基于鲁棒滑模观测器的多电机卷绕系统故障检测和隔离

Fault detection and isolation for multi-motor web-winding system via robust sliding mode observers

针对多电机卷绕系统故障检测和隔离问题,提出了基于滑模观测器的鲁棒故障检测和隔离策略.首先,考虑到多电机卷绕系统运行环境变化引起的摩擦系数、杨式模量、半径和转动惯量等参数的不确定性,通过引入非线性未知输入干扰对系统不确定参数进行描述,将多电机卷绕系统转化为包含执行器故障的一般不确定非线性系统;其次,构造鲁棒滑模观测器并用作故障检测观测器,通过滑模控制来抑制未知输入干扰,使观测器具有鲁棒性,在此基础上,结合多观测器故障隔离的思想,提出了可以同时对多个执行器故障进行检测和隔离的方法;然后,根据李雅普诺夫理论得到使观测器系统指数收敛的充分条件,同时给出了观测器增益的求解方法;最后,仿真结果验证了所提方法的鲁棒性和有效性.

The robust fault detection and isolation strategy for multi-motor web-winding system with changing operating environment is put forward. Firstly, taking into account the changes of the friction coefficients, Young＇s modulus,radii and the moments of the roller, etc. which are caused by the changing operating environment, the unknown input disturbance is introduced to describe the system parameters uncertainties of the multi-motor web-winding system, based on which a generalized nonlinear system dynamic model with actuator faults and uncertainties is constructed. Next, a robust sliding mode observer is developed as fault detection observer to detect actuator faults. Furthermore, multi-observers are used for the multi actuator faults isolation purpose. Then, according to the Lyapunov theory, the sufficient conditions of exponential convergence for the designed observers are derived. Meanwhile, the observer gain matrix is presented. Finally,the simulation is conducted to verify the robustness and effectiveness of the proposed method.