燃气轮机排气温度异常检测及诊断

摘要

针对燃气轮机运行过程中的健康维护问题，提出了一种基于模糊C均值聚类的燃气轮机排气温度异常检测方法。以某型工业燃气轮机为例，采用模糊C均值聚类算法对排气温度的监测数据进行了聚类分析，得到了燃气轮机不同运行状态下的排气温度特征模式，并在此基础上实现了燃气轮机排气温度异常状态下的故障诊断分析。研究结果表明，燃气轮机甩负荷及其热通道部件的损坏失效均能够对排气温度产生不同程度的影响，模糊C均值聚类算法可以有效实现燃气轮机排气温度的异常检测，为燃气轮机性能退化预测及故障诊断提供决策参考。%Aiming at the problem of health maintenance under gas turbine operating conditions, a fuzzy C⁃means clustering approach was applied to realize the anomaly detection of gas turbine exhaust gas temperature in this pa⁃per. Taking an industry gas turbine as an example, the clustering analysis of exhaust gas temperature based on fuzzy C⁃means clustering approach was studied to obtain the feature pattern of exhaust gas temperature when the gas tur⁃bine was in different operating conditions. Based on this, a diagnosis study was carried out to analyze the exhaust gas temperature in the abnormal state. The results showed that both the gas engine's load rejection and the damage failure of hot section parts have effects on the exhaust gas temperature of the gas turbine to varying degrees. The fuzzy C⁃means clustering approach can effectively realize the anomaly detection of gas turbine exhaust gas tempera⁃ture, which provides decision references for gas turbine performance degradation prediction and fault diagnosis.