The use of self-organising maps to investigate heat demand profiles

District heating companies are responsible for delivering the heat produced in central heat plants to the consumers through a pipeline system. At the same time they are expected to keep the total heat production cost as low as possible. Therefore, there is a growing need to optimise heat production through better prediction of customers needs. The paper illustrates the way neural networks, namely self-organised maps can be used to investigate long-term demand profiles of consumers. Real-life historical sales data is used to establish a number of typical demand profiles.