一类不确定离散时间系统的鲁棒故障诊断滤波器优化设计方法

Optimization approach to robust FDF design for uncertain discrete-time systems

研究线性不确定离散时间系统的鲁棒故障诊断滤波器设计问题。基于新提出的性能指标函数,将不确定离散时间系统的故障诊断滤波器设计问题归结为H∞优化问题,并通过选择适当的后滤波器和观测器增益矩阵,得到未知输入和模型不确定性鲁棒性的最优解。

The problem of robust fault detection filter design for uncertain discretetime linear systems with both modelling errors and l2 norm bounded unknown input is studied. By introducing a new optimization performance index, the design of observerbased robust fault detection filter is formulated as an H∞ optimization problem, which is solved by suitably selecting stable postfilter and observer gain matrix. This method maximizes the sensitivity of residual to fault and guarantees the robustness to model uncertainty as well unknown input.