Abstract:

A method based on incomplete binary decision tables for fault tree construction is proposed to solve the multi-state problem and the reduction of decision tables. This method starts from the listing of all the possible event sequences by means of decision tables with the condition stub of events and the condition entry of 0 and 1. The output events can be deduced according to interaction of parts under circumstances of different event sequences. On basis of the working principle, the irrelevant and repeated events can be ruled out so as to simplify the rule. And in way of achieving the union sets of the event sequences with the same output events, the causal relationship among events can be achieved. Verified by the fault tree modeling of a nose wheel steering system, the result achieved in this way is consistent with the one in FMEA(Failure Mode and Effects Analysis) and the fault diagnosis handbook and the scale of calculation can be reduced effectively.

Key words: decision tables, fault tree, function flow, structure function, fault propagation