Multi Objective Optimization of Machining Parameters in End Milling of AISI1020
Jignesh G. Parmar1, Komal G. Dave2
1Jignesh G Parmar*, Ph.D Scholar, Department of Mechanical Engineering, Gujarat Technological University, Ahmedabad (Gujarat), India.
2Dr. Komal G Dave, Professor, Department of Mechanical Engineering, Lalbhai Dalpatbhai College of Engineering, Ahmedabad (Gujarat), India.
Manuscript received on May 23, 2021. | Revised Manuscript received on May 29, 2021. | Manuscript published on June 30, 2021. | PP: 54-63 | Volume-10, Issue-8, June 2021 | Retrieval Number: 100.1/ijitee.H92250610821 | DOI: 10.35940/ijitee.H9225.0610821
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: In current research, artificial neural network (ANN) and Multi objective genetic algorithm (MOGA) have been used for the prediction and multi objective optimization of the end milling operation. Cutting speed, feed rate, depth of cut, material density and hardness have been considered as input variables. The predicted values and optimized results obtained through ANN and MOGA are compared with experimental results. A good correlation has been established between the ANN predicted values and experimental results with an average accuracy of 91.983% for material removal rate, 99.894% for tool life, 92.683% for machining time, 92.671% for tangential cutting force, 92.109% for power and 90.311% for torque. The MOGA approach has been proposed to obtain the cutting condition for optimization of each responses. The MOGA gives average accuracy of 96.801% for MRR, 99.653% for tool life, 86.833% for machining time, 93.74% for cutting force, 93.74% for power and 99.473% for torque. It concludes that ANN and MOGA are efficiently and effectively used for prediction and multi objective optimization of end milling operation for any selected materials before the experimental. Implementation of these techniques in industries before the experimentation is useful to reduce the lead time, experimental cost and power consumption also increase the productivity of the product.
Keywords: Word; ANN; Prediction; MOGA: Multi Objective Optimization; Modelling; DOE.