Laser cutting quality prediction based on pareto genetic algorithm

Abstract

Prediction and optimization of cutting quality is an important method to improve the cutting quality. Aiming at the prediction of quality characteristic parameters for pulsed Nd: YAG laser cutting, a prediction algorithm based on pareto genetic algorithm is used in this paper. KW(Kerf Width) and MRR(Material removal rate) are selected as the optimization objective, and the multi-objective optimization model is established in this paper. The theoretical analysis and experimental results show that the algorithm can be used for KW and MRR prediction in pulsed Nd: YAG laser cutting. A large number of forecast data show the rules as follows. The effects of three types of combined parameters( gas pressure and pulse width, pulse width and pulse frequency, pulse width and cutting speed) on KW are obvious, while the effects of combined parameters, pulse width and pulse frequency, pulse frequency and cutting speed are more obvious on MRR. The study in this paper can provide theoretical guidance and parameters for prediction and optimization of quality in laser cutting.

Multi-Objective Optimization of Quality in Laser Cutting Based on Response Surface Model

“…If it is set too large, the computation amount will be too large, and the search speed of Pareto algorithm will be influenced; if be set too small, the good solution will be missed, and the accuracy will be affected. To balance the precision and the speed, the improved Pareto solution filter used in reference [13] is adopted, and the simultaneous optimization of KW and MRR is conducted in this paper.…”

Optimization of tungsten carbide processing parameters for laser heating and ultrasonic vibration composite assisted cutting

“…Noteworthy is that a choice of cutting parameters exerts a direct effect on the processing efficiency and quality, as well as production costs, so multiple methods of cutting parameters' optimization, such as response surface method (RSM), GA, orthogonal experiment method, are used. 20–25 The orthogonal experiment has a simple structure and may operate with continuous or discrete variables. However, since it fails to clarify the interaction between coupled factors, the response is optimized only at the current level by determining the optimal values.…”