ELM算法在微电网超短期负荷预测的应用

Application of ELM Algorithm in the Ultra-Short Term Load Prediction of Microgrid

针对微电网中负荷预测的特点提出的多模型极限学习法(ELM),将其应用于系统超短期负荷预测。该算法采取建立多个前馈神经网络模型,基于更新准则将模型集分成即时更新模型和批次更新模型这两个部分,以此进行数据筛选和分析。此算法的目的在于节省训练时间,提高预测的速度,同时保证预测的精度。

For the characteristics of microgrid load prediction, put forward a multi-model extreme learning method, which is applied to the system ultrashort term load prediction. The algorithcm adopts to set up mulitiple former feedfor- ward netrual network models. In view of renewing norms, divide the models into real renewing model and batch renewing model, taking this as screening and analysis of the data. The aim of the algorithem results in saving time, increasing pre- diction speed and ensuring predicting precision.