基于HPSO优化BP神经网络的刀具磨损状态识别

针对BP神经网络容易陷入局部极值导致识别精度低的问题,文章提出了一种基于混合粒子群算法(HPSO)的BP神经网络优化算法。在刀具磨损监测实验过程中,采集刀具切削的声发射(AE)信号,利用小波包分解算法对AE信号进行滤波,并进行特征提取。将频带能量特征和切削参数分别作为主特征和辅助特征,并对其对归一化处理。采用混合粒子群优化算法(HPSO)对BP神经网络预测模型进行优化,利用优化后的模型对测试样本进行模式识别,结果表明,优化后的HPSO-BP模型能够有效地降低神经网络陷入局部极值的情况,提高刀具磨损识别精度。

Tool Wear Recognition Based on HPSO Optimized BP Neural Network

In view of the problem that BP neural network is easy to fall into local extremum,which leads to low recognition accuracy,this paper proposes a BP neural network optimization algorithm based on hybrid particle swarm optimization(HPSO).Acoustic emission(AE)signals of tool cutting are collected during tool wear monitoring experiments,Wavelet packet decomposition algorithm is used to filter AE signal and extract features.The extracted eigenvalues of frequency band energy and cutting parameters are taken as main features and auxiliary features then into non-dimensional sequences.Using hybrid particle swarm optimization algorithm optimizes the BP neural network prediction model,and the optimized model is used for pattern recognition of test samples.The results show that the optimized HPSO-BP model can effectively improve the situation that the neural network falls into local extremum and the accuracy of tool wear state recognition.