States that predicting the thermal load for the next 24 hours is essential for optimal control of HVAC systems using thermal cool storage. Notes a joint U.S. Japanese research project to investigate four generally-used prediction methods to examine the basic models with variations and to compare the accuracy of each model. States the results indicate that an artificial neural network (ANN) produces the most accurate thermal load predictions. The ANN model was then applied to two measured building loads from another research project and the results confirmed its accuracy.

KEYWORDS: year 1995, Calculating, expert systems, controls, energy storage, cooling, research, comparing, testing, accuracy, cooling load