Research On 110kV Distribution Network Optimization Based On Improved Genetic Algorithm

The safe and stable operation of power grid is dependent on the rational distribution network structure.For effectively reducing construction investment,network loss and power grid failure,we can optimize the distribution network structure to improve reasonable use of resources and achieve reduction of the maintenance and protection of lines and equipment,as well as the cost of construction investment effectively,meanwhile the economic benefits of state and power enterprises can be thus improved.In recent years,most of researches focus on the medium and low voltage distribution network,however,110 k V distribution network is an important link for up-down connection,whose investment and operation and maintenance cost occupy a heavy proportion.Therefore,this thesis carries out the optimization study of 110 k V distribution network structure based on a modified genetic algorithm.Firstly,the safety and stability calculation and analysis of the current status of the planning area is carried out herein based on the power system analysis software PSD-BPA.The main characteristics and stability levels of the system stability problem are determined,and the overall optimization scheme of the network structure is designed.Furthermore,in order to realize reasonable division of power supply area and solve cross-regional problems,a new distribution network optimization zoning strategy is proposed by employing the discrete binary particle swarm optimization(PSO),which can not only meet the economic objectives of operation loss and construction cost,but also consider the reliability.A mathematical model is established with target of the shortest load path and constraints of the capacity load ratio and power supply radius.As a result,the proposed model is applied to solve the zoning scheme of planning area,and compare with Tabu algorithm,the result shows that the proposed zoning algorithm has superiority and feasibility.Furthermore,a single stage optimization model of the grid structure is established,and a simulated annealing algorithm based modified genetic algorithm is proposed with target of minimum comprehensive annual costs and constraints of reliabilities of voltage drop,radial network,etc.,for solving the planning area after partition and thus obtaining the optimal network structure.The optimized network structure is analyzed and compared with the basic genetic algorithm and immune genetic algorithm,the results illustrate that the proposed modified genetic algorithm has obvious advantages in convergence speed andtarget economy,as well as ideal practicability and effectiveness in solving the optimization problem of 110 k V distribution network.