A Modified Particle Swarm Optimization for Form and Position Errors of Mechanical Parts

Abstract

By studying the evaluation principles of mechanical parts’ form and position errors and many kinds of error algorithms, presented in recent patents and scientific articles, a project was proposed by using a modified particle swarm optimization (PSO) with linearly decreasing inertia weights to evaluate form and position errors. With its important parameters being analyzed based on the features of form and position errors’ evaluation, the modified PSO was proved to be of high accuracy and good robustness. The experiments were carried out for evaluating the form and position errors of a group of straight lines and planes produced by harmonic technologies, and the results indicated that the evaluation project was reliable. Finally, the modified algorithm was used to evaluate the roundness error and the cylindricity error of a transmission shaft and the evaluation results show the effectiveness of the modified PSO, compared with the results of a three coordinate measuring machine (CMM).