Remote monitoring of tools for prediction of tool wear in cutting processes was considered, and a method of implementation of a remote-monitoring system previously developed was proposed. Sensor signals were received ...Remote monitoring of tools for prediction of tool wear in cutting processes was considered, and a method of implementation of a remote-monitoring system previously developed was proposed. Sensor signals were received and tool wear was predicted in the local system using an ART2 algorithm, while the monitoring result was transferred to the remote system via intemet. The monitoring system was installed at an on-site machine tool for monitoring three kinds of tools cutting titanium alloys, and the tool wear was evaluated on the basis of vigilances, similarities between vibration signals received and the normal patterns previously trained. A number of experiments were carried out to evaluate the performance of the proposed system, and the results show that the wears of finishing-cut tools are successfully detected when the moving average vigilance becomes lower than the critical vigilance, thus the appropriate tool replacement time is notified before the breakage.展开更多
针对地铁车轮磨耗数据时间跨度较长引起的长期依赖问题,为了进一步提升预测精度,提出一种将麻雀搜索算法(sparrow search algorithm,SSA)优化双向长短期记忆网络(bidirectional long short term memory,Bi LSTM)的改进BiLSTM(SSA-BiLSTM...针对地铁车轮磨耗数据时间跨度较长引起的长期依赖问题,为了进一步提升预测精度,提出一种将麻雀搜索算法(sparrow search algorithm,SSA)优化双向长短期记忆网络(bidirectional long short term memory,Bi LSTM)的改进BiLSTM(SSA-BiLSTM)网络模型,用于地铁车轮磨耗预测。首先,利用麻雀搜索算法对双向长短期记忆网络算法的神经元个数、迭代次数、输入批量和学习率等超参数在给定范围内进行寻优,得到参数最优值;然后,以参数最优值来构建改进BiLSTM网络模型,对车轮磨耗进行预测分析;最后,以车轮踏面磨耗和轮缘磨耗作为研究对象,将某地铁1车厢1号车轮的现场实测历史磨耗数据作为输入,对该模型进行训练及验证分析,并与多层感知机(multilayer perceptron,MLP)、LSTM、BiLSTM以及SSA-LSTM模型的预测结果进行对比。研究结果表明:SSA-Bi-LSTM模型的车轮磨耗预测精度更高,与LSTM、BiLSTM以及SSA-LSTM网络模型相比,踏面磨耗的平均绝对百分误差(mean absolute percentage error,MAPE)分别降低了13.28%、10.32%、1.47%,轮缘磨耗分别降低了9.5%、0.46%、0.02%;分别对同一地铁2号、4号车厢的1号位置车轮磨耗进行预测,并与磨耗实测数据进行对比,踏面磨耗的平均绝对百分比误差分别为1.34%、1.42%,轮缘磨耗的平均绝对百分比误差分别为0.18%、0.19%,验证了本文所提模型具有良好的泛化性,为地铁轮对智能化管理提供理论支持,延长车轮使用寿命。展开更多
基金supported by Changwon National University in 2009-2010
文摘Remote monitoring of tools for prediction of tool wear in cutting processes was considered, and a method of implementation of a remote-monitoring system previously developed was proposed. Sensor signals were received and tool wear was predicted in the local system using an ART2 algorithm, while the monitoring result was transferred to the remote system via intemet. The monitoring system was installed at an on-site machine tool for monitoring three kinds of tools cutting titanium alloys, and the tool wear was evaluated on the basis of vigilances, similarities between vibration signals received and the normal patterns previously trained. A number of experiments were carried out to evaluate the performance of the proposed system, and the results show that the wears of finishing-cut tools are successfully detected when the moving average vigilance becomes lower than the critical vigilance, thus the appropriate tool replacement time is notified before the breakage.
文摘针对地铁车轮磨耗数据时间跨度较长引起的长期依赖问题,为了进一步提升预测精度,提出一种将麻雀搜索算法(sparrow search algorithm,SSA)优化双向长短期记忆网络(bidirectional long short term memory,Bi LSTM)的改进BiLSTM(SSA-BiLSTM)网络模型,用于地铁车轮磨耗预测。首先,利用麻雀搜索算法对双向长短期记忆网络算法的神经元个数、迭代次数、输入批量和学习率等超参数在给定范围内进行寻优,得到参数最优值;然后,以参数最优值来构建改进BiLSTM网络模型,对车轮磨耗进行预测分析;最后,以车轮踏面磨耗和轮缘磨耗作为研究对象,将某地铁1车厢1号车轮的现场实测历史磨耗数据作为输入,对该模型进行训练及验证分析,并与多层感知机(multilayer perceptron,MLP)、LSTM、BiLSTM以及SSA-LSTM模型的预测结果进行对比。研究结果表明:SSA-Bi-LSTM模型的车轮磨耗预测精度更高,与LSTM、BiLSTM以及SSA-LSTM网络模型相比,踏面磨耗的平均绝对百分误差(mean absolute percentage error,MAPE)分别降低了13.28%、10.32%、1.47%,轮缘磨耗分别降低了9.5%、0.46%、0.02%;分别对同一地铁2号、4号车厢的1号位置车轮磨耗进行预测,并与磨耗实测数据进行对比,踏面磨耗的平均绝对百分比误差分别为1.34%、1.42%,轮缘磨耗的平均绝对百分比误差分别为0.18%、0.19%,验证了本文所提模型具有良好的泛化性,为地铁轮对智能化管理提供理论支持,延长车轮使用寿命。
