Research On Reliability Evaluation Method Of Distribution Network

The distribution network is a bridge connecting the transmission system and users.Researching the reliability assessment method of the distribution network is of great significance and practical value to improve the reliability of the power supply.In order to make the distribution network reliability evaluation results closer to the real situation,this paper mainly takes the distribution network reliability evaluation method as the research object.Several improved schemes of reliability evaluation methods for distribution network are proposed,and these methods are compared and analyzed.The main results of this article are as follows:(1)Because the Monte Carlo method in the traditional method is more flexible,it is suitable for simulating random changing load characteristics,and it is also recognized as a more accurate algorithm in the traditional method.In this paper,the Monte Carlo random sampling method is used first to establish a reliability evaluation model of distribution network based on time-series Monte Carlo simulation method,and the relevant indicators of distribution network reliability are calculated.(2)With the help of neural networks,they have the characteristics of massively parallel processing and self-learning.The BP neural network optimized by genetic algorithm and particle swarm optimization algorithm is applied to the reliability assessment process of distribution network.It is less difficult to build a reliability assessment model for distribution networks using neural networks.By appropriately transforming the structure of the distribution network,it can be calculated and evaluated using neural networks.This method can overcome the problem of heavy calculation of Monte Carlo method,which improves the evaluation efficiency.(3)In order to fully mine the complex characteristic relationships in the distribution network data,this paper proposes a new method of applying deep neural networks(deep belief networks)to the reliability assessment of distribution networks.Considering that the particle swarm algorithm can well optimize the parameters of the neural network and has the effect of global approximation.In this paper,particle swarm optimization is used to adaptively adjust theparameters of the deep belief network model.A deep belief network is found that is suitable for reliability assessment of distribution networks.The reliability evaluation results calculated by the found optimal deep belief network model are compared with the reliability results calculated by Monte Carlo simulation model and optimized BP neural network model.The comparison results show that,compared with the shallow neural network and the traditional Monte Carlo method,the deep neural network can more effectively and accurately express the internal characteristic relationship between data samples after learning the distribution network data.Compared with the other two methods,the deep belief network model optimized by the particle swarm optimization algorithm has achieved a good improvement in both the accuracy and computation speed of the reliability assessment of the distribution network.It shows the superiority and advancedness of the deep belief network of particle swarm optimization applied to the reliability assessment of distribution network.