在燃气轮机气路诊断方法中，通常用热力模型获得气路部件特性参数的偏移描述气路性能退化。由于在线测量数据误差的存在，实机测量数据与热力模型仿真数据具有一定偏差，这种局限性限制了精确气路诊断的实现。为了更准确地诊断燃气轮机的气路性能退化，本文提出了一种基于数据协调的燃气轮机气路诊断方法，构建数据协调方程与仿真值和测量值均方误差相结合的目标函数，以数据协调值代替原本的仿真值，采用粒子群优化算法（PSO）获得部件特性参数偏移的精确解。利用仿真模拟的退化故障案例，开展了气路诊断仿真试验，试验结果表明该方法最大相对偏差小于0.96%，优于传统无数据协调的气路诊断方法。

The gas path diagnosis of gas turbines usually uses the thermal model to obtain the offset of the characteristic parameters of the gas path components to describe the gas path performance degradation. Due to the existence of online measurement data errors, the simulation results of the thermal model always have a certain deviation from the measurement results of the actual operation of the unit. This limitation limits the realization of accurate gas path diagnosis. To effectively diagnose the gas path of the gas turbine, this paper proposes a gas path diagnosis method based on data coordination. Constructing an objective function combining the data-coordinated equations with the mean-square error of the simulated and measured values, replacing the original simulated values with the data-coordinated values. The particle swarm optimization algorithm (PSO) is used to obtain the accurate solution of the component characteristic parameter offset. The simulation test of gas path diagnosis is carried out based on the simulated degenerated fault cases. The test results show that the maximum relative deviation of this method is less than 0.96%, which is better than the traditional method of gas path diagnosis without data coordination.

国家科技重大专项（J2019-I-0003-0004)