Distribution Network Planning Based On Improved Ant Colony Algorithm

As one of the important issues of the planning of the power system, the power distributionnetwork planning is not only a hot but a difficult issue which deeply attracts widespreadattention of many domestic and foreign experts and scholars. Terminal as the whole powersystem network, it is directly connected to the consumers, the high quality of its constructionnot only directly related to the stability and security of the society, but also the health of theeconomic development. The distribution network planning is generally mainly related to threeaspects, which are Grid planning, Reactive Power Optimization and Reconfiguration of thedistribution network. Its purpose is to reduce the abnormal loss of power energy, achieveenergy conservation and improve the quality of power supply. Distribution network affectsaspects of the society, as a correspondence, its planning problems is also limited by manysocial factors. So this planning problem is a combinatorial optimization problem. At this stage,the size of the distribution network has become bigger and bigger, the complexity and difficultyof its planning is significantly increasing.Based on the ant colony algorithm, this paper analyzes the distribution network structureof the power system grid, mainly focuses on the following works:(A) This paper expounded the purpose, importance and present status of the distributionnetwork planning; followed by the instructions, analysis and comparison of several commonlypower flow calculation method algorithm which used in the distribution network. For thespecial structure of the distribution network and the demands of different power flowcalculations for different radiation network, the author selects the trend forward and backwardsubstitution algorithm and made some improvements in the power flow calculation process.Through two examples of the power flow calculation whose simulation based on computerconfirmed that the improved power flow calculation algorithm is more suitable to solvedistribution power flow calculation.(B) The ant colony algorithm has characteristics of strong robustness and positivefeedback which other algorithms has none, and its optimization mechanism is very suitablefor solving the combinatorial optimization problem. With the famous TSP problems, theauthor studied the basic flow of the ant colony algorithm, and made the decision of selectingthis intelligent algorithm as the calculation method to solve distribution network planning inhis paper. Taking special characteristics of the algorithms’ optimization into account, thispaper introduced the ideas of self-adaptation and minimum-maximum algorithm, andproposed that making the average ant number on the path associated to the pheromone update mechanism for achieving the purpose of improving the way of pheromone update which tomake up for the slow speed of algorithm optimization and easy to fall into local optimum inthe process of distribution network planning.(C) According to the characteristics of distribution network structure, this paperartificially reconstructed the optimization solution space of artificial ants, so that it caneffectively applied to solve the distribution network structure planning optimization problem.Corresponding to the different constraints have different distribution network planningmathematical model, the model this article established is to seek the lowest cost plan in theplanning year grid target. And then the author selected the algorithm of this paper as theoptimal planning algorithm.Finally, this paper selected a10kV distribution network structure planning example toverify the improvement work above. Based on the Matlab.R2010b system computer simulationexperiments, the experiment results show that the improved algorithm can effectively solve theproblem of distribution network structure planning and optimization, not only can produce abetter planning result which saving the cost of distribution network planning, but also havecharacteristics of speed faster and better convergence performance in the optimization.