采用变系数模型改进空间相关性风速预测

摘要

Spatial correlation method has been paid more and more attention for wind power prediction in recent years. In order to improve ultra-short-term wind power prediction effect, this paper presents a kind of method for improving spatial correlation wind speed prediction by using the variable coefficient model. It firstly introduces a modeling approach for spatial correlation wind speed prediction, and then according to change of wind speed sequence relationship with time, it introduces the variable coefficient model. By means of weighted least squares method, it makes parameter estimation and proposes a kind of weight function form. Based on simulating experiment, it studies influence of forgetting factor in the weight function on prediction error and verifies effectiveness and feasibility of this variable coefficient model in improving spatial correlation wind speed prediction.%空间相关性方法是近年来逐渐受到重视的风电功率预测方法.为提高超短期风电功率预测效果,提出了采用变系数模型改进空间相关性风速预测的方法.首先介绍了一种空间相关性风速预测的建模方式,然后根据风速序列间关系随时间变化的情况引入变系数模型,通过加权最小二乘法进行参数估计,给出一种权函数形式,最后的仿真试验研究了权函数中的遗忘因子对预测误差的影响,验证变系数模型提高空间相关性风速预测的有效性,说明采用变系数模型来改进空间相关性风速预测是一条可行的思路.