Deep Learning-Based Short-Term Load Forecasting Approach in Smart Grid With Clustering and Consumption Pattern Recognition

Different aggregation levels of the electric grid’s big data can be helpful to develop highly accurate deep learning models for Short-term Load Forecasting (STLF) in electrical networks. Whilst different are proposed STLF, they based on small historical datasets and not scalable process large amounts as energy consumption grow exponentially distribution This paper proposes a novel hybrid clustering-based approach STLF at transformers’ level with enhanced scalability. It investigates gain training time performance terms accuracy when modeling is employed STLF. A k-Medoid algorithm clustering whereas forecasting generated clusters load profiles. The transformers similarity profile. reduces since it minimizes number required many transformers. developed neural network consists six layers employs Adam optimization using TensorFlow framework. day-ahead hourly horizon forecasting. tested 1,000-transformer substation subset Spanish containing more than 24 million records. results reveal that model has superior compared state-of-the-art methodologies. delivers an improvement around 44% while maintaining single-core processing non-clustering models.