An Optimization on the DIN1.2080 Alloy in the Electrical Discharge Machining Process Using ANN and GA

Subject Areas :

Masoud  Azimi ¹ (Department of Mechanical Engineering, Khomeinishahr Branch, Islamic Azad University, Isfahan, Iran)
Amin  Kolahdooz ² (Young Researchers and Elite Club, Khomeinishahr Branch, Islamic Azad University, Isfahan, Iran)
Seyyed Ali  Eftekhari ³ (Young Researchers and Elite Club, Khomeinishahr Branch, Islamic Azad University, Isfahan, Iran)

Received: 2017-04-05 Accepted : 2017-06-06 Published : 2017-02-01

Keywords: Genetic algorithm, optimization, Taguchi method, ANN, EDM, Optimum determining,

Abstract :

Electrical Discharge Machining (EDM) process is one of the most widely used methods formachining. This method is used to form parts that conduct electricity. This method of machininghas used for hard materials and therefore selects the correct values of parameters which are soeffective on the quality machining of parts. Reaching to optimum condition of the DIN1.2080 alloy(D3) machining is very important due to the rapid and widespread use of different industry such asMolding, lathe tools reamer, broaching, cutting guillotine and etc. Therefore the purpose of thisstudy is to consider the effect of the inlet parameters such as current, voltage, pulse on time andpulse off time on the machining chip rate and optimize the inlet parameters for D3 alloy. So toreach better result after doing some experiments to predict and optimize the rate of removing chip,neural network method and genetic algorithm are used. Then optimizing input parameters tomaximize the rate of removing chip are performed. In this condition by decreasing time, the productcost is decreased. In this condition, the optimum parameters are obtained under the current of 20(A), 160 (V), pulse on time of 100 (ms) and pulse off time of12 (ms). At this condition, the rate ofmachining chip is obtained 0.063 (cm3/min). Also, surveying the level of error and its accuracy areevaluated. According to the obtained error value that is about 5.18%, the used method is evaluatedsuitable for genetic algorithm.