串行分组深度学习运行状态分析与故障预测  

Operation State Analysis and Fault Prediction Model Based on Serial Grouping Deep Learning

作　　者：钱虹 孙勃 郭媛君[2,3] 凌君 杨之乐 冯伟 QIAN Hong;SUN Bo;GUO Yuan-jun;LING Jun;Yang Zhi-le;FENG Wei(School of Automation Engineering,Shanghai University of Electric Power,Shanghai 200090,China;Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences,Shenzhen 518055,China;State Key Laboratory of Nuclear Power Safety Monitoring Technology and Equipment,Shenzhen 518172,China;Department of Automation,Shanghai Jiaotong University,Shanghai 200240,China;Key Laboratory of System Control and Information Processing,Ministry of Education of China,Shanghai Jiaotong University,Shanghai 200240,China;Shanghai Engineering Research Center of Intelligent Control and Management,Shanghai Jiaotong University,Shanghai 200240,China)

机构地区：[1]上海电力大学自动化工程学院,上海200090 [2]中国科学院深圳先进技术研究院,广东深圳518055 [3]核电安全监控技术与装备国家重点实验室,广东深圳518172 [4]上海交通大学自动化系,上海200240 [5]上海交通大学系统控制与信息处理教育部重点实验室,上海200240 [6]上海交通大学上海工业智能管控工程技术研究中心,上海200240

出　　处：《控制工程》2023年第5期936-943,共8页Control Engineering of China

基　　金：国家自然科学基金资助项目(51607177)。

摘　　要：在现代工业生产运行中,如何充分挖掘海量的多源异构生产运行数据,实现异常工况的快速检测和故障预测,有效提高工业生产设备的可靠性,仍然是研究的难点。提出一种基于串行分组深度学习的工业生产运行状态分析与故障预测模型,针对时间序列突变故障设计了串行分组深度学习网络框架,实现目标对象的故障检测与预测,及时发出故障预警。通过对某造纸厂数据以及风力发电频率监测数据进行测试,并与传统神经网络预测模型进行对比分析,表明了所提算法的准确性,为提高生产设备使用寿命、减低工业生产成本,提高安全稳定运行起到重要作用。In the modern industrial production operation,a research difficulty still exists on how to fully analysis the massive multi-source heterogeneous production data,realize the rapid detection of abnormal conditions and fault prediction,and effectively improve the reliability of industrial production equipment.A serial grouping deep learning model of industrial production operation state analysis and fault prediction are proposed,according to time series abrupt faults.By testing the data of a paper mill and wind power frequency monitoring data,and comparing with the traditional neural network prediction model,the accuracy of the proposed algorithm is demonstrated,which plays an important role in improving the service life of production equipment,reducing the cost of industrial production,and improving the safety and stability of industrial operations.

关 键 词：故障预测 串行分组深度学习 长短期记忆模型 卷积网络 主成分分析 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]