Machine Learning for Short-Term Load Forecasting in Smart Grids

A smart grid is the future vision of power systems that will be enabled by artificial intelligence (AI), big data, and Internet things (IoT), where digitalization at core energy sector transformation. However, grids require managers become more concerned about reliability security systems. Therefore, planners use various methods technologies to support sustainable expansion systems, such as electricity demand forecasting models, stochastic optimization, robust simulation. Electricity plays a vital role in supporting reliable transitioning This paper deals with short-term load (STLF), which has an active area research over last few years, handful studies. STLF predicting one hour 24 h advance. We extensively experimented several methodologies from machine learning complex case study Panama. Deep advanced paradigm field continues have significant breakthroughs domain areas forecasting, object detection, speech recognition, etc. identified main predictors short term: previous week’s load, day’s temperature. found deep regression model achieved best performance, yielded R squared (R2) 0.93 mean absolute percentage error (MAPE) 2.9%, while AdaBoost obtained worst performance R2 0.75 MAPE 5.70%.