Optimization of electrical discharge machining process using combined artificial neural networks and heuristic algorithm

In this paper the effect of input electrical discharge machining process variables on AISI۲۳۱۲ hot worked steel is modeled and optimized. The objective is to find a combination of process variables to minimize tool wear rate and surface roughness and maximize material removal rate simultaneously. In order to establish the relationships between the input and the output parameters, back propagation neural network used. In the last section of the study, particle swarm optimization algorithm has been employed for optimization of the multiple response characteristics. Using the proposed optimization procedure, proper levels of input parameters for any desirable group of process outputs can be identified. The results indicate that the proposed modeling technique and PSO algorithm are quite efficient in modeling and optimization of the process variables in order to the desired outputs.

Keywords:

Electrical discharge machining (EDM) process , Design of experiments (DOE) , Optimization , Back propagation neural network (BPNN) , Particle swarm optimization (PSO) algorithm