A fast hybrid roundness evaluation algorithm based on computational geometry and particle swarm optimization for profiles with massive points

Abstract

The advancement of measurement has led to rigorous demands for the accuracy and efficiency of roundness evaluation of minimum zone circle (MZC) for profile with massive points. Existing algorithms struggle to balance efficiency and accuracy in such cases. A fast hybrid roundness evaluation (HCGPSO) algorithm for MZC based on the computational geometry (CG) and particle swarm optimization (PSO) algorithm is proposed, which incorporates a CG-based global best (gBest) selection method into the PSO algorithm and refines the termination criterion. A comparison with the published studies demonstrated that the HCGPSO algorithm were accurate and result are displayed to 10-15 mm. The impact of target accuracy, the harmonic and the number of points contained in profile on the performance of proposed algorithm are analyzed. With sub-nanometer target evaluation precision, HCGPSO enhances efficiency by an average of 64 % for profiles containing the highest harmonics from 30 undulations per revolution (upr) to 500 upr with 1500 points. For profiles with 300 to 4000 points at 150 upr, an average efficiency enhancement of 60 % was observed. For profiles with 4000 points at 500 upr from four workpieces, the calculation time is reduced by 55 % on average compared to the PSO algorithm, with increased accuracy and stability.