Annual Variation Characteristics of Air Conditioning Operating Behavior and Its Impact on Model Application in Office Buildings

Abstract

Air conditioning (AC) is an important component of building energy consumption. Reducing building AC energy consumption has attracted significant research interest worldwide. Studies have shown that the AC control behavior of users is a key factor affecting building AC energy consumption; however, the existing research on the dynamic laws for the AC control behavioral changes of users over a long period is limited. Therefore, taking a typical open office as an example, this study collected measured data spanning different years, and explored the temporal variation characteristics of AC operating behavior in office buildings. Based on a dynamic model framework constructed with a three-parameter Weibull function and a time superposition function, this study conducted modeling and analysis of dynamic AC operating behaviors in the same open-plan office across different years. First, the AC operating behavioral model was trained in parallel using field measurement data from different years to quantitatively analyze the patterns and extent of changes in occupants’ AC operating behaviors. Subsequently, AC operating data from a fixed year was used as a test set to examine the impact of behavior changes on the prediction accuracy of the AC operating behavioral model through indicators such as open rate, on–off profiles, confusion matrices, and open rate under different time periods/temperatures. Results indicate that, due to behavioral changes, the maximum difference in the probability of AC opening under the same temperature can reach 96.8%. These behavior changes occur not only in varying intensity but also function as influencing factors. If behavior changes are ignored, prediction accuracy for AC open rates decreases by approximately 15%. This study reveals a method for dynamically adjusting the AC operating behavior model and improving its accuracy, which can significantly improve the accuracy of AC operating behavior modeling, the practical application effect of the behavior model, and reduce the energy consumption and carbon emissions of buildings.