Research On Voltage Optimization Method Of Distribution Network Based On Parallel Improved Genetic Algorithm

In recent years,the extreme natural disasters such as ice storms and snowstorms occurred frequently which resulting in large-scale power outages for a long time and bringing economic losses to society.Hence,it is essential to enhance the resilience of power systems to cope with such events.As the distribution network directly participates in power supply to a number of critical loads,how to optimize the operation of the distribution network to ensure the quality of power supply for critical loads after disasters is of great significance to improve the resilience of power systems.In this dissertation,the distributed generation and controllable equipment are used to optimize the distribution network.According to the characteristics of optimization problem of the distribution network,the solution method is studied and improved,and the parallelization of the algorithm is realized to reduce the calculation time,so as to enhance the ability of power systems to resist disasters and recover after disasters.An optimization model is proposed to maximize the ability to supply power to critical loads in distribution networks,which coordinating the operation of different types of devices,such as the energy storage system,capacitor banks,and so on,to optimize the voltage of critical loads,active power loss and cost of devices.Genetic algorithm is selected to solve the model,which is a mixed integer nonlinear programming model.And the algorithm is improved by self-adaptive distribution index of crossover and mutation,expanding the number of population and two-stage solving method,which overcomes the drawbacks of standard genetic algorithm,such as slow convergence,low precision and easy to plunges into the local optimum.In view of the optimization of distribution network operation requests to adjustof the output of controllable devices rapidly,which requires high computing performance.In this dissertation,GPU is applied to accelerate algorithm to reduce the computing time.The parallelism of the improved genetic algorithm is analyzed,and the storage space and block structure division of GPU are studied based on the algorithm.The procedures of improved genetic algorithm,such as initialization population and fitness calculation,are parallelized by GPU.The correctness of improved genetic algorithm and the acceleration of GPU are tested in the IEEE 33-bus and IEEE 118-bus system.The results prove the correctness of the algorithm and the improvement of the convergence speed and precision,GPU can accelerate the solving speed of the algorithm and reduce the calculation time.In addition,the optimization of the voltage quality of the critical loads is proved by examples,which can effectively enhance the resilience of power systems.