论文部分内容阅读
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。
Precise wind speed prediction is the basis of wind power prediction, which is of great significance to ensure the grid operation of wind farms and to maintain the safety and stability of the power system. Aiming at the strong volatility, randomness and unpredictability of time series of wind speed, a new short-term wind speed prediction method based on weighted regular limit learning machine (WRELM) is proposed. Firstly, the historical wind speed, wind direction and temperature, pressure, humidity and other meteorological factors were used to construct the candidate feature set. The mRMR criteria was used to select the feature set with the highest correlation with the wind speed sequence as the predicted input , So as to determine the training set and test set of the forecasting network and establish WRELM. Training data of the training set is used to train the network parameters to construct the WRELM forecasting model. Finally, the WRELM network is used to forecast the short-term wind speed. The test of wind speed data of wind farm verifies the effectiveness of this method and can be used in short-term wind speed prediction.