Optimization of MRP quartz glass process parameters based on BP-PSO

Abstract

ABSTRACT The BP-PSO was used to obtain the optimal process parameters of magnetorheological polishing quartz glass. Specifically, the orthogonal experiments investigated the effect of process parameters on processing properties. A BP prediction model was constructed with process parameters as input and comprehensive evaluation Z as output. Based on the model, PSO was used to optimize process parameters, and experimental validation was conducted. It turns out that the relationship of MRR with workpiece speed is linear, with working gap and polishing disc speed are sinusoidal, and with yaw speed is quadratic. The relationship between Sa and four process parameters are quadratic. The R2 of the model is 0.9609. The error of the predicted Z is less than 3.45%. The optimal combination predicted is: workpiece speed 700r/min, working gap 3.1mm, yaw speed 115mm/min, and polishing disc speed 60r/min. The predicted Z is 39.96, the actual Z is 41.93, with an error of 4.9%.