A hybrid AI model for forecasting electricity volume to optimize water supply company efficiency

Abstract

Most water supply companies consume a large amount of electricity to ensure technological processes of water purification and distribution. However, even though Vodokanals are a large consumer of electricity, the forecasting of electricity consumption is still not given priority. An accurate forecast of the amount of electricity consumption will allow optimization of the distribution of consumption, reducing the values of peak consumption and in general reducing the electricity costs. In this study, deep learning methods are proposed to predict the daily electrical load during a month. Where the performance of deep learning artificial neural networks and hybrid neural networks are compared, this study, based on the comparison of various deep learning methods, proposes to increase the effectiveness of the application of artificial neural networks by their hybridization, to forecast the daily electrical load in the monthly period. We combined the gray wolf optimizer (GWO) and the group data processing method (GMDH) to predict the optimal amount of electrical load in water utilities. Keywords—extrapolation; forecasting; observation data; plotting position formulas; uncertainty.