Short-Term Dynamic Prediction Method Of Wind Farm Output Power And Its Applications

Wind power is an ideal pollution-free electric power and the strongest growing form of renewable energy. Wind power has entered a rapid progress stage and its application is more and more widely. But wind power has the disadvantages of intermittence and volatility, which will bring adverse effect on optimal scheduling of power system, reactive power optimization and challenge to the safety and stabilization of wind power. Large grid-connected wind farm will affect power quality, the transient angle stability, frequency stability and voltage stability, which restricts the scale of wind power development. Wind power prediction is an effective approach for the above problem. Especially, the short term wind power prediction has great significance for the safe and stable operation of power system and the development of power market. Short-term wind power prediction is more needed in China, but the research in this field is at the beginning, this paper makes research and discussion in the field. This paper studied short-term wind power prediction methods and its application in power system scheduling, the main work includes the following several aspects.①The wind characteristic parameters, wind power generation characteristic, wind power characteristic and wind power factor was studied in this paper, and the pretreatment methods of output power, which required data collection, was determined. It is making sure that the reliability of the data, so as to lay a good foundation for follow-on work.②Based on the traditional wind power static prediction model, the prediction becomes poor because of the weak correlation between training data and real-time measurement data as the time go on, this paper proposes a short-term wind power output prediction approach based on the self-adjusting strategy, different from conventional prediction models, the proposed approach incorporates discriminant error function to realize the function of model parameters self-adjusting, thus increasing the prediction accuracy and realizing the dynamic prediction of wind power output. The simulation results demonstrate the proposed dynamic prediction approach can provide more accurate wind power prediction. And the accuracy and effectiveness of the proposed approach will be more apparent, when the wind power volatility is larger or the prediction time is longer. In addition, the dynamic adjustment strategy proposed by this paper has good versatility; it does not involve the inner character of basic prediction model, and could be applied to any static prediction models based on the time series data.③There are important applications about the short-term prediction of wind power in the field of power system reactive power optimization and optimal operation. Aiming at large-scale wind power grid, this paper proposed a scheduling method of joint of wind power and thermal power, which is improved priority. Generation cost of thermal power unit is set to optimization objective, and using particle swarm optimization algorithm to update the output of generating unit, the simulation results demonstrate the effectiveness of the proposed method. Aiming at small-scale wind power grid, this paper proposes a strategy about complements between wind power and hydroelectric power in micro grid operation, the output of wind power is mainly considered, and the output of hydroelectric power as a compensation way, this paper also contains the dispatching function of power grid. The operation strategy proposed in this paper, which maximized the value of wind energy and water energy and improved the quality and reliability of power supply of the micro grid, is effective and feasible.The study of this paper has great significance on improving the operation efficiency of wind power, the market competitiveness of wind power, the safe and stable operation of power system.