Heating and cooling load prediction using a neural network system

Many models have been proposed to identify and predict system behavior, but the modeling is generally difficult, especially in the case that the system is complex and has the characteristics of nonlinearity. An artificial neural network has the capability of learning the system behavior, so the authors applied it to heating and cooling load prediction. Kohonen´s feature map was chosen as a network model, and the extended learning vector quantization (LVQ) which realizes an associative memory was adopted as a learning algorithm. The predictive results were good, and the authors were able to confirm the feasibility of the model in the field of load prediction.