Short-Term Electricity Load Forecasting Using K-Means Clustering - Artificial Neural Networks Hybrid Model: Case Study Of Benin Electricity Community (CEB)

• Published in: 2021 IV International Conference on High Technology for Sustainable Development (HiTech) pp. 01 - 05
• Main Authors: Guenoupkati, Agbassou, Salami, Adekunle Akim, Kodjo, Mawugno Koffi, Nano, Kossi
• Format: Conference Proceeding
• Language: English
• Published: IEEE 07.10.2021
• Subjects: Artificial neural networks; Electricity load; Forecasting; Hybrid model; Interconnected power grid; K-Means algorithm; Load forecasting; Multilayer Perceptron; Multilayer perceptrons; Predictive models; Short-term forecasting; Supply and demand; Temperature
• DOI: 10.1109/HiTech53072.2021.9614236

In this work, a hybrid model based on K-Means Clustering and Artificial Neural Networks is proposed to predict the Republics of Benin and Togo electricity consumption supplied by the CEB interconnected power grid. This approach consists in first classifying the load data according to different day profiles. Two profiles corresponding to working and non-working days are elaborated for this purpose. In a second step, each class of data is predicted using Multilayer Perceptron (MLP) Artificial Neural Networks (ANNs). The results obtained from this mixed approach are compared with those obtained from a generalized model of Multilayer Perceptron Artificial Neural Network. The results show that the proposed hybrid model gives best performance for CEB electricity load prediction.