Multi-objective Reactive Power Optimization Of Power System With Random Wind Power Generation

Reactive power optimization is reasonably adjusting related equipment to optimize the distribution of reactive power to reduce active power loss,improve power quality, and strengthen the economy, stability and security of power system. Multi-objective reactive power optimization is a multi-variable, multi-constrained nonlinear programming problem, and there is generally game among its multiple targets, so it is difficult to achieve the optimum at the same time. With intermittent energy sources such as wind power infiltrate power grid gradually, wind turbine output is uncertain caused by the randomness of wind speed, which is making the problem more complicated and transforming the problem of reactive power optimization with wind power into a probabilistic optimal power flow.The main difficulties of multi-objective reactive power optimization with wind power are that the wind power uncertainty output and the selection of a suitable optimization algorithm. To solve this problem, this paper introduces two methods.The first method, using stochastic programming theory, the Latin hypercube sampling method is used to generate a random sample stratified wind speed scenarios to active power loss and node voltage deviation minimum optimization objectives and build multi-objective reactive power optimization based on chance constrained programming mathematical models. Under the premise of satisfying a given objective function and state variable constraints confidence level, a decomposition-collaborative thinking of co-evolutionary algorithm is introduced. The reactive power optimization problem is decomposed into a series of sub-optimization problems of mutual cooperation, according to the different types of control variables which divided into different populations within the various populations, which are used to optimize by genetic algorithm optimization separately, and cooperation between populations has led the evolve throughout the whole ecosystem.The second method, a new concept of optimal scenario analysis method is proposed, which is to find out the most representative of the best scenes by Latin hypercube sampling method to generate random scenarios to replace other scenes. Interior point method(IPM) is suitable for solving the continuously differentiable problem, while genetic algorithm(GA) is suit to solve the discrete optimization problems, what is more, continuous variables and discrete variables coexistence in the problem of reactive power optimization, so this paper propose a hybrid algorithm based on interior point method and genetic algorithm, which the original reactive power optimization problem is decomposed into a continuous optimization sub-problem and a discrete optimization sub-problem, which are solved by IPM and GA respectively and take advantage of each other.In addition, the paper also discuss the nature of the wind power generations node and flow calculation method under different control modes. The paper takes the IEEE-30 bus system with three kinds of wind farms as example, which can be solved by these two proposed methods. The simulation results show that the proposed model and methods in this paper are reasonable and effective.