论文部分内容阅读
针对航空发动机压气机健康监测提出了一种基于线性矩阵不等式(LMI)和H∞优化理论的航空发动机压气机传感器鲁棒故障诊断的方法.在航空发动机具有模型不确定性和外界噪声的情况下,应用基于神经网络的线性拟合方法实现航空发动机压气机离散模型的建立;并通过LMI和H∞优化问题的求解得到未知输入观测器的设计参数,实现具有强鲁棒性的传感器故障诊断.该方法比以前研究中未知输入观测器故障诊断方法的优点在于能够同时处理模型不确定性和外界噪声.应用ALSTOM公司提供的燃气涡轮压气机模型进行了仿真验证,在压气机具有白噪声模型误差和正弦外界干扰的情况下,实现对小于测量范围2%的传感器故障的检测和诊断.
Aeroengine compressor robust fault diagnosis method based on linear matrix inequality (LMI) and H∞ optimization theory is proposed for aero-engine compressor health monitoring. In the case of aero-engine with model uncertainty and external noise , A linear model fitting method based on neural network is used to establish the discrete model of the aeroengine compressor. The design parameters of the unknown input observer are obtained by solving the LMI and H∞optimization problems, and the robustness of the sensor fault diagnosis is achieved. The advantage of the proposed method is that it can deal with the model uncertainty and the external noise at the same time. The gas turbine compressor model provided by ALSTOM Company is used to verify the simulation results. When the compressor has white noise model error And sinusoidal interference outside the case, to achieve less than 2% of the measurement range sensor failure detection and diagnosis.