To diagnose faults of stepping motors in running, a fault reasoning machine using fuzzy neural network (FNN) and its singlechip implementation have been designed.

In the paper,the authors have introduced a fault reasoning method (fault directional graph) used in space propulsive system,and a corresponding fault reasoning model has been set up.

By analysing the three typical fault mechanism of the motor, the paper has put forward a diagnostic method based on the BP neural network.

(2) The main fault mechanism and characteristics of steam turbine generator unit was researched.

The paper analyzed the failure mechanism of the important and most probably failed parts of FOG,such as lamp-house and fiber coil,a.

The failure mode and failure mechanism of hydraulic pump are analyzed in this paper,and then the temperature,pressure and rotating speed are decided primary factors for synthesis stress and the conclusion that the synthesis stress can be used in validating the life of the pumps is received.

In this paper, from the point of view of building the fault diagnostic system for spacecrafts, some fundamental concepts of symptom, failure, fault criterion, fault mode, fault model and failure mechanism are given, and difference and connection among them are discussed.

By using the technology of the fault converse inference,expert system,bus snooping and so on,the system s reliability and flexibility is greatly enhanced.