On-line fuzzy-rule-based thermal control of a sub-scaled multi-room building facility

Abstract This study presents the development and experimental validation of robust fuzzy-rule-based strategies for real-time thermal control of a sub-scaled, multi-room building facility. Their main objective is to maintain prescribed temperature setpoints in each room despite varying internal and external disturbances. The experimental testbed consists of eight rooms equipped with airflow dampers and light bulbs as heat sources. Each room is modeled as a single-input single-output (SISO) system, where the manipulated variable is the airflow rate and the controlled output is the average room air temperature. For each strategy, the control system uses time-dependent data, including the temperature error, its derivative, and cumulative integral, to inform the fuzzy logic controllers with room air-temperature readings for their actuation. Triangular membership functions and experimentally derived if-then rules constitute the basis of the fuzzy sets used in developing the control laws, with the Mamdani inference method being employed to generate the corresponding output, and the centroid technique to perform the defuzzification process. Two configurations were tested: a standalone (SA) scheme, which uses only individual room data, and a parallel-structure (PS) scheme, that incorporates thermal feed-back from adjacent rooms. Experimental results show that all fuzzy controllers can successfully regulate room temperatures within ±1°C of the desired setpoints, even under disturbances and changing temperature requirements, with the hierarchical controller offering superior accuracy and response time, highlighting its potential for real-world multi-zone building energy applications.