贝叶斯假设理论检测发动机传感器故障

贝叶斯多重假设检验是将被检测传感器的M个可能状态,作相应M个假设H_i,其先验概率分别为P(H_i)(i=1,2,…,M),故障决策就是从给定观测量M,寻求H_j为真,由贝叶斯风险函数H_i(i=1,2,…,M,i≠j)个假设中的最小值确定最可能发生的假设H_l。 

AEROENGINE SENSOR FAILURE DETECTION BY BAYESIAN MULTIPE HYPOTHESIS TESTING

Analytical redundancy of the advanced sensor failure detection isolation and fault sensor output reconstruction obviously enhance the reliability of aircraft engine controls.The Bayesian multiple hypothesis testing is applied to detection isolation and accommodation of the hard failure.The hypothesis conditioned errors,the covariance matrices of the cost function and the optimal estimate are obtained from a bank of Kalman filters.The example using the microcomputer based simulation indicates that the detective rate,the accuracy of output reconstruction and the detectable minimum range and so on has the advantage over the Kalman filter residual in hard failure.