The Short-Term Power Consumption Forecasting Based on the Portrait of Substation Areas

The short-term load forecasting of substation areas is significant in varieties of scenarios for the power grid. The daily power consumption (DPC) forecasting during high temperature period is especially significant for the soaring demand. In this paper, a portrait-based multivariate regression model (PMRM) is proposed with the idea to forecast the short-term DPC from prior knowledge of the portraits of substation areas. A label system of DPC is modeled and the portrait of each substation area is derived by clustering method based on the label system. Then the PMRM is performed for DPC forecasting in each cluster respectively. The case study applies PMRM and the benchmark model using the real DPC data of 161 substation areas in Shanghai and validates the effectiveness and priority of the PMRM.