一种基于BP网络的多模型预测主动容错控制方法

Predictive Active Fault-Tolerant Control of Multiple Models Based on BP Networks

针对非线性受控对象,给出了一种基于BP神经网络的多模型预测主动容错控制方案,设计了故障诊断与决策机制的算法步骤,在基于BP网络建立系统正常或故障状态动态模型库的基础上,利用预测控制算法得到控制律,达到对非线性系统主动容错控制的目的.仿真结果表明所述方法的可行性和有效性.

A new active fault-tolerant control method of nonlinear system is given on the basis of BP networks and predictive control strategy. An algorithm of fault detection and diagnosis was put forward. The predictive control strategy of tuning the controller＇s parameters is used. Simulations are given to demonstrate the effectiveness of the proposed approach.