Search

Due to the severe outdoor PM2.5 pollution in China, many people have installed air‐cleaning systems in homes. To make the systems run automatically and intelligently, we developed a recurrent neural network (RNN) that uses historical data to predict the future indoor PM2.5 concentration. The RNN architecture includes an autoencoder and a recurrent part. We used data measured in an apartment over the course of an entire year to train and test the RNN. The data include indoor/outdoor PM2.5 concentration, environmental parameters and time of day. By comparing three different input strategies, we found that a strategy employing historical PM2.5 and time of day as inputs performed best. With this strategy, the model can be applied to predict the relatively stable trend of indoor PM2.5 concentration in advance. When the input length is 2 h and the prediction horizon is 30 min, the median prediction error is 8.3 µg/m3 for the whole test set. For times with indoor PM2.5 concentrations between (20,50] µg/m3 and (50,100] µg/m3, the median prediction error is 8.3 and 9.2 µg/m3, respectively. The low prediction error between the ground‐truth and predicted values shows that the RNN can predict indoor PM2.5 concentrations with satisfactory performance. [ABSTRACT FROM AUTHOR]

[Display omitted] • Reinforcement learning saves energy by 52% for mixed-mode ventilation buildings. • Explainable AI can prompt new insights in mixed-mode ventilation building control. • Optimizing window opening area and cooling setpoint can increase energy saving. • Reinforcement learning utilizes internal thermal mass to increase cooling efficiency. The application of mixed-mode ventilation (MMV) in the tropics is challenging, given its hot and humid climate. Consequently, there are limited periods when operating in natural ventilation (NV) is desirable. Furthermore, the potential to use NV diminishes at locations characterized by generally light winds. Given the complex interactions in MMV, this study aims to use reinforcement learning (RL) to identify relationships that can further reduce cooling energy use while maintaining occupant thermal comfort. The control variable variables considered include the percentage of window opening area and dynamic cooling setpoint temperature. Results show that RL can achieve a 52% reduction in cooling energy while maintaining good thermal comfort and indoor air quality compared to the existing baselines that typically involve switching between NV and air-conditioning based on outdoor conditions. We then developed an Explainable AI (XAI) framework to prompt building scientists toward new insights into MMV control in the tropics, which consists of shapley additive explanations (SHAP) and decision tree. The SHAP is able to improve the transparency of RL strategy by revealing the impact of each input on the final decision and the decision tree can extract the key control rules from RL. Using the XAI framework, this study also identified that the RL algorithm was taking advantage of the internal thermal mass to increase cooling efficiency. The extracted rules are further applied to the actual testbed, showing the feasibility of the rules extracted by XAI approach. [ABSTRACT FROM AUTHOR]

Energy consumption has dramatically increased in buildings over the past decade due to population growth, more time spent indoors, increased demand for building functions and indoor environmental quality, and global climate change. Building energy use currently accounts for over 40% of total primary energy consumption in the U.S. and E.U. Nevertheless, significant energy savings can be achieved in buildings if they are properly designed, constructed and operated. For this reason, building energy efficiency can provide key solutions to energy shortages, carbon emissions and their serious threat to our living environment. This paper offers a brief overview of building energy-consumption situations, relevant energy-saving approaches, and the influence of global climate change. Building energy-consumption situations based on data derived from international energy reports are initially compared between the U.S., China and the E.U. Both similarities and differences are found in aspects of building energy end-uses and final energy fuel-types among these top three building energy consumers. We then introduce the current concept of the zero-energy building (ZEB). State-of-the-art approaches for ZEB technologies are summarized in three categories: passive energy-saving technologies, energy-efficient building service systems and renewable energy production technologies. The feasibility of these technologies is reviewed. In addition, we briefly discuss the influence of global climate change on the evolution of building energy use in the future. We find that climate change significantly impacts building energy performance, particularly in space heating and cooling. Improvements on building envelope and ventilation can play an important role in reducing space heating and cooling consumption levels. We also provide some suggestions for further developing ZEBs. [ABSTRACT FROM AUTHOR]

• The prediction of occupancy and window-opening behaviour using ML models was reviewed. • Temperature is the most important predictor variable for window-opening behaviour. • Logistic regression does not perform well for the "cut-off" temperature. • Integrating occupancy into HVAC system control can reduce energy consumption by 23%. This study carries out a literature review on studies using machine learning (ML) models to predict occupancy and window-opening behaviour and their application in smart buildings. For occupant number/level prediction, the indoor CO 2 concentration is an important predictor variable. To further improve the model performance, it is recommended to use feature engineering to create additional CO 2 -related predictor variables. For the window-opening behaviour, the outdoor temperature, indoor temperature and wind speed are the top three most important predictor variables. The window-opening probability usually increases as the outdoor temperature increases. It is not recommended to use both indoor and outdoor temperatures as predictor variables when buildings are free-running. In regard to model selection, although logistic regression is the most common model for the window-opening behaviour, it does not perform well for the "cut-off" temperature. From another aspect, it can be inferred that although an artificial neural network may have better prediction accuracy, its transferability might not be as good as that of traditional ML models. It is recommended to compare the performances of different ML models using a few criteria to find the most suitable model. In the future, to further improve the prediction accuracy, the model can take building characteristics and occupant features into consideration. In regard to the application of ML models, the energy consumption of heating, ventilation and air-conditioning (HVAC) systems can be reduced by 23% on average by optimizing the control strategy of HVAC systems according to the occupancy information predicted by the ML models. [ABSTRACT FROM AUTHOR]

To evaluate the separate impacts on human health and establish effective control strategies, it is crucial to estimate the contribution of outdoor infiltration and indoor emission to indoor PM2.5 in buildings. This study used an algorithm to automatically estimate the long‐term time‐resolved indoor PM2.5 of outdoor and indoor origin in real apartments with natural ventilation. The inputs for the algorithm were only the time‐resolved indoor/outdoor PM2.5 concentrations and occupants' window actions, which were easily obtained from the low‐cost sensors. This study first applied the algorithm in an apartment in Tianjin, China. The indoor/outdoor contribution to the gross indoor exposure and time‐resolved infiltration factor were automatically estimated using the algorithm. The influence of outdoor PM2.5 data source and algorithm parameters on the estimated results was analyzed. The algorithm was then applied in four other apartments located in Chongqing, Shenyang, Xi'an, and Urumqi to further demonstrate its feasibility. The results provided indirect evidence, such as the plausible explanations for seasonal and spatial variation, to partially support the success of the algorithm used in real apartments. Through the analysis, this study also identified several further development directions to facilitate the practical applications of the algorithm, such as robust long‐term outdoor PM2.5 monitoring using low‐cost light‐scattering sensors. [ABSTRACT FROM AUTHOR]

The indoor environment influences occupants' health. From March 1, 2018, to February 28, 2019, we continuously monitored indoor temperature (T), relative humidity (RH), and CO2 concentration in bedrooms via an online system in 165 residences that covered all five climate zones of China. Meanwhile, we asked one specific occupant in each home to complete questionnaires about perceived air quality and sick building syndrome (SBS) symptoms at the end of each month. Higher CO2 concentration was significantly associated with a higher percentage of perceived stuffy odor and skin SBS symptoms. Higher relative humidity was associated with higher percentage of perceived moldy odor and humid air, while lower RH was associated with a higher percentage of perceived dry air. Occupants who lived in residences with high RH were less likely to have mucosal and skin SBS symptoms (adjusted odds ratio (AOR): 0.73–0.78). However, the benefit of high humidity for perceived dry air and skin dryness symptoms is weaker if there is a high CO2 concentration level. [ABSTRACT FROM AUTHOR]

Ammonia is a harmful gas in semiconductor cleanrooms. Activated carbons are widely used to remove ammonia. However, the existence of ambient moisture significantly impacts ammonia adsorption on adsorbents. Therefore, it is necessary to understand their adsorption performance on activated carbons to improve the product yield in semiconductor cleanrooms. Although many ammonia and water adsorption performance studies have been conducted, there has been little experimental research on low-pressure ammonia and water adsorption performance. In this study, several traditional isotherm models were applied to fit experimental isotherm data. The double-site (DS) Langmuir isotherm model provided good performance for ammonia adsorption on M-1, M-3, and M-4. The Toth isotherm model provided the best fit for ammonia adsorption on M-2. The Dubinin–Serpinsky (DS) isotherm model provided the best fit for water adsorption on the four investigated materials even at low pressure. In regard to predicting adsorption capacity, the previous prediction approach may limit the accuracy of the models by using a single set of characteristics to predict the adsorption capacity at different pressures. To solve this, an attention-based recurrent neural network (RNN) is proposed in this study. The attention mechanism is able to assign different weights to each material characteristic for different partial pressures. Due to this strength, the attention-based RNN realizes the prediction of adsorption isotherms by only taking material characteristics as inputs. The mean absolute percentage error (MAPE) of the attention-based RNN model in the prediction of adsorption capacity was 4.09%–18.68% and 3.68%–20.58% for ammonia and water, respectively, indicating that the well-trained model provides a reasonable prediction. The results predicted by the attention-based RNN model are consistent with the experimental data in terms of adsorption isotherm type, further confirming the reliability of the RNN model. [Display omitted] • Toth isotherm model fits best for ammonia adsorption on M-2 (RMSE= 0.034). • DS Langmuir isotherm model fits best for ammonia adsorption on M-1, M-3, and M-4. • DS isotherm model fits best for water adsorption on all adsorbents. • The attention-based RNN provides a reliable prediction for adsorption isotherms. [ABSTRACT FROM AUTHOR]

Aortic fenestration (AF) uses puncture and a dilation balloon to create a tear in the intimal flap, which can directly relieve ischemia syndrome and reduce hypertension in the false lumen. The selection of a dilation balloon as well as the area of the created tear applied in reality depend on clinical experience, so we aim to provide a quantitative guidance and reference for doctors to better plan the treatment of aortic fenestration. In this study, the area of the created tear was virtually enlarged to at least 10 different values for four cases including one ideal case, and a computational fluid dynamic approach was applied to simulate blood flows in the aorta. The area ratio (AR) between the created tear and entry tear was introduced to express the enlargement of the created tear. The quantitative hemodynamic results indicate that the AR should be controlled to be larger than 7.0, but not too big to obtain the best treatment for acute aortic dissection (AD) case. Additionally, we assessed that AR might also be a risk factor for the prediction of dissection propagation. [ABSTRACT FROM AUTHOR]

The ventilation modalities in most Chinese residences are infiltration and opening windows. We measured infiltration rates and air change rates at night, with no attempt to change occupants' behaviors, of urban residences in five climate zones of China during four seasons. Using the CO2 decay method, we found the median infiltration rate for 294 residences to be 0.34 h−1. Using occupant‐generated CO2 as tracer gas, we determined air change rates over the course of 1 year in 46 bedrooms at night from mass balance considerations. In 54% of the measurements, windows were closed, so ventilation was only by infiltration. Windows were mainly closed when the outdoor temperature was below 15°C and above 26°C. The median infiltration rates did not differ appreciably among seasons and climate zones and were always less than 0.45 h−1. [ABSTRACT FROM AUTHOR]

Highlights • 2D-PIV experiments testing different exhaust airflow rates and mounting heights of range hood were conducted. • The range hood-driven flows under isothermal and heating conditions were compared. • The spill characteristics of the cooking plume were analyzed. • Instantaneous air distributions and vortex identification were analyzed. • The uncertainty of the data was verified as a basis for numerical simulation. Abstract High-performance range hoods can quickly exhaust cooking smoke and fumes outdoors. Quantifying the detailed information that range hood-driven flow provides can be an essential prerequisite to optimize the hood's performance. Direct measuring instruments cannot easily determine the velocity due to large velocity gradients and high temperatures. Particle image velocimetry (PIV) is a non-intrusive velocimeter-based method with high spatial resolution and in a global domain. Thus, we applied large-scale 2D-PIV and measured the range hood-driven flow in three sections under isothermal and heating conditions. This study systematically investigated the influence of the exhaust flow rate and mounting height on the range hood-driven flow. The measurement results under isothermal conditions show the actual airflow driven by a range hood cannot be characterized as a simple exhaust and have self-similarity for different exhaust airflow rates. The results also indicate that the mounting height only slightly impacts the velocity boundary conditions of the range hood exhaust surface. The results under the heating conditions clearly reveal that the time-averaged spill length is larger than the average value of the dynamic spill length. This phenomenon indicates that the relationship between time-averaged velocity fields and spill length is not a simple linear scaling relationship. The unsteady flow in the kitchen should be considered to determine the true capture efficiency. Moreover, uncertainty analysis ensures that the benchmark flow data for the follow-up kitchen environment and range hood design simulations. [ABSTRACT FROM AUTHOR]

Renewable energy usage is continuing to increase as many countries worldwide are aiming to reach peak carbon emission and achieve carbon neutrality in the near future. One inherent problem with renewable energy is that its generation profile does not often fit well with the electricity usage profile. Therefore, it is of utmost importance that terminal users help to adjust the usage profile. Thermal energy storage (TES) systems have become an important means of adjusting the electricity usage profile of buildings. The operation strategy for TES must be carefully optimized to maximize its economic profile. To this end, we developed a framework for TES operation strategy optimization by integrating deep learning and physics-based modeling. The deep learning model, an attention-based dual-gated recurrent unit (A-dGRU) network, can learn the cooling load change trends from historical data and achieve state-of-the-art performance in hourly cooling load prediction for the next day with a coefficient of variation of the root mean square error of 0.08. For the TES modeling, we took the nonlinear change in the ice-charging rate into consideration based on the heat-transfer model; this change has often been ignored in previous studies. The high prediction accuracy and reliability of the TES model guarantee that the optimal strategy can be achieved by the framework. Compared to the basic TES operation strategy, we confirmed that the optimal operation strategy can further increase the cost savings by 11.2% for the entire ice-cooling season. In summary, the framework proposed in this study performs well in reducing the operation cost of a cooling plant based on the current electricity price tariff. The framework is expected to help the grid fit the electricity generation and usage profile. • A framework for optimization of the TES operation strategy was developed. • The TES optimization framework integrates deep learning and physics-based modeling. • The cost saving was increased by 11.2% under the optimal strategy. • An open-source physics-based ice-storage cooling system model was created. [ABSTRACT FROM AUTHOR]

This study conducted on-site measurements of indoor volatile organic compounds (VOCs) in 251 occupied residences in China, with multiple visits throughout a whole year. Over 1000 samples were collected for measurement of VOCs in 8 cities, covering different climate regions. Overall, the concentrations of total VOCs (TVOCs) in occupied residences are in the range of 104–1151 μg/m3, with 20% of the samples over the Chinese standard of 600 μg/m3. A higher concentration was evident in the summer (mean = 705 μg/m3) compared to other seasons, especially winter (mean = 289 μg/m3). The TVOCs of residences in areas with central heating (severe cold regions and cold regions) are generally higher than those in areas without central heating. In winter, temperature was the dominant factor, whereas in summer, the building infiltration rate was the key factor influencing the indoor TVOC levels. The TVOCs concentration was also found to be directly proportional to the city economy level. Twenty-nine VOC species with a detection frequency higher than 40% were identified in all samples. Toluene is the most common VOC, with the highest detection rate (90%). The median concentration for a single VOC was between 1 and 14 μg/m3. Aldehydes were found to be the largest contributors to total VOCs in the Chinese residential buildings (mass proportion 22%), followed by benzene series (20%), alkenes (18%), and alkanes (15%). Aldehydes, especially long-chain saturated carbonyls, are likely to be the characteristic VOCs in the Chinese dwellings, with Chinese cooking as the major emission source. In addition, n-butane/i-butane showed maximum concentration in some residences (approximately 105 μg/m3 higher than other VOCs) owing to cooking fuel. Unlabelled Image • Multi-entry indoor VOCs survey in 251 Chinese residences • Significant higher VOC concentration in summer than in winter was observed. • Higher VOC concentration in northern China due to central heating and ventilation behavior • Aldehydes due to Chinese cooking are likely to be the characteristic VOCs of Chinese dwellings. [ABSTRACT FROM AUTHOR]