Share

IIR document

A hybrid prediction model for hotel cooling load considering dynamic occupancy rate.

Author(s) : WEI X., ZHAO A., ZHOU X., SUN F., FENG Z.

Type of article: IJR article

Summary

Optimal control of the chiller system in hotel air conditioning requires accurate short-term cooling load forecasts. However, existing studies on hotel building cooling load prediction mainly rely on historical load-related data, neglecting the impact of dynamic occupancy rates, which leads to reduced prediction accuracy. This study aims to enhance the accuracy of short-term cooling load forecasting in hotel buildings. A grey relational analysis method is first employed to comprehensively select influencing factors, identifying dynamic occupancy as one of the key contributors to short-term cooling load variation. Building on this, a hybrid prediction model named ETO Transformer LSTM is proposed, which integrates the self-attention mechanism of the Transformer with the sequential modeling capability of LSTM. Dynamic occupancy is incorporated as a core coupling factor, and the Equilibrium Optimizer (ETO) algorithm is utilized for hyperparameter tuning.Validation using real-world operational data from a large hotel in northern China demonstrates that dynamic occupancy significantly improves forecasting accuracy. The inclusion of occupancy information reduces the Mean Absolute Error (MAE),
Mean Absolute Percentage Error (MAPE), and Root Mean Square Error (RMSE) by 20.56 kW, 7.51 %, and 18.44 kW, respectively. Compared with other forecasting models, the proposed hybrid model further reduces MAE, MAPE, and RMSE by 25.34 kW, 9.3 %, and 18.68 kW, respectively. Meanwhile, among various forecasting models, ETO-transformer-LSTM achieves shorter prediction times and demonstrates advantages in terms of computational complexity.

Available documents

Format PDF

Pages: 15 p.

Available

  • Public price

    20 €

  • Member price*

    Free

* Best rate depending on membership category (see the detailed benefits of individual and corporate memberships).

Details

  • Original title: A hybrid prediction model for hotel cooling load considering dynamic occupancy rate.
  • Record ID : 30034346
  • Languages: English
  • Subject: Technology
  • Source: International Journal of Refrigeration - Revue Internationale du Froid - vol. 179
  • Publication date: 2025/11
  • DOI: http://dx.doi.org/https://doi.org/10.1016/j.ijrefrig.2025.08.023

Links


See other articles in this issue (36)
See the source

Indexing