The Short-term Load Forecasting Method of Transformer Based on Prophet-LightGBM
Short-term load forecasting of transformer plays an important role in optimal operation,fine management,energy saving and loss reduction of distribution network.In order to accurately forecast the transformer load in the short-term,we proposed a short-term load forecasting method of transformer based on Prophet-LightGBM.It used the i-dea of secondary forecasting and combined the advantages of Prophet and LightGBM to achieve more accurate forecas-ting.Firstly,we analyzed the characteristic of the transformer load and the specific changes of a transformer load in a week.Then,we used Prophet's algorithm for historical load fitting and the first prediction to extract the characteristic items of the transformer load with more regular changes.Finally,we used the above-mentioned characteristics and LightGBM algorithm to forecast the transformer load for the second time,so as to achieve the goal of predicting the over-all change of transformer load.The experimental results show that compared with the prediction results of Prophet and LightGBM,the proposed method has higher prediction accuracy in three types of transformers.
万方数据
维普
