基于麻雀搜索算法优化支持向量机的刀具磨损识别

针对微小深孔钻削刀具磨损状态检测的工程需求,提出了基于钻削声信号的麻花钻头磨损状态识别方法。根据不同磨损程度的麻花钻在钻削过程中的声信号,使用经验模态分解(empirical mode decomposition, EMD)将声信号分解成若干个固有模态函数(intrinsic mode functions, IMFs),通过时频联合分析探索刀具磨损与声信号特征之间的关联规律;再使用麻雀搜索算法(sparrow search algorithm, SSA)优化支持向量机(support vector machine, SVM)的参数,并利用SVM实现基于声信号特征的刀具磨损状态识别。实验结果表明,微小深孔钻头磨损程度与钻削声信号特征之间存在非线性耦合关系,声信号高频特征对钻头磨损程度的变化非常敏感;采用经过SSA优化后的SVM算法,基于优选的IMF特征能够准确识别钻削刀具磨损状态,识别准确率可达98.246%。

Tool Wear Recognition Based on Sparrow Search Algorithm Optimized Support Vector Machine

Aiming at the engineering requirements of wear state detection of small deep hole drilling tools, a wear state recognition method of twist drill head based on drilling acoustic signal is proposed. According to the acoustic signals of twist drills with different wear degrees during drilling, the acoustic signals are decomposed into several intrinsic mode functions (IMFs) by empirical mode decomposition (EMD), and the correlation between tool wear and acoustic signal characteristics is explored by time-frequency joint analysis. Then the parameters of support vector machine (SVM) are optimized by sparrow search algorithm (SSA), and the tool wear state recognition based on acoustic signal features is realized by SVM. The experimental results show that there is a nonlinear coupling relationship between the wear degree of micro-deep hole drill and the characteristics of drilling acoustic signal, and the high-frequency characteristics of acoustic signal are very sensitive to the change of drill wear degree. The SVM algorithm optimized by SSA can accurately identify the wear state of drilling tool based on the optimized IMF features, and the recognition accuracy can reach 98.246 %.