Automated commissioning of air-handing units using fuzzy air temperature sensors

Abstract

This chapter describes a new approach to automating the commissioning of air-handling units based on the use of a fuzzy mixed air temperature sensor. The fuzzy sensor is developed using a fuzzy identification scheme and training data obtained by simulating the air temperature and flow around the sensing element(s) with a computational fluid dynamics package. The fuzzy sensor removes the bias errors and provides an indication of the uncertainty associated with the air temperature measurement at the current operating conditions. Sensor fusion is used to reduce the measurement uncertainty. Fuzzy reference models, which take account of sensor bias, are identified using the output from the fuzzy sensor together with training data generated by both CFD and conventional lumped-parameter simulations. The reference models, which provide a semiqualitative description of the plant behavior when it is fault-free and when particular faults are present, are incorporated into a fuzzy model-based fault diagnosis scheme. A version of the fuzzy sensor, which uses the return damper position as an auxiliary input, is also developed to reduce the levels of uncertainty, associated with estimating the average mixed air temperature.