Study Of Short-term Prediction And Multi-objective Optimal Dispatching Of Grid-connected Wind Power

Wind power has strong intermittency and fluctuation, the proportion of wind power capacity increases in the power system, which will bring some negative influences to the grid, and to a certain extent, affect the power quality, such as voltage sag and swell, harmonic frequency deviation, and lowering the reliability of system scheduling operation while increasing the cost. To increase the forecast level of wind power and to establish and solve the logical grid connected wind power scheduling model both can improve the wind power utilization level and reduce the impact of wind power integration. Under this background, based on the research of wind power forecasting and multi objective power system dispatch optimization with wind power connected to the grid, the main work of this paper is as follows:(1) Starting from the characteristics of wind power, a new kind of wind power forecasting model based on the ensemble empirical mode decomposition combined with the wavelet neural network is proposed. First, using the ensemble empirical mode decomposition to reduce the non smoothness of wind power signal, second, mining the chaotic characteristics of sub sequence signal with phase space reconstruction, then using wavelet neural network for each sub series to model and predict, last superimposing the predicted results of each sub sequence to obtain the final results. The case study shows that the the proposed combination prediction model has higher prediction accuracy and great potential in engineering application.(2) For the shortcoming of the traditional modeling progress which ignored the prediction model learning effect, combined with adaptive particle swarm algorithm and learning effect feedback mechanism, a wind speed forecasting model combined ensemble empirical mode decomposition and least square support vector machine is established. First, using the ensemble empirical mode decomposition to break out the real different scale trend or fluctuation in wind speed signal step by step, then constructing the forecast model of least squares support vector machine for the decomposed sub sequences respectively, and integrated optimizing the model parameters based on adaptive disturbance particle swarm algorithm and model learning effect feedback mechanism. At the end of each sub sequence prediction results are superposed to obtain wind speed forecasting value. The case study shows that the combined prediction method has good prediction effect. (3) An elite cloud variability multi-objective backward learning particle swarm algorithm is proposed to solve the multi-objective grid connected wind power system dispatching. By by introducing alternative set to get a historical Pareto optimal solution set and the global Pareto optimal solution set, to improve the global search ability of the algorithm; by introducing elite cloud variability and backward learning to improve the performance of the particle swarm algorithm, to reduce the probability of the algorithm getting into a local optimum; then via perturbation operation to find the true Pareto front; last selecting compromised optimum based on the theory of fuzzy. This algorithm was used to the scheduling and testing system of an IEEE6-machine30-node, the simulation results verify the validity and the advance of this algorithm. At the same time, the research results of this paper believe that in the planning of wind power capacity, more needed is to consider the security and stability of power system as well as the environmental, economic and other factors.(4) Power system dynamic environmental economic dispatch considers the ties of each time section, to simultaneously optimize the minimum fuel costs of the conventional unit and the minimum pollution emissions as goal. That is more in line with the actual operation of power system dispatching. However, there is little literature and Research on power system dynamic environmental economic dispatch when wind power connected to the grid. Under this background, this chapter explores the use of an artificial fish swarm algorithm to solve the model of multi-objective quantum based on the establishing the model of power system dynamic environmental economic dispatch when wind power connected to the grid In order to achieve the comprehensive optimal.(5) In this paper, a mixed Pareto Optimal Solution Set method is proposed based on Elite-Cloudy Mutation Multiobjective Improved Oppstion Particle Swarm Optimization and Mutiobjective Quantum Atificail Fish School Algorithms, whose basic methodology is parallelly optimizing the two algorithms to get the optimal solution sets then mixing and modifying these sets to get a new Pareto optimal solution. This algorithm has been examined and applied to optimize the economic dispatch model of Guangxi power grid and got good optimal results. The proposed algorithm has been compared with the original dispatch scheme. The result clearly shows that the proposed algorithm is able to produce good economic and social benefits in that it reduced55535.91yuan of power purchase cost2667.9533tons of CO2emission and22.30368tons of SO2emission on the experiment day.