Comparison of conventional and fuzzy control of indoor air quality in buildings

This article compares various techniques for the control of indoor air quality in a naturally ventilated building: ON-OFF, proportional-inregral-derivative (PID), proportional-integral with deadband (PIdb), and fuzzy control. The control system for the fuzzy controllers is a typical two input (carbon dioxide concentration and its time derivative), one output (change in window opening area) system. Two different fuzzy controllers are modeled: one based on Mamdani's max-min inference, and the other using Gupta's multivariable structure. The results show that the fuzzy controllers and the PIdb controller are much more suitable than the PID or the ON-OFF controllers. The number of cycles of the controlled parameter (i.e., the window opening fluctuations) is actually much smaller. Minor differences between the two fuzzy controllers are observed. (C) 1996 John Wiley and Sons, Inc.