Forecasting Energy Consumption of a Mining Plant Using Artificial Neural Networks

This study addresses the issue of forecasting active and reactive power consumption at a mining
and processing plant, aiming to improve the efficiency of energy resource management. It explores existing
approaches to modelling and analysing electricity consumption, including methods for forecasting active
power to accurately assess the energy needs of industrial enterprises, as well as methods for estimating
reactive power required to compensate for reactive components and stabilize grid parameters. Using real
electricity consumption data from an industrial enterprise, the completed training and research demonstrates
the potential to forecast energy consumption for the next 24 to 48 hours using artificial neural networks with
nonlinear autoregressive architecture. It also provides rationale for pre-processing initial data to enhance
forecast accuracy. These approaches contribute to reducing capital and operating costs, improving the
reliability and stability of energy systems, and optimising operating efficiency.