水电机组变工况性能退化评估与非线性预测研究

Some problems are encountered for the condition monitoring and fault diagnosis of hydropower unit. These problems are summarized as follows:(1) It is difficult to carry out effective diagnosis due to the complex operating conditions, many measuring points and few fault samples. (2) The dynamic performance differences under variable operating conditions are ignored when setting the static alarm threshold. (3) The decision-making information is scarce although storing huge amounts of data. In this project, the optimization of performance sensitive characteristic parameters, surface modeling of characteristic parameters in health condition, performance degradation assessment and nonlinear prediction methods for hydropower unit are researched. The proposed performance degradation assessment and nonlinear prediction method can adaptively respond to the change of operating conditions (active power and working head change) for hydropower unit. The evaluation standards of operating conditions are presented. The proposed method dedicated to rapidly and effectively detect unit's hazards. Through the project, some original research findings can be obtained. It can further improve and develop the theory and practical applications of condition based monitoring and predictive maintenance in hydropower unit.

水力发电机组状态监测和故障诊断研究存在以下问题：运行条件复杂，测点多，故障样本少，难以进行有效诊断；设置的静态报警阈值，忽略了变工况条件下机组的动态性能差异；拥有海量数据，决策指导信息缺乏。本项目拟在水电机组性能敏感特征参数优选、健康状态特征参数曲面建模、性能退化评估与非线性预测方法等方面展开深入研究，提出能够自适应水电机组出力和水头变化的变工况性能退化动态评估和非线性预测理论与方法，建立运行状态评价标准，以便能及早地发现机组设备可能存在的安全隐患，获得一些原创性的研究成果，为实现水电状态检修和预知维修奠定坚实基础，具有十分重要的理论创新意义和工程应用价值。

水电机组运行条件复杂，振动信号具有很强的非平稳性；故障样本少，难以进行有效诊断；设置的静态报警阈值，忽略了变工况条件下机组的动态性能差异；拥有海量数据，决策指导信息缺乏。本项目研究了复杂背景噪声下水电机组振动信号的去噪方法，提出了基于ALIF、VMD、ITD、NA-MEMD的非平稳振动信号解调方法，对压力脉动进行了有效的时频分析，研究了水电机组状态参数非线性预测模型，开发了基于健康样本的机组异常检测系统，并对复杂工况下旋转机械故障诊断理论进行了深入的研究。部分研究成果已应用到三峡集团所属水电站和风电场。