摘要
针对传感器故障,提出了一种基于改进的BP神经网络的集成故障诊断方法。在测量回路中引入“等价偏差”向量,用改进的BP网络建立传感器故障模型,对系统的状态和故障参数进行在线估计,然后将故障参数与修正的Bayes分类算法(MB算法)相结合,进行传感器故障在线检测、分离和估计。对连续搅拌釜式反应器(CSTR)的仿真结果表明,该集成故障诊断方法能够对多重传感器故障进行快速准确的分离和估计,并对传感器故障具有容错性。
An integrated fault detection and diagnosis approach to sensor faults based on improved BP neural networks is presented, “Equal departure” vector is introduced into the measure circuit, and a BP neural network is used to estimate the state and fault parameters of the constructed model for sensor faults. The estimated fault parameters are processed by the Bayes algorithm to realize online sensor fault detection, isolation, and estimation. The simulation for CSTR shows that the presented approach can isolate and estimate the multiple sensor faults quickly and accurately and the integrated system has tolerantability to sensor faults.
出处
《控制工程》
CSCD
2005年第5期429-431,共3页
Control Engineering of China
关键词
故障诊断
BP神经网络
状态估计
容错控制
fault detection and diagnosis
improved BP neural network
state estimation
fault tolerant control
