The effective monitoring of tool wear status in the milling process of a five-axis machining center is important for improving product quality and efficiency,so this paper proposes a CNN convolutional neural network m...The effective monitoring of tool wear status in the milling process of a five-axis machining center is important for improving product quality and efficiency,so this paper proposes a CNN convolutional neural network model based on the optimization of PSO algorithm to monitor the tool wear status.Firstly,the cutting vibration signals and spindle current signals during the milling process of the five-axis machining center are collected using sensor technology,and the features related to the tool wear status are extracted in the time domain,frequency domain and time-frequency domain to form a feature sample matrix;secondly,the tool wear values corresponding to the above features are measured using an electron microscope and classified into three types:slight wear,normal wear and sharp wear to construct a target Finally,the tool wear sample data set is constructed by using multi-source information fusion technology and input to PSO-CNN model to complete the prediction of tool wear status.The results show that the proposed method can effectively predict the tool wear state with an accuracy of 98.27%;and compared with BP model,CNN model and SVM model,the accuracy indexes are improved by 9.48%,3.44%and 1.72%respectively,which indicates that the PSO-CNN model proposed in this paper has obvious advantages in the field of tool wear state identification.展开更多
低截获概率(low probability of intercept,LPI)雷达作为一种具有强抗干扰能力及低截获特性的新型雷达,对其精准高效识别已成为雷达对抗一方波形识别的难点。针对该方向主流分类器卷积神经网络(convolution neural network,CNN)的结构...低截获概率(low probability of intercept,LPI)雷达作为一种具有强抗干扰能力及低截获特性的新型雷达,对其精准高效识别已成为雷达对抗一方波形识别的难点。针对该方向主流分类器卷积神经网络(convolution neural network,CNN)的结构智能寻优问题,提出一种基于粒子群优化(particle swarm optimization,PSO)算法-CNN的波形识别算法。该算法利用PSO的寻优特性,可实现较大范围内自动搭建不定层数、不定层类别及层内参数的CNN结构并进行迭代寻优;采用识别精度及网络复杂度相结合的衡量指标,可根据需求调整两者比重以实现对精度与轻量性的选择。该算法获取的CNN结构实现了比9种经典CNN结构更好的LPI雷达波形识别效果,同时避免了波形识别时人工选定CNN超参数缺乏智能性、客观性的问题,提高了选用CNN结构的适配性及高效性。展开更多
基金financed with the means of Basic Scientific Research Youth Program of Education Department of Liaoning Province,No.LJKQZ2021185Yingkou Enterprise and Doctor Innovation Program (QB-2021-05).
文摘The effective monitoring of tool wear status in the milling process of a five-axis machining center is important for improving product quality and efficiency,so this paper proposes a CNN convolutional neural network model based on the optimization of PSO algorithm to monitor the tool wear status.Firstly,the cutting vibration signals and spindle current signals during the milling process of the five-axis machining center are collected using sensor technology,and the features related to the tool wear status are extracted in the time domain,frequency domain and time-frequency domain to form a feature sample matrix;secondly,the tool wear values corresponding to the above features are measured using an electron microscope and classified into three types:slight wear,normal wear and sharp wear to construct a target Finally,the tool wear sample data set is constructed by using multi-source information fusion technology and input to PSO-CNN model to complete the prediction of tool wear status.The results show that the proposed method can effectively predict the tool wear state with an accuracy of 98.27%;and compared with BP model,CNN model and SVM model,the accuracy indexes are improved by 9.48%,3.44%and 1.72%respectively,which indicates that the PSO-CNN model proposed in this paper has obvious advantages in the field of tool wear state identification.
文摘低截获概率(low probability of intercept,LPI)雷达作为一种具有强抗干扰能力及低截获特性的新型雷达,对其精准高效识别已成为雷达对抗一方波形识别的难点。针对该方向主流分类器卷积神经网络(convolution neural network,CNN)的结构智能寻优问题,提出一种基于粒子群优化(particle swarm optimization,PSO)算法-CNN的波形识别算法。该算法利用PSO的寻优特性,可实现较大范围内自动搭建不定层数、不定层类别及层内参数的CNN结构并进行迭代寻优;采用识别精度及网络复杂度相结合的衡量指标,可根据需求调整两者比重以实现对精度与轻量性的选择。该算法获取的CNN结构实现了比9种经典CNN结构更好的LPI雷达波形识别效果,同时避免了波形识别时人工选定CNN超参数缺乏智能性、客观性的问题,提高了选用CNN结构的适配性及高效性。
