基于粒度熵的SDG故障推理方法

he fault diagnosis method based on qualitative SDG deep knowledge model has better completeness, but lower resolution. Because it ignores the fact that the effect is different between different nodes in a SDG graph. To solve this problem, a method which added the quantitative information of the effect between nodes into the qualitative SDG reasoning by granularity entropy was proposed, according to which the priority could be given to the multi-potential fault origin and the resolution could be improved. In this method, the value of the effect can be obtained by calculation of attribute state values in fault diagnosis decision table. The shortcoming of the previous fuzzy SDG fault diagnosis method can be avoided.