Climate responsive cooling control using artificial neural networks

The building envelope is influenced by climatic factors as thermal radiation, solar radiation, convection heat and infiltration heat. Their peak occurs at different times. Obtaining an equivalent thermal resistance of the building envelope is a challenge considering heat loss/heat gain of building envelope towards climate responsive cooling control. Considering heat flow at the zone using EnergyPlus software brings climate responsive cooling control. The Artificial Neural Network (ANN) model was developed which deciphers the building envelope heat flow using data obtained from EnergyPlus. Using ANN, model predictive controller and Gray box model of the building cooling system, thermal performance was obtained by simulations using Simulink, MLE + , BCVTB and EnergyPlus. The ANN envelope heat load predictor improves energy efficiency over the temperature based model in which the climate heat flow is determined using the equivalent thermal resistance and the atmospheric temperature. An Energy saving of 6.25% with 1.05% error for Chennai 5.19% with 2.21% error for Trichy and 7.52% with 0.08% error for Shillong climate was obtained.