The Research Of Distribution Network Planning Based On Improved Genetic Algorithm And The Software Development

Genetic Algorithm (GA) is an adaptive and probabilistic search method of global optimization. Because it does not rely on the specific area of the problem and has high robustness, the research about GA becomes in-depth in power system planning in recent years. However, in engineering applications, the distribution system network developed by GA has some problems, such as heavy preparing work in primary planning. Moreover, it’s difficult in reflecting the nature of programming problems and combining adequately the information in the planning procedure.According to the current research result, the thesis presented a new real-number encoding based on the numbers of load nodes in planning area and short feeder lines collection as the essence to execute distribution network planning automation. Relying on the improved CHC strategy, the genetic operation adopted improved Fragment Reordering Crossover operator considering the shift operator and Swap Mutation operator. For the stability of the optimizing process, the mutation took part in the genetic operation just in the late iteration. The whole process blending the decoding and the characteristic of distribution network’s structure ensured each solution efficient and reasonable. And the (μ+λ) selector that owns highest selective pressure guaranteed the rate of convergence. In addition, basing on the principle of open-loop operation and closed-loop design, the improved GA took the tie-lines and link switches into the consideration of final total investment on the basis of the optimal distribution network structure and validated the optimal network reliability and the link switch selection plans in the crucial line fault.On the theoretical basis above, I have accomplished the programming based on the MFC to realize the function of distribution network planning and developed the GIS software basing on the ArcEngine platform to make the result visualization, the spatial geographic information storage and the rendering of power flow.Summing up the above, the optimal planning scheme doesn’t need the collection of selected lines and can reflect the essence of the distribution network planning. The optimal result has certain actual significance. Finally, the analyzed case proved the automation strategy feasible and practical.