The Research Of Multi-objective Wind Farm Reactive Power Optimization Based On Improved Artificial Fish And Genetic Hybrid Algorithm

With the increase of wind turbine grid capacity, how to maintain wind farm grid system voltage stability has become a very important problem that people attention to. The main purpose of the wind power system reactive power optimization is how to optimize the grid reactive current distribution by adjusting the control variables and reactive equipment and to optimize the distribution of system reactive power current, thereby reducing system active network loss, improvising voltage quality and increasing the voltage stability.The conventional reactive power optimization mathematical model of the main system takes minimum network loss as objective function, in order to realize the purpose of economic operation of the system. But each node voltage value in the optimization results can lead to reactive power output close to the limit. It contradicts the system voltage security. In this paper, considering the economy and safety of power grid operation, putting forward the multi-objective reactive power optimization mathematical model. The paper considers both the system active network loss, the voltage quality and voltage stability.Wind power system reactive power optimization problem has characteristic of multiple variables, multi-constrained conditions, mix of continuous and integer variables, etc. It belongs to nonlinear optimization problems. Solving the problem with traditional mathematical programming method has great difficulty. Artificial fish swarm intelligence is a kind of new optimization algorithm, is simple in principle, parameters, less and faster convergence speed. This paper has in-depth study of artificial fish algorithm and its application in wind farm reactive power optimization,uses artificial fish algorithm and hybrid algorithm of genetic algorithm and improved it. It is applied to solve the multi-objective reactive power optimization problem.This paper introduces the basic genetic algorithm and artificial fish genetic hybrid algorithm of the basic principle and method, and then introduces the non dominated sorting multi-objective artificial fish algorithm and its improved algorithm involves the basic principle and process of the algorithm. Multi-objective hybrid algorithm adds the global optimal information to population location state changes and introduces the devour behavior in order to improve of artificial fish algorithm global searching performance, increase the calculation precision of the algorithm, speed up the convergence speed and reduce the complexity. It has improved algorithm global performance and avoided individual trend of assimilation, precocious, easily trapped in local optimum, ignoring the global optimal solution。The algorithm selects the optimal solution in the Pareto optimal solution concentration by using a kind of multiple attribute decision. This article lists the considering network loss and voltage deviation and voltage stability margin for multi-objective reactive power optimization mathematical model of wind farms. This paper makes a simulation test on the system IEEE14 nodes. By comparison, improved artificial fish genetic hybrid algorithm and optimization is better than the basic fish algorithm and NSGA‐ Ⅱ, and improved artificial fish genetic hybrid algorithm on the convergence stability and efficiency has common advantage.This paper get the following results:(1)Using the multi-objective reactive power optimization mathematical model with comprehensive consideration of economy, security and stability of wind farm system to do wind farm reactive power optimization.(2) Artificial fish and genetic hybrid algorithm is given, and the hybrid algorithm is further improved. It simplifies the structure, overcomes the local convergence, and enhances the ability to search the global optimal value.(3) Mix improved artificial fish genetic algorithm is applied to multi-objective reactive power optimization of wind farms.(4) Through the analysis of the results of simulation data, it proved that improved artificial fish genetic hybrid algorithm is simple, easy to implement, overcoming the local convergence, enhancing the ability of getting the global optimal value, having good convergence performance...