新的基于LMI的鲁棒故障检测滤波器设计方法

New design approach to robust fault detection filter based on LMI

提出一种新的基于LMI的滤波器设计方法,能够解决在线性时不变系统中进行故障检测时所遇到的多目标优化问题。通常在故障检测中,H∞范数用于描述残差对于外界干扰的鲁棒性,而H-指标用于描述残差对于故障的灵敏度。通过引入残差的误差值,将多个指标(H∞和H-指标)统一转换成H∞范数,从而将故障检测滤波器的设计问题转换成对H∞范数优化问题;然后用一种新颖的基于LMI的方法来求解;最后,将本算法应用于某导弹发生故障的纵向平面模型,仿真结果表明了该算法能有效、及时地检测出故障。

Presented a new LMI-based filter scheme which could solve the multi-objective optimation problem for fault detection in LTI plants. In fault detection field, used H∞ norm to describe the robustness of the residual against unknown uncertainties and disturbances, and the H∞ index described the best sensitivity of the residual to the fault. The main contributions consisted of the formulation of the multi-objective optimization FD problem as a H~ norm optimization problem by using the residual error instead of residual, and the new LMI-based technique as the solution. Eventually, applied the design scheme in an aerocraft＇ s longitudinal model and achieved good performance.