A fault diagnosis system which utilized a bank of Kalman filters was developed for aeroengine sensor double fault detection and isolation.This approach used multiple Kalman filters

each of which was designed based on a specific hypothesis that fault the has occurred in two of the sensors.In addition

the measurements of the normal sensors were used as the input of the Kalman filter.In the event that a double fault does occur

all filters except the one using the correct hypothesis would produce large estimation errors

from which a specific fault is isolated.Furthermore

if the proper Boolean calculation is designed according to the feature of the estimation errors

the fault diagnosis system can also make detection and isolation to the single sensor fault.In order to validate the diagnosis system

a nonlinear engine simulation for both double and single sensor faults at cruise operating condition was conducted.The results indicate that the proposed fault diagnosis system is promising for reliable detection and isolation of aeroengine sensor fault.