基于改进LSTM-AdaBoost的铣刀磨损量预测

针对铣刀磨损量预测时精度低的问题，提出一种基于黑寡妇算法（BWO）优化的长短期记忆神经网络（LSTM）与AdaBoost集成学习算法相结合的铣刀磨损量预测方法。在铣刀磨损振动信号中提取时域、频域以及时频域多域特征。通过BWO算法优化LSTM的核心参数，并将优化后的LSTM网络与AdaBoost算法进行结合，构建铣刀磨损量预测模型。最后用PHM Society 2010铣刀全寿命周期的振动数据进行实验。研究结果表明：所提方法能够有效地预测出铣刀磨损量变化值，优化后模型的平均绝对误差百分比为3.436%、均方根误差为6.471、决定系数 R2 为0.935。该方法能够获得准确率更高的铣刀磨损量预测值，预测效率更高。

Improved LSTM-AdaBoost Based Milling Tool Wear Prediction

Aiming at the problem of low accuracy when predicting milling tool wear,a milling tool wear prediction method was proposed based on a long short-term memory neural network (LSTM) optimized by the black widow algorithm (BWO) combined with the AdaBoost integrated learning algorithm.The time domain,frequency domain and time-frequency domain multi-domain features were extracted from the milling tool wear vibration signal.The core parameters of the LSTM were optimized by the BWO algorithm,and the optimized LSTM network was combined with the AdaBoost algorithm to build a milling tool wear prediction model.Finally,experiments were conducted with the whole life cycle vibration data of the milling tool of PHM Society 2010.The results show that the proposed method can effectively predict the variation of milling tool wear,and the mean absolute percentage error of the optimized model is 3.436%,root mean squared error is 6.471 and R2 is 0.935 when compared with the pre-optimized model.This method can obtain the prediction value of milling cutter wear with higher accuracy and higher prediction efficiency.