Analysis of Deep Learning Sequence Models for Short Term Load Forecasting
MetadataShow full item record
CitationDağ, O., & Nefesoğlu, O. (2023, March). Analysis of Deep Learning Sequence Models for Short Term Load Forecasting. In Computational Intelligence, Data Analytics and Applications: Selected papers from the International Conference on Computing, Intelligence and Data Analytics (ICCIDA) (pp. 104-116). Cham: Springer International Publishing.
Short Term Load Forecasting (STLF) is an essential part of generator scheduling in power plants. Better scheduling is crucial for both economic and environmental aspects. In this study, two different deep learning (DL) model architectures based on Long Short Term Memory (LSTM) and Convolutional Neural Networks (CNN) were tested and compared on load datasets from Spain and Turkey. In addition, a novel Long Short Term Memory (LSTM) model with embedding layer was proposed and compared with these two models. Optimum model design choices and practices were discussed for achieving reliable and robust results. It was showed that datasets with different characteristics require different model design choices. The simulations of the proposed DL method were carried out with Python and the performance parameters were also presented. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.