Approach for Mining Fault Rules of Power Grid based on the Combination of Rough Set Theory and Association Rule

With the increasing of fault information transmission capacity in power grid, the volume of information which needs to be concerned by dispatchers has greatly increased, consequently making it difficult to identify the fault signal and analyze the cause of the accident quickly for dispatchers in massive fault information. To settle this problem, this paper uses a novel approach that combines rough set theory with association rule for mining fault rules in a large number of historical fault data of power grid. Firstly, it builds distributed original decision tables according to regions. And then it uses the information entropy algorithm in condition attribute reduction. Lastly, it applies the improved Apriori algorithm of association rule to fault rules mining based on the reduction decision table. In this way the problems of redundancy of massive fault information can be solved and complexity of rules extraction can be simplified effectively. It also improves the efficiency of fault rules mining.