Adaptive fuzzy temperature control for hydronic heating systems

This paper describes an innovative adaptive fuzzy control (AFC) algorithm for regulating the room temperature in a hydronic heating system. The hydronic heating system consists of a boiler that provides central heating water to a radiator or radiators in a home or apartment. Room temperature is regulated with a motor that drives a variable-position valve on each radiator that controls the flow of hot water through the radiator. The AFC automatically learns the steady-state radiator valve positions for five operating set points and uses this information in a fuzzy controller that commands the radiator valve position. Laboratory tests under typical operating conditions show that the AFC simultaneously improves the control quality while reducing the battery consumption when compared with a conventional proportional-integral (PI) feedback controller. For a 70-hour laboratory test with typical operating set points, the control quality increased 32% while the estimated battery consumption decreased 35%. The AFC's memory and processing requirements are suitable for embedded microprocessors, so it is a realistic replacement for a conventional PID algorithm.