Statistical-mathematical procedure to determine the cooling distribution of a chiller plant

This paper presents a procedure to determine the cooling capacity distribution of the chillers composing a chiller plant using a statistical analysis of the building cooling demand. The mathematical-statistical procedure uses tools such as frequency histograms, box-and-whisker plots, stem-and-leaf plots, the generalized least squares method, and finally an iterative factorial procedure to generate from the processed information. Besides the manufacturer’s data, all possible chiller plant combinations considering design constraints. The procedure was verified in a hotel facility. Eight thermal demand profiles were simulated. Statistical analysis yielded a range of individual capacities between 100–353 kW. The procedure generated 189 refrigeration plant combinations between 2 to 5 chillers, with a safety factor (SF) between 10%–20%. The highest number of combinations considered plants comprising three and four chillers, reaching 50 and 70 chiller plant options, respectively