Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Where should the thermal image sensor of a smart A/C look?-Occupant thermal sensation model based on thermal imaging data
https://orcid.org/0000-0003-0043-4127
Abstract Applying thermal imaging sensor to air conditioner for monitoring human thermal sensation and achieving dynamic settings may satisfy occupants' thermal needs while saving energy. The existing studies are mostly based on single-view imaging to build the model and ignore the possible differences in body surface temperature on thermal sensation response by gender, etc., which may have many limitations. Subject experiments were conducted in an artificial climate chamber to obtain subjective questionnaires and thermal images of the exposed frontal face, lateral face, top of the head, forearm, and hand dorsum of 27 subjects in this study. By applying machine learning classification algorithms and global optimal regression algorithms, the temperature collection zones that can accurately reflect the thermal sensation of both genders in each view were analyzed, and a two-stage thermal sensation assessment model applicable to multiple views was developed. Of the various imaging views, the frontal view of the face is the best, followed by the lateral view of the face, the top view of the head, and the forearm/hand dorsum view. For male and female, the mean absolute errors of the thermal sensation assessment model established were 0.41–0.49 and 0.50–0.53 thermal sensation units. In addition, gender differences were found in the response of head surface temperatures to thermal sensation. The results obtained can provide a reference for the application of thermal image sensor to smart air conditioners.
Graphical abstract Display Omitted
Highlights For evaluating thermal sensation using thermal image sensor, front view is the best. Two-stage model is developed by combining machine learning classifier with regression. The model evaluates thermal sensation via thermal imaging for smart air conditioner. The model takes into account gender differences in surface temperature sensitivity. The mean absolute error of the model is less than 0.5 thermal sensation units.
Where should the thermal image sensor of a smart A/C look?-Occupant thermal sensation model based on thermal imaging data
https://orcid.org/0000-0003-0043-4127
Abstract Applying thermal imaging sensor to air conditioner for monitoring human thermal sensation and achieving dynamic settings may satisfy occupants' thermal needs while saving energy. The existing studies are mostly based on single-view imaging to build the model and ignore the possible differences in body surface temperature on thermal sensation response by gender, etc., which may have many limitations. Subject experiments were conducted in an artificial climate chamber to obtain subjective questionnaires and thermal images of the exposed frontal face, lateral face, top of the head, forearm, and hand dorsum of 27 subjects in this study. By applying machine learning classification algorithms and global optimal regression algorithms, the temperature collection zones that can accurately reflect the thermal sensation of both genders in each view were analyzed, and a two-stage thermal sensation assessment model applicable to multiple views was developed. Of the various imaging views, the frontal view of the face is the best, followed by the lateral view of the face, the top view of the head, and the forearm/hand dorsum view. For male and female, the mean absolute errors of the thermal sensation assessment model established were 0.41–0.49 and 0.50–0.53 thermal sensation units. In addition, gender differences were found in the response of head surface temperatures to thermal sensation. The results obtained can provide a reference for the application of thermal image sensor to smart air conditioners.
Graphical abstract Display Omitted
Highlights For evaluating thermal sensation using thermal image sensor, front view is the best. Two-stage model is developed by combining machine learning classifier with regression. The model evaluates thermal sensation via thermal imaging for smart air conditioner. The model takes into account gender differences in surface temperature sensitivity. The mean absolute error of the model is less than 0.5 thermal sensation units.
Where should the thermal image sensor of a smart A/C look?-Occupant thermal sensation model based on thermal imaging data
https://orcid.org/0000-0003-0043-4127
Abstract Applying thermal imaging sensor to air conditioner for monitoring human thermal sensation and achieving dynamic settings may satisfy occupants' thermal needs while saving energy. The existing studies are mostly based on single-view imaging to build the model and ignore the possible differences in body surface temperature on thermal sensation response by gender, etc., which may have many limitations. Subject experiments were conducted in an artificial climate chamber to obtain subjective questionnaires and thermal images of the exposed frontal face, lateral face, top of the head, forearm, and hand dorsum of 27 subjects in this study. By applying machine learning classification algorithms and global optimal regression algorithms, the temperature collection zones that can accurately reflect the thermal sensation of both genders in each view were analyzed, and a two-stage thermal sensation assessment model applicable to multiple views was developed. Of the various imaging views, the frontal view of the face is the best, followed by the lateral view of the face, the top view of the head, and the forearm/hand dorsum view. For male and female, the mean absolute errors of the thermal sensation assessment model established were 0.41–0.49 and 0.50–0.53 thermal sensation units. In addition, gender differences were found in the response of head surface temperatures to thermal sensation. The results obtained can provide a reference for the application of thermal image sensor to smart air conditioners.
Graphical abstract Display Omitted
Highlights For evaluating thermal sensation using thermal image sensor, front view is the best. Two-stage model is developed by combining machine learning classifier with regression. The model evaluates thermal sensation via thermal imaging for smart air conditioner. The model takes into account gender differences in surface temperature sensitivity. The mean absolute error of the model is less than 0.5 thermal sensation units.
Where should the thermal image sensor of a smart A/C look?-Occupant thermal sensation model based on thermal imaging data
Lyu, Junmeng (Autor:in) / Du, Heng (Autor:in) / Zhao, Zisheng (Autor:in) / Shi, Yongxiang (Autor:in) / Wang, Bo (Autor:in) / Lian, Zhiwei (Autor:in)
Building and Environment ; 239
08.05.2023
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch