基于支持向量数据描述的油动机故障预警研究

Research on Hydraulic Servomotor Fault Warning Based on Support Vector Data Description

将以机器学习为基础的设备故障预警技术应用到油动机系统状态监测上,实现设备早期故障的诊断,可有效提高油动机系统运行的可靠性。通过总结调门油动机常见的故障现象,分析归纳了故障的机理和特点,并根据故障预警目标优化了调门油动机多传感器网络,搭建了油动机运行状态数据采集系统。设计了调门油动机故障注入试验方案,获取了不同故障阶段的油动机状态原始数据,基于SVDD算法实现了调门油动机内泄漏故障趋势的准确预报及严重程度的定量评价。研究成果可为同类型汽轮机EH系统状态监测及故障预警提供参考。

The equipment failure warning technology based on machine learning is applied on the condition monitoring of the hydraulic servomotor system to realize the diagnosis of early equipment failure and improve the operation reliability of the hydraulic servomotor.By concluding the common failure phenomenon of the hydraulic servomotor,the mechanisms and features of the failure are analyzed and summarized,and hydraulic servomotor multi-sensor network is optimized based on the target of the fault warning.Meanwhile,a data acquisition system for the operation state of the hydraulic servomotor is built.A test scheme of hydraulic servomotor fault injection is designed to obtain the original state data of the hydraulic servomotor on different fault conditions.Based on the SVDD algorithm,the accurate prediction of the internal leakage fault trend of the hydraulic servomotor and the quantitative evaluation of its severity are realized.The research results can provide reference for the condition monitoring and fault warning of the EH system in the same type of steam turbines.