A Statistical Approach for Selecting Buildings for Experimental Measurement of HVAC Needs

Abstract

This article presents a statistical methodology for selecting representative buildings for experimentally evaluating the performance of HVAC systems, especially in terms of energy consumption. The proposed approach is based on the k-means method. The algorithm for this method is conceptually simple, allowing it to be easily implemented. The method can be applied to large quantities of data with unknown distributions. The method was tested using numerical experiments to determine the hourly, daily, and yearly heat values and the domestic hot water demands of residential buildings in Poland. Due to its simplicity, the proposed approach is very promising for use in engineering applications and is applicable to testing the performance of many HVAC systems.