Reactive Power Optimization Of Regional Power Grid With Wind Farm

Wind power is one of the Renewable energy technologies that has relatively more mature technology, lower cost and develops rapidly. China, with a vast territory, long coastline, has rich wind resource. The proved usable wind energy is at least approximately one billion kilowatt, equivalent to more than40Installed capacity of Three Gorges Hydropower Station.On one hand, because of cleanliness without any pollution and limitless, wind power has wide development prospects compared with traditional fossil fuel energy; on the other hand, integration of large-scale wind power brings great challenges to security, stability, economic operation and power quality of traditional power grid. Especially large amount of wind power integration on power system may cause frequent changes of power flow, leading to the original reactive power optimization unsuitable for the existing situation any more. So a new power generation and operation plan is needed considering the character of wind power.In the actual system, load changes with a certain regularity and usually has small fluctuation in a certain period; however, wind speed changes in a weak regularity, and even may change dramatically within a short period of time. Also the adjusted times of on-load tap changing and capacitor is limited in one day, so it is not suitable to do static reactive power optimization according to a certain moment of load and wind levels. On the contrary, it is necessary to take into account the optimization effect within a certain period of time scale comprehensively considering the changes of load and wind speed. Dividing-stage strategy is used to handle the randomness and volatility of wind speed. First divide the research period into several parts according to the prediction of wind speed, then get the wind power expectation of every part to be used for optimization calculation.Reactive power optimization is a very complex non-linear mixed integer programming problem. Optimized object includes discrete variable such as adjustable transformer tap, capacitor banks and continuous variable such as generator reactive power output. So far domestic and foreign scholars have proposed a variety of methods to solve this problem. These methods can be summarized as the traditional algorithms and artificial intelligence algorithms. Each algorithm has its own advantages and disadvantages; it is a new direction of reactive power optimization to look for a hybrid optimization algorithm combining two or more algorithms. In this paper traditional algorithms and artificial intelligence algorithms are combined to solve the problem. The original problem is divided into two sub-problems:discrete optimization and continuous optimization. First adjustable transformer tap and capacitor banks are optimized by PSO; then generator reactive power output is optimized by interior point method with discrete variable not changed, and then return to the discrete optimization phase, iterate like this until the program converges. Respective advantages of the traditional algorithms and artificial intelligence algorithms are used sufficiently.As a new software development method, mixed-language programming can make full use of respective advantages of different programming language, improve development efficiency, and increase the flexibility of programming language. Use PSO-interior point algorithm to solve the mathematical model of dynamic reactive power optimization based on dividing-stage strategy. Python and Matlab combined with PSS/E are used to analysis IEEE30bus system and northern grid of Dongying. The results show that the proposed method is feasible, effective and suitable to solve reactive power optimization of regional power grid with wind farm.