基于加权正则极限学习机的短期风速预测

精确的风速预测是风电功率预测的基础,对保障风电场并网运行和维护电力系统的安全、稳定具有重要意义。针对风速时间序列强烈的波动性、随机性,难以预测的特点,建立了一种基于加权正则极限学习机(WRELM)的短期风速预测新方法。首先,采用与风速相关性大的历史风速、风向以及温度、气压、湿度等气象因素构成候选特征集;采用最大相关最小冗余(mRMR)准则选取与风速序列相关性最大的特征集作为预测输入,由此确定预测网络的训练集和测试集,建立WRELM;采用训练集数据训练网络参数,构建WRELM预测模型;最后,采用WRELM网络预测短期风速。通过风电场实测风速数据试验,验证了该方法的有效性,可用于短期风速预测实践。

Short-term Wind Speed Prediction Based on Weighted Regular Extreme Learning Machine

Accurate wind speed prediction is the basis of wind power prediction. It is of great significance to the connected grid operation of wind farm and the safe stability of power system. In view of the strong volatility and randomness of wind speed time series, a new method of short-term wind speed prediction was established based on the weighted regular extreme learning machine (WRELM). First, the wind speed and wind direction time series which have high correlation with the wind speed were taken into account. Besides the meteorological factors were also taken as the candidate sets, such as temperature, pressure, humidity and so on. Then the maximal relevance minimal redundancy (mRMR) principle was used to select the maximum serial correlation properties as the prediction inputs. Therefore, the trained set and test set of the prediction network were determined to establish WRELM. The network parameters were trained by the data of training set. And the WRELM prediction model was built. Finally, the WRELM network was adopted to predict short-term wind speed. The data from the wind farm was carried out to do the experiment and it proved the effectiveness of the new method.