基于混沌相空间重构理论的风电场短期风速预测.pdf

第 38 卷 第 21 期 电力系统保护与控制 Vol.38 No.21 2010 年 11 月 1 日 Power System Protection and Control Nov. 1, 2010 基于混沌相空间重构理论的风电场短期风速预测 吕 涛，唐 巍，所 丽 （中国农业大学信息与电气工程学院，北京 100083） 摘要：风力发电具有波动性、间歇性和随机性的特点，大容量的风力发电接入，会对电力系统的安全、稳定运行带来严峻挑 战，进行短期风速预测对并网风力发电系统的运行具有重要意义。根据风速具有混沌特性，采用相空间重构理论对短期风速 进行预测。由嵌入时间窗Γ和 m 、τ 的关系，确定了m 和τ 的多组可行匹配，并找出一个最佳匹配进行相空间重构。在 选取参考点时既考虑相点欧式距离又考虑其空间及时间上的相关性，有效克服“伪邻近点”的影响，提高了预测精度。预测 模型采用了一阶局域预测模型和BP神经网络两种模型。算例分析结果验证了所提出方法的可行性和有效性。 关键词：风速时间序列；短期风速预测；混沌特性；相空间重构；神经网络 Prediction of short-term wind speed in wind farm based on chaotic phase space reconstruction theory Lü Tao，TANG Wei，SUO Li （College of Information and Electrical Engineering，China Agricultural University，Beijing 100083 ，China ） Abstract ：Wind power generation has the features of volatility ，intermittent and random ．Involved in high-capacity wind power will bring serious challenges to power system safe and stable operation ．Thus，short-term prediction of wind speed has an important significance to wind power system operation．According to property of wind speed ，the phase space reconstruction technology is used for short-term prediction of wind speed ．Based on the fundamental relation of the delay time window Γ and m 、τ ，several sets of optimal combinations of m andτ are advanced，and the optimal combination is found ．An effective method is presented to choose reference point ，which determines near phase point by euclidean distance and correlation degree among phase point ．Some false neighboring points are kicked off to improve forecasting accuracy ．On the selection of forecasting model，we use one-order local forecasting method and BP neural network model ．Case study results validate the feasibility and effectiveness of the prop