摘要
针对一类具有不确定输入的状态时滞系统,基于时滞依赖H∞滤波器探讨了系统的鲁棒故障诊断(RFD)问题.首先,引入一个参考模型,形成广义残差系统,以此广义残差信号对扰动信号及其故障信号L2的增益来体现其对扰动的鲁棒性和对故障的灵敏性;其次,利用对偶原理,针对对偶系统,借助线性矩阵不等式(LM I)技术,研究基于状态观测器的时滞依赖RFD系统设计问题,并推导出具有时滞依赖解存在的条件及观测器增益矩阵的求解方法;最后,通过仿真实例验证了该方法的可行性和有效性.
A robust fault detection (RFD) scheme is proposed for a class of time-delay systems with uncertainty inputs. The fault detection filter is designed by using H∞ control theory. The reference model is introduced to formulate the generalized residual, and the robustness of residual to unknown inputs and the sensitivity to fault signal are reflected by an L2-gain. By using the dual principle and linear matrix inequality (LMI) approach, the design of observer-based RFD system is proposed. The existence condition and the solution of observer gain matrix, which are dependent on the size of the delay, are also given. Finally, an example is given to demonstrate the effectiveness of the proposed approach.
出处
《控制与决策》
EI
CSCD
北大核心
2005年第9期1012-1016,共5页
Control and Decision