Research On Power Forecasting And Optimal Dispatching In Wind Farm Incorporated Power System

Replacing fossil energy with renewable energy is the fundamental measure to solve the environmental pollution and energy crisis that humanity faces at this stage.According to the current situation,in the practice of renewable energy,wind power generation technology is becoming more and more mature,and the low cost of power generation is easy to achieve large-scale industrial development,and has gradually become the main force for sustainable development of energy and the environment.However,due to the influence of natural factors,the intermittent and volatility of wind power are more obvious,which in turn makes it difficult to grasp the uncertainty characteristics of grid-connected wind power generation.At the same time,the control and scheduling of power systems has great challenges.The gradual improvement of wind power forecasting levels and the rational construction of an optimized power dispatching model for wind power integrated system can greatly increase the level of wind power utilization,which in turn greatly reduces the impact on the power system after grid-connected wind power.Under the above metioned 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 research work and research results are as follows:(1)The wind power forecasting method based on neural network with conjugate gradient algorithm is proposed.Starting from the structural properties of traditional forecasting models,the problem of slow convergence speed and large amount of iterations,especially challenges when faced with large scale data are analysed firstly.Then,a new kind of wind power forecasting method based on conjugate gradient algorithm is built,which uses conjugate gradient algorithm to optimize the model parameters.This proposed method can over come shortcomings of traditional feedback forward neural networks based on steepest training,and reduce the probability that the traditional models are easy to fall into local convergence.It can improve the forecasting accuracy,training speed and generalization performance effectively.(2)Aiming at the problem that the traditional predictive model ignores the learning correction during the prediction process,an adaptive learning framework based on reinforcement learning algorithm and a wavelet neural network combined model prediction method is proposed.First,using the wavelet neural network with characteristic of multi-scale to construct prediction module of the method,then optimizing the model parameters based on adaptive frame structure and model learning effect feedback mechanism,and making the model to adapt to the changeable and complex operation.At the end of prediction results are obtained after learning optimization module.This method provides a direction for stabilizing the uncertainty of wind power and achieving scheduling operation exactly.The effectiveness and universality of the proposed model and method are verified by the analysis of the case.Accurate wind power forecasting can provide a more effective and scientific basis for the multi-objective optimal dispatching.Based on the achievements above,this dissertation makes a further study in the field of optimization of environmental economic dispatch.The main research results are summarized as follows:(3)Considering the uncertainty model for wind power forecasting and the economy and environment of system operation,the environmental economic dispatch model of wind farm incorporated system is established.Using the nondominated sorting genetic algorithm with elitist strategy(NSGA-II)to solve multi-objective optimal solution sets and optimal trade-off solutions.The feasibility and effectiveness of the proposed model are verified by coal-fired generating sets with wind power integrated system.(4)The dynamic environmental economic dispatch model of wind farm incorporated power system is established.The dynamic model considers the ties of each time section,to simultaneously optimize the minimum costs and the minimum pollution emissions as goal.It explores the use of NSGA-? 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 optimal result.