Very Short-Term9 Short-Term and Mid-Term Load Forecasting for Residential Academic Institute: A Case Study

In recent years number of smart meters installed in the distribution systems around the world has increased manifold. The large-scale data collected by these smart meters record real-time energy consumption across different nodes of the system. This constitutes a very large set of information. Availability of this very large amount of smart meter data opens up new avenues for multiple operations like load forecasting, demand side management (DSM), error identification etc. For DSM, availability of forecasted load data is indispensable. This paper presents a load forecasting technique that works well for very short term i.e. hour ahead or 15 min ahead load forecasting along with day ahead, month ahead and season ahead case i.e. short term as well as mid-term load forecasting technique. The forecasting is shown for the smart meter data of a practical system available at NIT Patna campus. The uniqueness about NIT Patna campus system is that it is a combination of all possible types of loads like commercial, residential as well as small industry due to the presence of multiple laboratories in the institute. Applicability of the proposed method on NIT Patna campus data shows the method’s applicability to any type of distribution system.