. An improvement of 18.5 %, 17 Page 79
% and 3.5 % in biodiesel yield of karanja, mahua and polanga oils respectively is observed by using hybrid RSM-GA technique as compared to RSM. It is observed from table 6 that biodiesel yield of more than 90 % is achieved for karanja and mahua oils by using hybrid RSM-GA. However the biodiesel yield of polanga oil (82.5 % by weight) is found to be lower than 90 % (by weight). Hence another optimization technique like hybrid ANN-GA needs to be applied to polanga oil for checking the possibility of further enhancing its biodiesel yield. Table 2: Optimum genetic algorithm operators of karanja biodiesel yield in hybrid RSM-GA Population type Double vector Population size 8 Creation function Scaling function Crossover Function Mutation function Rank Scattered Elite Count 2 Cross over fraction 0.8 No. of Generations 50 Stall generations 50 Table 3: Optimum solution sets of trans-esterification process parameters of karanja oil using RSM- GA package S. No. EC Rt RT CC MS KBY (Predicted) 1. 15.72 53.93 58.55 0.53 339.05 92.77 Selected 2. 17.25 47.01 52.27 0.578 194.25 91.03 3. 16.8 42.95 55.6 0.62 208.14 90.572 4. 15.29 20 51.58 0.64 155.65 89.8 5. 15.12 56.17 57.03 0.5 407.67 88.88 6. 15 46.79 55.21 0.66 289.6 86.1049 7. 16.29 53.71 49.59 0.55 275.64 85.6 8. 15 42.8 57.06 0.62 312.18 83.43 9. 18.27 45.35 51.96 0.58 262.1 77.94 10. 19.55 52.55 54.21 0.69 327.51 74.06 Page 80
Figure 1: Karanja biodiesel yield predicted from hybrid RSM-GA technique Table 4: Optimum genetic algorithm (GA) operators of mahua biodiesel yield in hybrid RSM-GA Population type Double vector Population size 10 Creation function Scaling function Crossover Function Mutation function Rank Scattered Elite Count 2 Cross over fraction 0.8 No. of Generations 80 Stall generations 50 Page 81
Table 5: Optimum genetic algorithm (GA) operators of polanga biodiesel yield in hybrid RSM-GA Population type Double vector Population size 15 Creation function Scaling function Rank Crossover Function Scattered Mutation function Elite Count 2 Cross over fraction 0.8 No. of Generations 70 Stall generations 50 Table 6: Comparison between karanja, mahua and polanga biodiesels produced from RSM and hybrid RSM-GA techniques Technique Process parameters Biodiesel yield (wt. %) % EC Rt RT CC MS Predicted Actual improvement Karanja oil RSM-GA 15.75 54 60 0.5 340 92.77 90.5 18.5 RSM 17 50 50 1 320 68.50 72 Mahua oil RSM-GA 15.51 54.54 56.02 2.49 387.89 90.02 93 17 RSM 20 30 45 1 260 73.24 76 Polanga oil RSM-GA 18.65 52.56 55.23 1.25 395 84 82.5 3.5 RSM 17 33 55 1.6 290 82.91 79 4. CONCLUSION i. The biodiesel yields of 72 %, 76 % and 79 % (by weight) are obtained for karanja, mahua and polanga oils respectively by performing the confirmatory experiments using RSM approach as referred from table 1. Page 82
ii. The biodiesel yields of 90.5 %, 93 % and 82.5 % (by weight) are obtained for karanja, mahua and polanga oils respectively by using hybrid RSM-GA approach as referred from table 6. REFERENCES [1] Deb, M., Banerjee, R., Majumder, A., & Sastry, G. R. K. (2014). Multi objective optimization of performance parameters of a single cylinder diesel engine with hydrogen as a dual fuel using pareto-based genetic algorithm. International Journal of Hydrogen Energy, 39(15), 8063-8077. [2] Dhingra, S., Bhushan G., & Dubey, K. K. (2013a). Development of a combined approach for improvement and optimization of karanja biodiesel using response surface methodology and genetic algorithm. Frontiers in Energy, 7(5), 495 505 [3] Dhingra, S., Bhushan G., & Dubey, K. K. (2013b). Performance and emission parameters optimization of mahua (madhuca indica) based biodiesel in direct injection diesel engine using response surface methodology. Journal of Renewable and Sustainable Energy, 5, 063117, DOI: 10.1063/1.4840155. [4] Dhingra, S., Bhushan G., & Dubey, K. K. (2014a). Understanding the interactions and evaluation of process factors for biodiesel production from waste cooking cottonseed oil by design of experiments through statistical approach. Frontiers in Energy (in press). [5] Dhingra, S., Bhushan G., & Dubey, K. K. (2014b). Multi-objective optimization of combustion, performance and emission parameters in a jatropha biodiesel engine using Non-dominated sorting genetic algorithm-ii. Frontiers of Mechanical Engineering,9(1), 81-94 [6] Dhingra, S., Bhushan G., & Dubey, K. K. (2016a). Comparative performance analysis of jatropha, karanja, mahua and polanga based biodiesel engine using hubrid genetic algorithm. Journal of Renewable and Sustainable Energy, 8, 013103, DOI:10.1063/1.4939513. [7] Dhingra, S., Bhushan G., & Dubey, K. K. (2016b). Validation and enhancement of waste cooking sunflower oil based biodiesel production by the trans-esterification process. Energy Sources, part A, 38(10), 1448-1454. Page 83
[8] Dhingra, S., Dubey, K. K., & Bhushan, G. (2014c). A Polymath Approach for the Prediction of Optimized Transesterification Process Variables of Polanga Biodiesel. Journal of the American oil Chemist s Society, 91(4), 641-653 [9] Dhingra, S., Dubey, K. K., & Bhushan, G. (2014d). Enhancement in Jatropha-based biodiesel yield by process optimization using design of experiment approach. International Journal of Sustainable Energy, 33 (4), 842-853. [10] Taghavifar, H., Khalilarya, S., & Jafarmadar, S. (2014). Diesel engine spray characteristics prediction with hybridized artificial neural network optimized by genetic algorithm. Energy, 71, 656-664. [11] Zarei, A., Amin, N. A. S., Talebian-Kiakalaieh, A., & Zain, N. A. M. (2014). Immobilized lipasecatalyzed transesterification of Jatropha curcas oil: Optimization and modeling. Journal of the Taiwan Institute of Chemical Engineers, 45(2), 444-451. Page 84