Research On Reactive Power Optimization Of Distribution Network With Distributed Generation

Reactive power optimization of distribution network is a kind of reactive power regulation method under a certain system operating mode and constraint conditions to achieve a single or a plurality of performance indicators optimal, so can implement optimizing the allocation of reactive power, keeping the balance of the reactive power, achieving economic operation of distribution network. The traditional reactive power optimization of distribution network is mainly to adjust the tap position of on-load tap changer transformer and the number of shunt compensation capacitors.In recent years, the problem of energy and climate change is becoming more and more serious, the distributed generation technology has attracted wide attention in the world because of its environmental protection and economical efficiency. The access of distributed generations has changed the mode of power flow in the traditional distribution network, some types of distributed generations can regulate reactive power output (or absorption) within a certain range and is continuously adjustable, thus able to participate in reactive power optimization. Characteristics of distributed generations may have some impact on power flow calculation and reactive power optimization of distribution network, therefore, the research on reactive power optimization of distribution network with distributed generations is of great significance.The thesis first studies the processing method of different distributed generations in power flow calculation, divide them into four kinds of node types, participating in the process of backward forward sweep.The power flow calculation program was written and verified with typical system of IEEE 33-bus, and studied the effects that the type of distributed generation, location of the access point and output of reactive power may have on node voltage and active power loss of the system. Result shows that:distributed generations that output reactive power are better at increasing node voltage and reducing active power loss than distributed generations that absorb reactive power; Distributed generation which access point is far away from power source,at long branch and lower voltage level node is better at increasing node voltage and reducing active power loss; The active power loss decreases with the increase of the compensation capacity, but beyond a certain value, it will increase with the increase of the capacity.As for reactive power optimization of distribution network with distributed generations, because of the genetic algorithm may converge at local optimum solution, and cause the "stagnation". To avoid these problems, this thesis adopts the self-adaptive probabilities of crossover and mutation, self-adaptive population size and optimization of initial population. Based on these approaches, the reactive power optimization program was written and verified with typical system of IEEE 33-bus. Result shows that:these approaches can availably solve the problems of reactive power optimization of distribution network with distributed generations, improve calculation speed, make the result meet the requirement and solve the problem of "stagnation" in the late process of the algorithm effectively.