论文部分内容阅读
针对带有概率密度函数逼近误差的非高斯不确定奇异随机分布控制系统提出鲁棒故障诊断算法,在有模型不确定性和概率密度函数逼近误差的情况下设计故障诊断观测器估计故障信息。故障发生后,利用故障信息重构跟踪控制器使得输出概率密度函数仍能够跟踪期望概率密度函数。利用李雅普诺夫稳定性理论分析观测误差动态系统、闭环控制系统和跟踪误差动态系统的稳定性,相应的增益矩阵由线性矩阵不等式求解。仿真实例验证了算法对时变故障的有效性。
A robust fault diagnosis algorithm for non-Gaussian singular stochastic distributed control system with approximation error of probability density function is proposed. The fault diagnosis observer is designed to estimate the fault information with model uncertainty and approximation error of probability density function. After the fault occurs, using the fault information to reconstruct the tracking controller makes the output probability density function still able to track the expected probability density function. The stability of observation error dynamic system, closed-loop control system and tracking error dynamic system are analyzed by using Lyapunov stability theory. The corresponding gain matrix is solved by linear matrix inequality. The simulation results verify the effectiveness of the algorithm against time-varying faults.