Search

There is a growing need for accurate and interpretable machine learning models of thermal comfort in buildings. Physics-informed machine learning could address this need by adding physical consistency to such models. This paper presents metamodeling of thermal comfort in non-air-conditioned buildings using physics-informed machine learning. The studied metamodel incorporated knowledge of both quasi-steady-state heat transfer and dynamic simulation results. Adaptive thermal comfort in an office located in cold and hot European climates was studied with the number of overheating hours as index. A one-at-a-time method was used to gain knowledge from dynamic simulation with TRNSYS software. This knowledge was used to filter the training data and to choose probability distributions for metamodel forms alternative to polynomial. The response of the dynamic model was positively skewed; and thus, the symmetric logistic and hyperbolic secant distributions were inappropriate and outperformed by positively skewed distributions. Incorporating physical knowledge into the metamodel was much more effective than doubling the size of the training sample. The highly flexible Kumaraswamy distribution provided the best performance with R2 equal to 0.9994 for the cold climate and 0.9975 for the hot climate. Physics-informed machine learning could combine the strength of both physics and machine learning models, and could therefore support building design with flexible, accurate and interpretable metamodels. [ABSTRACT FROM AUTHOR]

Since indoor clothing insulation is a key element in thermal comfort models, the aim of the present study is proposing an approach for predicting it, which could assist the occupants of a building in terms of recommendations regarding their ensemble. For that, a systematic analysis of input variables is exposed, and 13 regression and 12 classification machine learning algorithms were developed and compared. The results are based on data from 3352 questionnaires and 21 input variables from a field study in mixed-mode office buildings in Spain. Outdoor temperature at 6 a.m., indoor air temperature, indoor relative humidity, comfort temperature and gender were the most relevant features for predicting clothing insulation. When comparing machine learning algorithms, decision tree-based algorithms with Boosting techniques achieved the best performance. The proposed model provides an efficient method for forecasting the clothing insulation level and its application would entail optimising thermal comfort and energy efficiency. [ABSTRACT FROM AUTHOR]

It has been found in recent years that using setpoint temperatures based on adaptive thermal comfort models is a successful method of energy conservation. Recent studies using adaptive setpoint temperatures incorporate international models from ASHRAE Standard 55 and EN16798-1. This study, however, has instead considered a regional Brazilian adaptive comfort model. This study investigates the energy demand arising from the use of a local Brazilian comfort model in order to assess the energy implications from the use of the worldwide ASHRAE Standard 55 adaptive model and various fixed setpoint temperatures. All of Brazil's climate zones, full air-conditioning, mixed-mode building operating modes, present-day climate change scenarios, and future scenarios—specifically Representative Concentration Pathways (RCP) 2.6, 4.5, and 8.5 for the years 2050 and 2100—have all been taken into account in building energy simulations. The use of adaptive setpoint temperatures based on the Brazilian local model considering mixed-mode has been found to significantly reduce energy consumption when compared to static setpoint temperatures (average energy-saving values ranging from 52% to 58%) and the ASHRAE 55 adaptive model (average values ranging from 15% to 21%). Considering climate change and the mixed-mode Brazilian model, the overall energy demand for the three groups of climatic zones (annual average outdoor temperatures ≤ 21 °C, > 21 and ≤ 25 °C and > 25 °C) ranged between 2% decrease and 5% increase, 4% and 27% increase, and 13% and 45% increase, respectively. It is concluded as a consequence that setting setpoint temperatures based on the Brazilian local adaptive comfort model is a very efficient energy-saving method. [ABSTRACT FROM AUTHOR]

Predicting the thermal sensations of building occupants is challenging, but useful for indoor environment conditioning. In this study, a data-driven thermal sensation prediction model was developed using three quality-controlled thermal comfort databases. Different machine-learning algorithms were compared in terms of prediction accuracy and rationality. The model was further improved by adding categorical inputs, and building submodels and general models for different contexts. A comprehensive data-driven thermal sensation prediction model was established. The results indicate that the multilayer perceptron (MLP) algorithm achieves higher prediction accuracy and more rational results than the other four algorithms in this specific case. Labeling AC and NV scenarios, climate zones, and cooling and heating seasons can improve model performance. Establishing submodels for specific scenarios can result in better thermal sensation vote (TSV) predictions than using general models with or without labels. With 11 submodels corresponding to 11 scenarios, and three general models without labels, the final TSV prediction model achieved higher prediction accuracy, with 64.7%–90.7% fewer prediction errors (reducing SSE by 3.2–4.9) than the predicted mean vote (PMV). Possible applications of the new model are discussed. The findings of this study can help in development of simple, accurate, and rational thermal sensation prediction tools. [ABSTRACT FROM AUTHOR]