Based On AHP And Intelligent Algorithm Of Grid Companies Operating Risk Assessment Research

With the deepening of electricity reform, for grid company that economic benefit will be more and more important, In order to maintain the safe operation of electric system and the sustainable development of grid company, it is especially necessary to evaluate the operational risks faced by grid company.In this paper, firstly we have a detailed analysis and research of the risks in the business process of grid company, and design a set of risk evaluation index system. Based on the risk evaluation index system and combining with AHP, ordering, cluster analysis, and interactive methods, we empower the risk indexes, have a comprehensive evaluation of the operational risks and give the quantitative results. On the one hand, the method need not to judge the consistency and reduce the computation; On the other hand, it avoids the impact of man-made and subjective factors on weight to the greatest extent.Secondly, we have a detailed analysis of the risk of purchasing electricity of grid company, and establish a model of the risk to purchase electricity taking the minimum conditional risk value and the maximization expected revenue as goals. This model make up the flaw of VaR not to be able to reflect the loss rear part information, guard against the small probability extreme risk, reduce the possibility of grid company to have the disastrous risk, and do not need the priori knowledge.Finally, aiming at multi-objective particle swarm algorithm insufficiency, and proposes an improved multi-objective particle swarm algorithm. By introducing the idea of local disturbance and variation operation to improve local search ability of MOPSO, and with the idea of non-dominated sorting genetic algorithm for the maintenance of an external file. Through experiment contrast, this algorithm can converge to the optimal front end while maintain the diversity. We use the improved intelligence algorithm to solve the model, each time a set of optimal solutions can be calculated, and the optimal solution evenly distributed in the optimal front end, which provide a reference of correct decisions for policy-makers.