论文部分内容阅读
研究一般非均匀采样数据系统鲁棒传感器故障检测设计问题.首先,基于输出时滞方法将非均匀采样数据系统转换成具有时变时滞输出的连续系统;然后,选择故障检测滤波器作为残差产生器,并将故障检测设计问题描述成一个多目标优化问题,即连续时间过程噪声和离散时间测量噪声对残差信号的H∞范数小于一个给定值,同时传感器故障对残差信号的l2增益大于一个给定值,基于输入输出方法以矩阵不等式的形式给出该多目标优化问题有解的充分条件;进一步的,提出一个迭代算法来权衡噪声鲁棒性与故障灵敏度,并将矩阵不等式转换成可解的线性矩阵不等式.最后,对某型飞控系统的仿真实验验证了所提方法的有效性.
The problem of robust sensor fault detection design for non-uniform sampling data systems is studied.Firstly, the non-uniform sampling data system is transformed into a continuous system with time-varying delay output based on the output delay method. Then, the fault detection filter is selected as the residual Generator, and the fault detection design problem is described as a multi-objective optimization problem, that is, the H∞ norm of the residual signal for continuous-time process noise and discrete-time measurement noise is less than a given value, and the sensor fault on the residual signal l2 gain is greater than a given value, based on the input and output methods in the form of matrix inequalities given sufficient conditions for the multi-objective optimization problem; further, an iterative algorithm to weigh the noise robustness and fault sensitivity, and the matrix Inequality into solvable linear matrix inequalities.Finally, a simulation experiment on a flight control system verifies the effectiveness of the proposed method.