基于改进型强跟踪卡尔曼滤波的燃气轮机组气路故障诊断研究

Gas Turbine Gas-path Components Health Diagnosis based on Improved Strong Tracking Kalman Filter

针对燃气轮机组气路故障诊断易受时变噪声干扰以及突变故障诊断精度不高等问题,本文提出一种基于改进型强跟踪卡尔曼滤波的燃气轮机组气路故障诊断方法。该算法通过引入气路部件先验知识,合理分配各通道的调节作用,从而提高了气路故障诊断的精度以及动态响应速度。以PG9171E型燃气轮机为研究对象,分别利用EKF(扩展卡尔曼滤波)、STF(强跟踪滤波)以及ISTF(改进型强跟踪滤波)对常见气路故障进行诊断,结果表明ISTF算法同时兼具良好的响应速度以及较高的精度。

For the gas path fault diagnosis susceptible to time-varying noise interference and the problem of poor gas path abrupt fault diagnosis,a method based on the improved strong tracking Kalman filter was proposed. The algorithm introduced a priori knowledge of gas path,and adjusted each child filtering channel reasonably,so as to improve thediagnostic accuracy and response speed. With the PG9171E gas turbine as the object of study,the EKF,STF and ISTF were utilized to diagnose common gas-path fault. The results showed that the ISTF has not only quick response,but also high estimation accuracy.