针对一类离散线性时变系统的故障诊断问题,提出一种新的故障检测与估计算法.该算法通过引入虚拟故障构建离散故障跟踪估计器,在选取的优化时域内,利用估计器输出和系统实际输出产生的残差信号,采用迭代学习算法来调节虚拟故障,使虚拟故障逼近系统中实际发生的故障,从而达到对系统故障诊断的目的.该方法不仅能检测出系统不同类型的故障,还可以实现对故障信号的精确估计.仿真结果验证了所提出方法的有效性. Aiming at the problem of fault diagnosis for a class of discrete time-varying linear systems, a new fault detection and estimation algorithm is proposed. This algorithm constructs a discrete fault tracking estimator through the introduction of virtual faults, and uses the output of estimator and system The residual signal is actually output and the iterative learning algorithm is used to adjust the virtual fault so that the virtual fault approaches the actual fault in the system to achieve the purpose of system fault diagnosis.The method can not only detect different types of faults in the system but also Can accurately estimate the fault signal.The simulation results verify the effectiveness of the proposed method.