Speaker
Description
Occupant-centric control has gained significant attention in recent years, with the primary focus on developing personal thermal preference models to ensure optimal comfort for individuals within a given space. The Predicted Mean Vote (PMV) serves as an evaluation index for thermal comfort; however, it does not fully capture the diverse thermal preferences of individual occupants. Consequently, there is a need for an adaptive control method that can cater to various occupants with different thermal preferences.
Data-based personal thermal preference models, which predominantly rely on machine learning techniques, have been the primary focus of many studies. These models, however, suffer from several limitations, such as the requirement of a sufficient amount of data and the need for complex data learning methods during the model creation process. This necessitates the development of alternative strategies that can overcome these challenges and provide a more inclusive and versatile thermal preference model.
In response to these limitations, this study proposes a novel method that eliminates the need for a large amount of data and complex calculations by utilizing a white box method. The white box method emphasizes the logical causal relationship between variables and thermal preferences, allowing for the creation of various thermal preference models without relying on extensive data sets.
Additionally, this study introduces a virtual person-based thermal preference model that simulates the thermal preferences of occupants until sufficient real data is gathered. The model adapts and evolves with each addition of new data, gradually becoming more representative of the actual participants' preferences. This approach offers a significant advantage over previous research, which did not consider occupant thermal preferences until enough data was collected.
As a result, this method allows for the consideration of occupant thermal preferences even when data is limited. As more data is collected, the model becomes increasingly tailored to the actual occupants, ensuring a more comfortable environment for all. This innovative approach addresses the shortcomings of data-based models and offers a more practical and efficient solution for occupant-centric control.
In conclusion, the development of personal thermal preference models that cater to individual differences is essential for occupant-centric control. Traditional data-based models, which rely heavily on machine learning, have inherent disadvantages due to their dependence on large data sets and complex calculations. By employing a white box method and a virtual person-based thermal preference model, this study overcomes these challenges and offers a more inclusive and adaptive solution. This method not only accounts for occupant thermal preferences with limited data but also evolves and refines itself as more data becomes available, ultimately resulting in a more accurate and representative model. Such an approach has the potential to revolutionize occupant-centric control and significantly improve the overall thermal comfort of occupants in various environments.
References
Zhang, H., & Tzempelikos, A. (2021). Thermal preference-based control studies: review and detailed classification. Science and Technology for the Built Environment, 27(8), 1031-1039.
| Keywords | Thermal Preference, Logistic Regression |
|---|
