Decoupling the complexities of air-conditioning cooling energy use in express hotels by data mining approaches

Abstract

The air-conditioning (AC) energy use in express hotels is stochastic with the high coupling relationships amongst AC usage, indoor temperature and energy consumption. Such complexities and stochasticity make it hard to facilitate energy saving with clear effect on indoor environment. However, lacking analyses of high-resolution occupants’ energy use makes it difficult to achieve such goals due to the split form of ACs and various thermal comfort of guests in express hotels. Therefore, this study made a serial analysis on the AC energy use in a more detailed scope. The stochastic AC usage, indoor temperature and AC energy consumption were quantified by proposed typical patterns with the cluster method. The stochasticity was described by four typical patterns for each aspect. After the quantifications, the relationships amongst these three aspects were decoupled by the proposed energy use decoupling model. Two data mining methods, namely, random forest method and decision tree method, were employed to achieve this purpose, respectively. With these models, the impacts of each variable on AC energy consumption and explicit relationships of operation rules for management are presented. Strictly limiting set point temperature higher than 23°C is the effective way to save energy for most of AC usage patterns. This study can provide a deeper understanding of AC energy use in express hotels, and benefits energy saving and facility operation in express hotels.