摘要
针对SVM预测刀具磨损量存在的参数不易确定的问题,提出了新的基于粒子群优化SVM的智能预测方法。在介绍粒子群算法和SVM回归模型基本理论的基础上,提出用自适应粒子群优化算法优化SVM参数的策略,采用小波包方法对切削声信号进行分解处理,建立了基于粒子群优化SVM的刀具磨损量预测模型。试验分析的仿真结果表明,所建立的刀具磨损量智能预测模型具有较强的推广能力和较高的预测精度。
Aiming at the problem of parameters is not easy to determine existing in tool wear prediction for SVM,putting forward a new intelligent prediction method of based on particle swarm optimization SVM. Based on the introduction of the basic theory of particle swarm optimization and SVM regression model,putting forward a strategy for optimizing SVM parameters based on particle swarm optimization,adopting the wavelet packet method to decompose the acoustic signals,we further establishing the prediction model of tool wear based on particle swarm optimization SVM. The simulation result of experimental analysis shows that the intelligent prediction model of tool wear has strong generalization ability and high prediction accuracy.
出处
《工具技术》
北大核心
2016年第11期109-112,共4页
Tool Engineering
基金
"十二五"国家科技支持计划(2014BD06B00)
关键词
刀具
磨损量预测
粒子群
支持向量机
tool
wear prediction
particle swarm
support vector machine
