Study On Short-term Prediction And Low-carbon Dispatching Of Grid-connected Wind Power

With the increasing global energy crisis and environmental pollution problems, clean, non-pollution wind power is developing rapidly around the world. China, as the world’s energy guzzler, the growth of wind power has emerged by leaps and bounds in recent years. By the end of2010, cumulative installed wind power capacity has reached44734MW in China, ranking first in the world. Wind power for the sustainable development of China’s energy is no longer "alternative". Green wind power can optimize the power structure of power system and promote the energy conservation of power industry and low carbon development. However, intermittent and stochastic volatility of wind power would put forward a new challenge to the safe and stable operation of power system.Reduce the impact of the grid interconnection of wind farms and achieve optimal scheduling between wind power and conventional thermal power to promote low carbon electricity production are important issues of wind power. This paper focus on wind power short-term forecast method, low-carbon dispatching of grid-connected wind power and intelligent optimization algorithms of low-carbon dispatching according to wind power forecast and power system operation dispatch.1) In order to avoid the defect of the SVR model which selects the learning parameters depending on human experience in wind power prediction, adaptive precocious determination criterion, hybrid perturbation operation and dynamic expansion shrinkage coefficient are added to QPSO (quantum-behaved particle swarm optimization) algorithm, then proposed ADQPSO (adaptive disturbance QPSO) algorithm to optimize selection of the SVR learning parameters and achieve automatic adjustment of parameter combination.2) Study the short-term prediction of wind power based on the combined forecasting method. Combine the time series method, BP Neural Network method, RBF Neural Network method and Support Vector Machine method, and then determine the weight coefficient by the minimum square of the errors, thus the short-term combined forecasting model of wind power is constructed.Give a detailed introduction about the principle, the basic approach and modeling steps of the combined forecasting method. A Case study of a wind farm in China shows the accuracy and stability of the combined forecasting method which provides a new way to wind power forecast.3) Study optimized dispatching of wind power integrated system on the basis of short-term wind power prediction. Build wind power integrated system low-carbon dispatching model with optimal energy-environmental efficiency and minimized generation resource consumption synthesizing environmental benefit and power generation. Fuzzy optimization method is used to deal with the two contradictory objective functions of optimization dispatch model. The model is solved with tabu search optimization PSO algorithm.4) Considering the strong fluctuations of the wind power and the error of the wind power short-term prediction, the constraint condition of the low-carbon dispatching model of wind power integrated system is expressed in the form of probability. Low-carbon dispatching model based on the multi-objective chance-constrained programming method is constructed. IBBO (improved biogeography-based optimization) algorithm is presented to improve the solving precision of the model. IBBO improves the performance of the BBO (biogeography-based optimization) using migration of cosine model, variation mechanism based on Cauchy distribution, hybrid migration operator and similar body detection technology. Example analysis shows that IBBO algorithm has good precision and robustness, the Pareto solution is evenly divided in a broad range. With the increase of wind power penetration power coefficient, extreme solution of optimal energy-environmental efficiency increases at the beginning and then decreases, but extreme solution of minimum power resource consumption is monotone decreasing.