Research On The Optimal Of Distributed Generation And Island Detection

In recent years, the earth’s resources increasingly exhaustion and the environment deteriorating, especially since2013, the atmospheric environment in our country is serious, same as the haze. The haze of Beijing, north China and east China regional, sounded the alarm for the serious air pollution. Premier Li Keqiang chaired a conference on energy conservation and emissions reduction and climate change, from eight aspects promoting energy conservation and emissions reduction and lowing carbon work in2014. One was to promote clean energy, namely, to speed up the development of nuclear power, wind power, solar, etc. Wu Xinxiong, head of the national energy administration said that we will increase photovoltaic6million kilowatts in2014. The development of distributed power supply and distributed power grid is imminent. The accessing of distributed power supply, not only reduce the pollution of the environment, but also improve the stability of power system and flexibility. However, need to optimize the access of the location and capacity of the distributed power. And also need to detect the system of the island accurately and effectively, and cut off it in time, to ensure the normal operation of the power grid. Thus optimized configuration of distributed power and island detection had a practical significance.Firstly, the basic concept of micro-grid and it’s characteristics were analyzed. And the basic structure of the micro-grid, system configuration and operation mode were described, then analyzed the island operation and grid operation.Secondly, the characteristics and basic principles of three kinds of distributed power which include of photovoltaic, wind turbines and gas turbines were analyzed, the output of each model of distributed powers and the multi-objective optimization model which included of the cost of construction, environmental factors and net. loss were established. The deficiency of the existing improved particle swarm algorithm was analysis, and proposed a linear gradient of learning factor and adaptive particle swarm optimization algorithm-LD-SAPSO. The results showed that LD-SAPSO not only convergence speed and search a wider range, and can keep the population diversity. As IEEE33for a example, used the algorithm of LD-SAPSO to optimize the allocation of distributed power and to select the optimal location and capacity.Then, the mechanism of power island and the danger of island effect were discussed in the paper, found that the most difficult to detect islanding was a parallel RLC load, mainly because of the parallel resonance, and the most serious was that when islanding occured, the parallel resonant frequency was equal to frequency of large power grids. Therefore, written in MATLAB parallel RLC load phase frequency characteristic curve program, the paper found that the loaded quality factor Qf was higher, the probability of failure was the islanding detection. Then the advantages and disadvantages of three existing islanding detection method was analyzed in the paper.Finally, based on the advantages and disadvantages of various island detection methods proposed improve the active frequency shift of the island. And through the QfO×Cnorm DNZ, the superiority of the improved algorithm was verified and selected the appropriate parameters. On this basis, builded a distributed power grid silos test simulation model in the Matlab/Simulink. Tests showed that the test time is not obvious difference for inductive load; the improved AFDPF algorithm could detect island phenomenon after0.202s,and AFD algorithm needed0.485s for capacitive load. In comparison, the improved AFDPF algorithm can more quickly and reliably detect the island.