基于逆系统多模型内模主动容错控制

Active Fault Tolerant Control Using Multi-Model Based on Inverse System Internal Model

针对现有非线性系统对故障容错的快速性和可靠性不足,从而导致容错失败的问题,为解决上述问题,研究了一类具有可能故障先验知识的非线性系统,提出了一种基于最小二乘支持向量机(LS-SVM)建模的逆系统多模型内模主动容错控制方法。采用LS-SVM对系统正常以及各种先验故障情形的逆系统进行建模,构造系统的逆模型库,根据逆系统方法将逆模型与被控系统串联形成伪线性系统,设计具有良好鲁棒性能的内模控制器。系统在线运行时,监控决策机制基于对系统性能指标的仿真计算,结果表明在无需改变内模控制器的情况下达到对非线性系统故障容错的快速性和可靠性,证明方法的有效性。

Becuace the existing nonlinear system is insufficient in quickness and reliability for fault-tolerant,which often results in failure,By researching a nonlinear system with priori knowledge of possible failure,the paper proposes an active fault-tolerant control using multi-models based on least square support vector machine（LS-SVM） inverse system internal model control method.First,the method uses LS-SVM to set up inverse system model for normal and faults with each kind of priori breakdown situation,then constitutes the inverse model bank.Based on inverse system method,the inverse model is connected to the system to form a pseudo-linear system and to design the internal model controller with good robust performance.When the system is running,the decision-making mechanism of monitor based on system performance index of calculation and analysis locates the movement patten which is called switching LS-SVM inverse model,so the system has always been connected with inverse model and remains unchanged pseudo-linear system,thus does not need to change the situation of internal model controller for nonlinear system,and achieves fault-tolerant of quickness and reliability.The effectiveness of this method can be seen from the simulation at last.