A Learning Predictive Model for Improving the Performance of Indoor Thermal Controls and Energy Use for Temporary Buildings

Purpose: For economic reasons, several temporary buildings have been planned and built in urban areas. Therefore, in addition to the consideration for material properties and recycling issues, it is time to investigate the sustainable method for improving the thermal comfort of users and the performance of energy use. Method: By utilizing the characteristics of temporary buildings commonly used in industry, this study aims to develop a control model that can maintain the quality of indoor thermal comfort and suppress the increase in excessive energy use. The improved control patterns are calculated by combining a conventional thermostat model with an adaptive process that analyzes the changes in human factors. By learning the control patterns as input variables, a predictive model is developed and its control performance is analyzed. Results: As compared to the conventional thermostat model, the performance of the learning predictive model is improved by about 63% for the consistency of thermal comfort and by about 8% for the reduction of energy use, respectively. The result confirms the possibility that significant energy savings can be achieved while maintaining the usability of temporary buildings where the operational economy is prioritized.