Modeling and optimizing lapping process of 440C steel by Neural Network and Multi-objective particle swarm optimization algorithm

The most essential problem in lapping process is low material removal rate which leads to increase in production costs and time. Thus, in this process, it's essential to select a condition that besides producing pieces with required flatness and roughness, has a high material removal rate. In this research, effects of parameters such as abrasive particle size, abrasive particles concentration in slurry, and lapping pressure on material removal rate, flatness and surface roughness were studied by experimental method in single sided lapping of flat workpieces made of 440c steel. In the following, effect of aforementioned parameters on material removal rate, flatness and surface roughness of lapped surface has been modeled using artificial neural network. Finally, by exerting multi-objective particle swarm optimization, simultaneous optimization of material removal rate, surface roughness and flatness of lapping pieces has been conducted and related Pareto front has been obtained. Obtained results show that by using Multi-objective particle swarm optimization algorithm we can produce workpieces with required surface roughness and flatness with high material removal rate. Consequently, by using this method moreover producing workpieces with desired quality, production cost and time would decrease.