The Research Of Dynamic Economic Dispatch Integrated With Wind Power System Based On Chemical Reaction Optimization Algorithm

Wind power generation is a renewable energy power generation which has the largest development prospect today,and its technology has been relatively mature.In order to make the most of wind power,and reduce the operating costs of the power system,it is significance to find a more reasonable dispatching model and a more effective solution for the power system.Based on the characteristics of the wind farm and the requirements of the system optimal operation,this thesis analyzes the economic dispatching model with the wind farm and the solution of model.Firstly,this thesis introduces the traditional dynamic economic dispatching model,and adopts chemical reaction optimization algorithm which has a capability of global search to solve the model.At the same time,an adaptive penalty function is used to optimize the constraints.It can dynamically adjust the penalty factor based on the degree of meet each constraint,to prevent the penalty value from becoming too small or too large.Through simulation on the four different dynamic economic dispatching models,the algorithm is validated its effectiveness in the field.Secondly,the instability influence of power system caused by wind power generation are analyzed.According to the characteristics of wind power such as fluctuation and uncontrollability,the deterministic dynamic economic dispatching model is formulated by introducing the increase and decrease spinning reserve capacity in consideration of wind power penetration.However,the chemical reaction algorithm has defects in convergence rate,a new elite opposition-based learning chemical reaction differential evolution optimization algorithm is proposed.As the adaptive differential evolution algorithm has advantages of simple structure,fast convergence speed and less control parameters,and the characteristics of the elite opposition-based learning strategy,they are combined and applied to the dispatching field.At the same time,multiple-objective optimal problem is converted into simple objective optimization by using a virtual solution theory which is based on the distance between multi-objective solution on the virtual ideal solution.The improved algorithm has better convergence speed by the simulation and comparing with the original algorithm.Finally,according to the deterministic dynamic economic dispatching model with wind power,this thesis propose an model which combines the spinning reserve constraints and randomness.The stochastic fluctuation of wind power is simulated and analyzed by the scene method.By using the way of Latin Hypercube Sampling,the prediction value of wind power active are sampled,and the sampling scenarios are reduced by the fast forward reduction technology of probability metric.The integrity of wind volatility can be expressed in the reduced scenarios.The randomness dynamic economic dispatching model is solved by the elite opposition-based learning chemical reaction differential evolution optimization algorithm.Simulation results show that the established model is effectiveness and economy.