Fault Diagnosis of Flight Control System Based on Correction High Dimensional Cloud Model
Yong Sun, LI Zeng-lu, Meng Hong, Wenwei Li, Yi Zhong-Kai, Guangyun Li, Jirong Xue · 2012
The reasoning error of existing high dimensional cloud model is large when some domains are correlated, and a correction high dimensional cloud model is proposed based on cloud model and multivariate normal distribution theory. The entropy and hyper entropy were replaced by covariance entropy and covariance hyper-entropy in the correction cloud model, which were used to describe the correlation of some domains. The error could be reduced by the new parameters when the correlation exists. The correction high dimensional backward and forward cloud generators are also proposed. The correction high dimensional cloud was used in fault diagnosis of flight control system. With the residual errors of correlated states as input and the degrees of fault as output, the correction high dimensional cloud multi-rule system could get exact results.