The Research For Idendifying The Vulnerability Of Power Grid Based On Multidimensional Association Rules

With the accumulation of diversity, multiplicity and uncertainty of power line malfunction, large-scale power outages have occurred frequently. However, it could dig out potential hazards and to develop relevant emergency measures, playing an important role in decision-making for the government departments and corporations. As the traditional ways, data-mining show inadequate in multidimensional and inefficient shortcomings, How to use the scientific and effective methods to resolve the grid of potential problems in the current. And to provide powerful tools to solve problems has become an urgent social needs for the electric power personnel.Aiming at a series of problem in traditional power network, such as to consider a number of factors insufficiency, The rules of the system are lack of variety, The setting of parameters and threshold is imperfect, Algorithm efficiency and accuracy are not higher. On the basis of study of fault attribute model in depth and to pursue a more simple, efficient, intuitive and accurate method for power system fault recognition. The technology of multidimensional Association rules as well as the analysis of electrical power system vulnerability has been launched the research, specifically as follows:1. Clarify the basic elements of association rules, By combing with the specific algorithm implementation of the Apriori and FP algorithm to analysis them, The advantages and disadvantages of the algorithm are presented, then the method about the optimization of the algorithm also has been proposed.2. Based on the classical decision algorithm for mining Association rules, the method of multidimensional Association has been proposed, which the way based on optimal frequent item sets, Through the analysis of the single dimension attribute, dimension between dimension attribute, and multidimensional attribute to obtain the maximal frequent item sets.3. The method of FP-Growth which based on the optimal frequents item sets is successfully applied to the identification of power grid line. By mining the association rules between the factors of grid line to provide reliable basis for the fault prediction and early warning for power system.4. Through the analysis of specific examples,the method about the feasibility and the validity have carried on the confirmation. The result indicated that this method has good practicability in the risk disaster forecasting and early warning of power network system. It can effectively avoid the high information redundancy generated by the conventional method, And to guarantee the accuracy and objectivity of the association rules to the maximum extent.