The development of a hybrid firefly algorithm for multi-pass grinding process optimization

Abstract

Industrial grinding processes always involve multiple sequential passes, such as rough, semi-rough, semi-finish, finish, and spark-out, for the surface and geometrical accuracy generation. The design of multi-pass grinding process parameters usually requires in-depth heuristic knowledge or complex process modeling. However, when multiple process output objectives must be achieved, either heuristic knowledge or analytical modeling is incapable in dealing with the large number of process parameters or decoupling the dependency of individual pass with its neighboring passes. In this paper, a hybrid Non-dominated Sorting Firefly Algorithm (NSFA) is proposed by incorporating the non-dominated sorting algorithm with the firefly algorithm. The developed NSFA is capable in searching the optimal whole set of grinding process parameters at improved convergence speed and with less iterations, which is proved by comparing with Non-dominated Genetic Algorithm (NSGA-II) and Non-dominated Particle Swarm Optimization (NSPSO). Finally, a variable-pass internal grinding for engineering ceramics is carried out to verify the efficacy of the proposed NSFA. With the process output objectives of minimal grinding time and geometric error, the optimized process by the NSFA can realize the achievement of all process quality objectives simultaneously with the grinding efficiency increased by 32.4%.