Method of medium-term industrial load forecasting of region

This paper presents a new medium-term forecasting method applied to load of a particular region. This is a hybrid method based on neural networks and technocenosis theory. The theory in its turn is based on idea to regard big infrastructure objects as an entity of technical products. The main advantage of describing method is that it allows to account system constraints and dependences, and structural range of power consumers. Necessary data for analysis was collected using corporate ERP system.