A Model-Driven Learning Approach for Predicting the Personalized Dynamic Thermal Comfort in Ordinary Office Environment

A vital requirement for all-air ventilation systems are their functionality to operate both in cooling and heating mode. This article experimentally investigates two newly designed air distribution systems, corner impinging jet (CIJV) and hybrid displacement ventilation (HDV) in comparison against a mixing type air distribution system. These three different systems are examined and compared to one another to evaluate their performance based on local thermal comfort and ventilation effectiveness when operating in heating mode. The evaluated test room is an office environment with two workstations. One of the office walls, which has three windows, faces a cold climate chamber. The results show that CIJV and HDV perform similar to a mixing ventilation in terms of ventilation effectiveness close to the workstations. As for local thermal comfort evaluation, the results show a small advantage for CIJV in the occupied zone. Comparing C2-CIJV to C2-CMV the average draught rate (DR) in the occupied zone is 0.3% for C2-CIJV and 5.3% for C2-CMV with the highest difference reaching as high as 10% at the height of 1.7 m. The results indicate that these systems can perform as well as mixing ventilation when used in offices that require moderate heating. The results also show that downdraught from the windows greatly impacts on the overall airflow and temperature pattern in the room.

Download Full-text

A machine learning approach to predict outdoor thermal comfort using local skin temperatures

Sustainable Cities and Society ◽

10.1016/j.scs.2020.102216 ◽

2020 ◽

Vol 59 ◽

pp. 102216 ◽

Cited By ~ 4

Author(s):

Kuixing Liu ◽

Ting Nie ◽

Wei Liu ◽

Yiqing Liu ◽

Dayi Lai

Keyword(s):

Machine Learning ◽

Thermal Comfort ◽

Outdoor Thermal Comfort ◽

Learning Approach ◽

Local Skin ◽

Machine Learning Approach ◽

Skin Temperatures

Download Full-text

Field experiments on energy consumption and thermal comfort in the office environment controlled by occupants’ requirements from PC terminal

Building and Environment ◽

10.1016/j.buildenv.2006.05.012 ◽

2007 ◽

Vol 42 (12) ◽

pp. 4022-4027 ◽

Cited By ~ 62

Author(s):

Yoshifumi Murakami ◽

Masaaki Terano ◽

Kana Mizutani ◽

Masayuki Harada ◽

Satoru Kuno

Keyword(s):

Energy Consumption ◽

Thermal Comfort ◽

Field Experiments ◽

Office Environment

Download Full-text

A Machine Learning approach to enhance indoor thermal comfort in a changing climate

Journal of Physics Conference Series ◽

10.1088/1742-6596/2042/1/012070 ◽

2021 ◽

Vol 2042 (1) ◽

pp. 012070

Author(s):

Tobias Kramer ◽

Veronica Garcia-Hansen ◽

Sara Omrani Vahid M. Nik ◽

Dong Chen

Keyword(s):

Thermal Comfort ◽

Data Science ◽

Prediction Models ◽

Computational Design ◽

Learning Approach ◽

Indoor Thermal Comfort ◽

Machine Learning Approach ◽

Early Case ◽

Building Characteristics

Abstract This paper presents an alternative workflow for thermal comfort prediction. By using the leverage of Data Science & AI in combination with the power of computational design, the proposed methodology exploits the extensive comfort data provided by the ASHRAE Global Thermal Comfort Database II to generate more customised comfort prediction models. These models consider additional, often significant input parameters like location and specific building characteristics. Results from an early case study indicate that such an approach has the potential for more accurate comfort predictions that eventually lead to more efficient and comfortable buildings.

Download Full-text

Teaching Model-Driven Engineering in a Master's Program

10.4018/978-1-6684-3702-5.ch075 ◽

2022 ◽

pp. 1586-1611

Author(s):

Alexandre Bragança ◽

Isabel Azevedo ◽

Nuno Bettencourt

Keyword(s):

Project Based Learning ◽

Learning Approach ◽

Master's Program ◽

Student Survey ◽

Model Driven Engineering ◽

Master Program ◽

Teaching Experiences ◽

Model Driven ◽

Software Complexity ◽

Software Professionals

Model-driven engineering (MDE) is an approach to software engineering that adopts models as the central artefact. Although the approach is promising in addressing major issues in software development, particularly in dealing with software complexity, and there are several success cases in the industry as well as growing interest in the research community, it seems that it has been hard to generalize its gains among software professionals. To address this issue, MDE must be taught at a higher-education level. This chapter presents a three-year experience in teaching MDE in a course of a master program in informatics engineering. The chapter provides details on how a project-based learning approach was adopted and evolved along three editions of the course. Results of a student survey are discussed and compared to those from another course. In addition, several other similar teaching experiences are analyzed.

Download Full-text