Development of control algorithms for optimal thermal environment of double skin envelope buildings in summer

This study examines diverse thermal control algorithms to control the openings and cooling systems of double skin envelope (DSE) buildings in summer. In order to examine the system performance of control algorithms, five control algorithms combining a conventional rule-based algorithm and four proposed algorithms including fuzzy logic (FL), artificial neural network (ANN), and two adaptive neuro-fuzzy inference systems (ANFIS) were developed. The system performance of the algorithms was compared to those from field measurements to validate prediction accuracy. Further simulations were performed for the DSE building by using the five validated control algorithms. The results indicate that the algorithm employing FL to operate cooling systems created the most acceptable and stable indoor temperature in which 99.98% of the test period was within the target indoor temperature with the narrowest ranges. Compared to other algorithms, the FL-based control algorithm for cooling system can potentially improve building energy efficiency demonstrating an amount of reduction in heat removal up to 49.4%. The ANN-based and ANFIS-based algorithms operated the cooling system more stably and effectively reduced the number of times that the cooling system was turned on and off.