The Application Of Fruit Fly Optimization Algorithm In Distribution Network Planning

Distribution network planning is an important component of power system plan. Rational plan can not only bring social benefits, but also bring the power system and the users of objective economic benefits. Distribution network planning can be attributed essentially to a series of mathematical optimization problem, so the problem of modeling directly affects the quality of distribution network planning programs.Firstly, the Fruit fly optimization algorithm is studied and analyzed for its shortcomings prone to premature and slow convergence of iterative late. Then we put forward the corresponding measures for improvement, the formation of new fruit fly optimization algorithm. First, a moderate step search algorithm has been improved step change in design of adaptive algorithm to change with the increase in the numeric of iterations step through the function test found that improved early fruit fly optimization algorithm to expand search range, while in later iterations with decreasing the step size is more conducive to find the optimal solution; second, in order to prevent the premature introduction of a mutation similar genetic algorithm, by making "escape" individual flies to escape from local optima solution, to a certain extent, prevent premature phenomena.Then we applied the improved fruit fly optimization algorithms in the distribution network. It found that the traditional distribution network planning models have ignored the limited numeric of distribution substations outlet, this will determine the location and distribution substations and distribution networks are considered separately formed. For this situation, we propose a load distribution network planning model based on the partition, the first based on the amount of the network load to determine a reasonable numeric of lines, and the use of K-means clustering algorithm is a regional breakdown of the network load. Then Fruit fly optimization algorithm with improved distribution substation location in a two-step iterative process evaluation Fruit fly individual:(1) based on the location and distribution substations each partition set a source load, based on the "first branch trunk " the principle source load as a starting point to establish a network connection between the sub-region load;(2) the total network load superimposed on the source partition load distribution substations to establish the connection with the source of radiation load between. Furthermore, in order closer to the actual situation, this article will introduce geography programming model through penalty factor and reward factor to influence the cost of investment planning solutions.In this paper, in order to verify the validity of the model and optimization algorithms, applications based on AutoCAD Object ARX secondary development, design and implement a simple distribution network planning system. Using pixel information integration model, so in AutoCAD points, lines and other elements of both geometric and electrical equipment information, including load information management system, a special area management, planning and process management, planning and program management and basic data management module. System operate in common, planners need to master the use of AutoCAD to complete the planning process, with strong practicality.