A data fusion scheme for building automation systems of building central chilling plants

Accurate and reliable building load measurement is essential for robust chiller sequencing control, building air-conditioning system performance monitoring and optimization. This paper presents a scheme adopting the data fusion technique to improve the quality of building cooling load measurement of building automation systems. The strategy uses two types of measurement information on the cooling load, i.e., “direct measurement” of building cooling load, which is calculated directly using the differential water temperature and water flow rate measurements, and “indirect measurement” of building cooling load, which is calculated using a model using the instantaneous chiller electrical power input. Capitalizing their own advantages and disadvantages, a data fusion algorithm is developed to merge these two types of data to remove outliers and system errors as well as to reduce the impacts of measurement noises. Meanwhile, a method is implemented to provide quantitative evaluation of the degree of reliability of the merged measurement. Validation of the data fusion algorithm is conducted using field data collected from a chiller plant in a high-rising building in Hong Kong.