基于MPC的I/O约束容错控制数值化设计方法

MPC Based Numerical Method for Fault-tolerant Control with I/O Constraints

针对一类离散系统的传感器故障,研究同时具有控制输入幅值及变化率约束、可测输出幅值约束下的主动容错控制问题。借鉴预测控制策略和多约束优化算法,仅利用故障传感器输出信号和控制输入信号,以及状态估计滤波器,通过构建二次型优化目标的多约束凸优化算法模型,在线递推计算容错控制输入信号,得到了一种基于MPC策略的数值化容错控制设计方法。通过改进现有MPC稳定性结果的不足,证明了闭环故障容错系统的渐近稳定性,且具有给定的I/O操作约束和优化的二次型性能。最后,通过对某汽车电子节气门的容错控制仿真实验,验证了该方法的有效性。

For the sensor fault case of a class of discrete-time systems, the active fault-tolerant control is investigated in consideration of the control input and its rate constraints, as well as the measured output constraint. Based on the model-based predictive control(MPC) method and the constrained optimization algorithm, the design of the fault-tolerant controller is converted into the development of constrained convex optimization with the quadratic objective function. The fault-tolerant control input is then calculated online in recursive way by using the control input in the past time, the faulty sensor output and state estimation filter.Thus, a numerical fault-tolerant control method is derived based on the MPC strategy. The asymptotical stability of the closed-loop system is proved by improving the existing results in MPC. The input-output(I/O)constraints and the minimized quadratic performance are also guaranteed. Finally, simulations on the fault-tolerant control of an electronic throttle show the validity of the proposed method.