Research On Optimal Planning Of Distributed Generation Based On SAGWO Algorithm

With the deepening of the energy revolution, the construction of clean and reliable distributed energy system is the general orientation of development during the new era which is of significance for improving the reliability of power supply, optimizing the energy structure of China, and enabling sustainable energy utilization. Hence, a reasonable distributed generation planning of placement and capacity becomes more important in this format.This paper firstly researches the intelligence algorithm which is suit able for distributed generation. And the paper analyzes the principle of a new intelligence algorithm called grey wolf algorithm in details, then discusses an improved grey wolf algorithm based on simulated annealing algorithm. Trial functions are tested through MATLAB to verify the feasibility of Simulated Annealing Grey Wolf Algorithm, and experimental results show that the algorithm has well convergency efficient, ability to escape from local optimum and fast convergence rate.In the process of distributed generation planning, the interconnected locations of distributed generation are directly defined in general. This conventional method may not accurately reflect the characteristics and requirements of practical power system. This paper takes advantage of the average improved voltage level index of distribution netwok and maximum penetration level to choose the installation nodes of distributed generation. Paper also uses the timing characteristics of distributed generations’ output and load demand to determine the effectiveness of siting method. The simulation for IEEE-33 bus system proves that the planning of interconnected locations can achieve the expected aim.On the basis of choosing the interconnected locations of distributed generation, this paper carries on research on distributed generation’s size. Renewable energy which are represented by photovoltaic power and wind power exist uncertainty and time-varying. An approach of time segments division is put forward to dispose the time-varying characteristic of distributed generations and load. At every time period, the modeling method of random variable based on Latin Hypercube Sampling are put forward considering the uncertainty of distributed generations and load. The probabilistic load flow is adopted by using the Monte Carlo sampling method. The mathematical model of goal function is established by considering the economic benefits of operators of distributed generations and distribution company. The experiment result indicates that the method for siting and sizing of DG can make full use of distributed generations’ advantages and promote DG development. The improved grey wolf algorithm can improve running speed and results which has good application prospects.