Abstract This study addresses the issue of intelligent energy forecasting in the hotel industry, with the primary objective of delivering accurate predictions of hotel energy demand to enhance the efficiency of energy resource planning and management systems. The anticipated benefits include a reduction in energy consumption, a decrease in associated greenhouse gas emissions, and an improvement in the sustainability metrics of hotel operations. The forecasting algorithm is based on the application of advanced intelligent methods, specifically designed to provide 24-hour energy consumption predictions with a 24-hour forecast horizon. The algorithm utilizes Long Short-Term Memory and Gated Recurrent Unit networks to achieve this goal. By incorporating key variables that influence energy consumption, the algorithm demonstrates significant capabilities that offer improvements over existing approaches. Notably, the proposed model is capable of online adaptation to changing conditions while maintaining an optimal balance between accuracy and simplicity. The evaluation of this method was conducted in a luxury hotel located in the Canary Islands, and the results indicate its potential for real-time implementation in an energy management system. Results also illustrate the impact of the forecasting model to improve the load displacement strategy when considering flexibility capacity in the hotel.
Arnay et al. (Thu,) studied this question.