基于支持向量机的风力发电超短期风速预测模型研究

发布时间：2018-05-07 07:13

本文选题：超短期风速预测 + 支持向量机　；参考：《华北电力大学》2014年硕士论文

【摘要】：能源问题已经成为一个全球性问题,风能作为一种清洁、可再生的新能源,其发电量不断的提高,但是风力发电本身具有波动性、不稳定性的特点并且在并网之后会给电力系统带来不良冲击,电力系统的安全性、平稳性和高效性受到了很大的影响。为了减小风电并网对电网的冲击,合理的调度风能资源,对风电场的风能资源进行准确预测是非常有必要的。对于中长期的风速预测研究,已经有很多学者参与进来并进行了广泛的研究,其预测效果还是不错的；但是对于短期和超短期的风速预测研究还没有达到很理想的效果,主要是由于风速的随机性和非平稳性的特点造成的。首先本文对风电事业的发展现状和背景做了简单介绍,对于现有的风速预测模型方法进行了详细说明。其次主要研究了基于支持向量回归机的风速预测模型,随后对支持向量机理论与支持向量回归机原理及其实现问题的细节进行了介绍。根据在线测得的风速数据进行单个周期的预测,进而扩展到了多个周期的预测,延长了风速预测的预测时间,实现了对某一点的风速值的预测。最后介绍了主成分分析理论和粒子群优化算法,并将其与支持向量机有机的结合起来,通过主成分析了解各个变量所占的比重,然后进行不同变量的输入模型的研究,实现了风速预测由点到面的升级,即实现了某一个面的平均风速值预测,提高了风速预测的精度。在论文的最后提出了一些不足和新方法的探索。
[Abstract]:The energy problem has become a global problem, wind energy as a clean, renewable new energy, its power generation is increasing, but wind power itself is volatile, The characteristics of instability will bring bad impact to the power system after the grid connection. The security, stability and efficiency of the power system are greatly affected. In order to reduce the impact of wind power grid, it is necessary to accurately predict the wind energy resource of wind farm. For the study of wind speed prediction in the medium and long term, many scholars have participated in the research and carried out extensive research, and its prediction effect is still good, but for the short-term and ultra-short-term wind speed forecasting research has not achieved a very good effect. It is mainly caused by the randomness and non-stationarity of wind speed. Firstly, this paper introduces the current situation and background of wind power industry, and gives a detailed description of the existing wind speed forecasting model. Secondly, the wind speed prediction model based on support vector regression machine is studied, and then the theory of support vector machine, the principle of support vector regression machine and its implementation are introduced in detail. Based on the wind speed data measured online, a single period is predicted, which is extended to the prediction of multiple periods. The prediction time of wind speed prediction is extended, and the wind speed value at a certain point is predicted. Finally, the principal component analysis (PCA) theory and particle swarm optimization (PSO) algorithm are introduced, which are combined with support vector machine (SVM) to find out the proportion of each variable, and then to study the input model of different variables. The wind speed prediction is upgraded from point to surface, that is, the average wind speed value of a certain surface is forecasted, and the precision of wind speed prediction is improved. At the end of the paper, some shortcomings and new methods are put forward.
【学位授予单位】：华北电力大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM614;TP181

【参考文献】