Pub Date : 2026-01-24DOI: 10.1016/j.enbuild.2026.117060
Harry Vallianos, Costa Kapsis
This study investigates the use of surrogate modeling to optimize retrofit strategies in multi-unit residential buildings (MURBs), including HVAC systems. A comprehensive synthetic dataset was generated using EnergyPlus simulations, parameterized across a wide range of building and system variables, including ten distinct HVAC configurations. Multiple surrogate modeling approaches were evaluated, including single-output and multi-output artificial neural networks (ANNs) as well as Light Gradient Boosting Machine (LGBM) models. The models were trained to predict key performance metrics: Energy Use Intensity (EUI), Thermal Energy Demand Intensity (TEDI), and Cooling Energy Demand Intensity (CEDI). Results show that multi-output ANN models, with HVAC system as a categorical input, achieved high accuracy (R2 > 0.997 and RMSE<2.2kWh/m2/year) and superior generalization compared to both single-output ANNs and LGBM models, while also reducing computational effort. The findings underscore the effectiveness of ANN-based surrogate models for rapid and accurate evaluation of retrofit scenarios involving diverse HVAC systems, supporting more efficient decision-making in building energy retrofits.
{"title":"Surrogate models for evaluating HVAC retrofit options in multi-unit residential buildings","authors":"Harry Vallianos, Costa Kapsis","doi":"10.1016/j.enbuild.2026.117060","DOIUrl":"10.1016/j.enbuild.2026.117060","url":null,"abstract":"<div><div>This study investigates the use of surrogate modeling to optimize retrofit strategies in multi-unit residential buildings (MURBs), including HVAC systems. A comprehensive synthetic dataset was generated using EnergyPlus simulations, parameterized across a wide range of building and system variables, including ten distinct HVAC configurations. Multiple surrogate modeling approaches were evaluated, including single-output and multi-output artificial neural networks (ANNs) as well as Light Gradient Boosting Machine (LGBM) models. The models were trained to predict key performance metrics: Energy Use Intensity (EUI), Thermal Energy Demand Intensity (TEDI), and Cooling Energy Demand Intensity (CEDI). Results show that multi-output ANN models, with HVAC system as a categorical input, achieved high accuracy (R<sup>2</sup> > 0.997 and RMSE<2.2kWh/m<sup>2</sup>/year) and superior generalization compared to both single-output ANNs and LGBM models, while also reducing computational effort. The findings underscore the effectiveness of ANN-based surrogate models for rapid and accurate evaluation of retrofit scenarios involving diverse HVAC systems, supporting more efficient decision-making in building energy retrofits.</div></div>","PeriodicalId":11641,"journal":{"name":"Energy and Buildings","volume":"355 ","pages":"Article 117060"},"PeriodicalIF":7.1,"publicationDate":"2026-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146048127","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-24DOI: 10.1016/j.enbuild.2026.117055
Qi Chang , Jie Kang , Lingfeng Tang
Building load forecasting is crucial to its operation optimization. Federated learning enables distributed building load forecasting while preserving data privacy. However, a global model via directly aggregating local models cannot effectively capture personalized energy consumption patterns. Moreover, fine-tuning the global model may face the risk of overwriting globally shared knowledge. To address this issue, this paper proposes a personalized federated meta-learning approach for distributed building load forecasting. The framework consists of two parts. The first is a federated meta-learning-based global model that incorporates meta-learned auxiliary variables to extract global model parameters optimized across buildings, thereby acquiring globally shared knowledge. The second is a multi-channel residual compensation model, trained locally to capture residuals between the global prediction and actual loads, which acquires personalized knowledge not represented by the global model. The final prediction is obtained by summing the outputs of the global and personalized models, effectively balancing global generalization and local personalization. The proposed method is validated on the real-world dataset Building Data Genome Project 2, with conformal prediction employed to quantify the model uncertainty. Experimental results demonstrate that the proposed model not only improves prediction accuracy but also provides reliable uncertainty estimation through conformal prediction-based intervals.
{"title":"A personalized federated meta-learning approach for distributed load forecasting of buildings","authors":"Qi Chang , Jie Kang , Lingfeng Tang","doi":"10.1016/j.enbuild.2026.117055","DOIUrl":"10.1016/j.enbuild.2026.117055","url":null,"abstract":"<div><div>Building load forecasting is crucial to its operation optimization. Federated learning enables distributed building load forecasting while preserving data privacy. However, a global model via directly aggregating local models cannot effectively capture personalized energy consumption patterns. Moreover, fine-tuning the global model may face the risk of overwriting globally shared knowledge. To address this issue, this paper proposes a personalized federated <em>meta</em>-learning approach for distributed building load forecasting. The framework consists of two parts. The first is a federated <em>meta</em>-learning-based global model that incorporates <em>meta</em>-learned auxiliary variables to extract global model parameters optimized across buildings, thereby acquiring globally shared knowledge. The second is a multi-channel residual compensation model, trained locally to capture residuals between the global prediction and actual loads, which acquires personalized knowledge not represented by the global model. The final prediction is obtained by summing the outputs of the global and personalized models, effectively balancing global generalization and local personalization. The proposed method is validated on the real-world dataset Building Data Genome Project 2, with conformal prediction employed to quantify the model uncertainty. Experimental results demonstrate that the proposed model not only improves prediction accuracy but also provides reliable uncertainty estimation through conformal prediction-based intervals.</div></div>","PeriodicalId":11641,"journal":{"name":"Energy and Buildings","volume":"356 ","pages":"Article 117055"},"PeriodicalIF":7.1,"publicationDate":"2026-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146048126","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-24DOI: 10.1016/j.enbuild.2026.117040
Jingjie Tan , Xiaojing Zhang , Ziyang Hao , Jingchao Xie , Jiaping Liu
To address the boundary ambiguity issue in building climate zoning for sea-land interlaced region at low-latitudes in China, this study proposes a collaborative zoning method that integrates density-based clustering with subsequent classification. The method enables the identification of arbitrarily shaped climate clusters and the precise treatment of noisy samples in sea-land transition zones, thereby overcoming the limitations of K-means and hierarchical clustering approaches, which assume spherical clusters and rigid boundaries. The ERA5 high-resolution meteorological data (with a spatial resolution of 0.25° × 0.25°) is used to construct a zoning index system incorporating three elements: temperature, precipitation, and radiation. By optimizing the dual-index joint classification, the proportion of boundary-disputed samples is reduced to 0.76%, with key thresholds identified as an annual total horizontal solar radiation of 1573 kWh/m2 and a coldest-month mean temperature of 15 °C. The zoning results show that the low-latitude regions in China can be distinctly partitioned along the coastline into a New Hot-Summer and Warm-Winter Zone (wherein buildings must adequately address the heat prevention requirements and rain protection during summer) and an Extreme Hot-Humid Zone (wherein buildings require year-round heat prevention, rain protection, and full-shading design). Building energy simulations across 19 typical cities reveal that the average building energy consumption is significantly higher in the Extreme Hot-Humid Zone (101.27 kWh·m−2·a−1) than in the New Hot-Summer and Warm-Winter Zone (57.07 kWh·m−2·a−1). Moreover, the rate of energy consumption changes peaks across the climate zone boundary. These simulation results effectively validate the accuracy of the new building climate zoning.
{"title":"Building climate zoning in the sea-land interlaced region based on the clustering method","authors":"Jingjie Tan , Xiaojing Zhang , Ziyang Hao , Jingchao Xie , Jiaping Liu","doi":"10.1016/j.enbuild.2026.117040","DOIUrl":"10.1016/j.enbuild.2026.117040","url":null,"abstract":"<div><div>To address the boundary ambiguity issue in building climate zoning for sea-land interlaced region at low-latitudes in China, this study proposes a collaborative zoning method that integrates density-based clustering with subsequent classification. The method enables the identification of arbitrarily shaped climate clusters and the precise treatment of noisy samples in sea-land transition zones, thereby overcoming the limitations of K-means and hierarchical clustering approaches, which assume spherical clusters and rigid boundaries. The ERA5 high-resolution meteorological data (with a spatial resolution of 0.25° × 0.25°) is used to construct a zoning index system incorporating three elements: temperature, precipitation, and radiation. By optimizing the dual-index joint classification, the proportion of boundary-disputed samples is reduced to 0.76%, with key thresholds identified as an annual total horizontal solar radiation of 1573 kWh/m<sup>2</sup> and a coldest-month mean temperature of 15 °C. The zoning results show that the low-latitude regions in China can be distinctly partitioned along the coastline into a <em>New Hot-Summer and Warm-Winter Zone</em> (wherein buildings must adequately address the heat prevention requirements and rain protection during summer) and an <em>Extreme Hot-Humid Zone</em> (wherein buildings require year-round heat prevention, rain protection, and full-shading design). Building energy simulations across 19 typical cities reveal that the average building energy consumption is significantly higher in the <em>Extreme Hot-Humid Zone</em> (101.27 kWh·m<sup>−2</sup>·a<sup>−1</sup>) than in the <em>New Hot-Summer and Warm-Winter Zone (</em>57.07 kWh·m<sup>−2</sup>·a<sup>−1</sup><em>)</em>. Moreover, the rate of energy consumption changes peaks across the climate zone boundary. These simulation results effectively validate the accuracy of the new building climate zoning.</div></div>","PeriodicalId":11641,"journal":{"name":"Energy and Buildings","volume":"355 ","pages":"Article 117040"},"PeriodicalIF":7.1,"publicationDate":"2026-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146048180","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-23DOI: 10.1016/j.enbuild.2026.117059
Panlong Liu , Yinghuai Liang , Shuhong Li , Yanjun Li , Wei Sheng
Cooling load forecasting of central air-conditioning systems is a critical prerequisite for improving building energy efficiency and optimizing system control. Although existing data-driven prediction methods do not require complex physical modeling, their performance is still limited by the quality and quantity of historical data. Based on transfer learning, this study proposed a cross-building domain-adaptive cooling load forecasting framework and developed two domain adaptation models for cooling load forecasting: the CORAL-LSTM model and the DANN-LSTM model. These two models are a Long Short-Term Memory (LSTM) model coupled with Correlation Alignment (CORAL) and an LSTM model coupled with Domain Adversarial Neural Networks (DANN), respectively. Through transfer learning, both models integrate domain adaptation enhancement capabilities to address the scarcity of historical data. The models were evaluated using real building data under simulated data scarcity conditions. Experimental results show that in data-scarce scenarios, when the number of source domain samples reaches 4 times or more that of the target domain, compared with traditional LSTM and Gated Recurrent Unit (GRU) models, the overall Performance Improvement Rate (PIR) of the CORAL-LSTM model ranges from 70.8% to 92.62%, and that of the DANN-LSTM model ranges from 23.8% to 68.57%. Meanwhile, the Mean Absolute Percentage Error (MAPE) of the CORAL-LSTM model ranges from 0.99% to 1.15%, and its Coefficient of Determination (R2) ranges from 0.97 to 0.99. Furthermore, this study creatively introduces equivalent parameters to solve the problem of inconsistent feature dimensions when applying transfer learning models to heterogeneous systems. It also uses the Shapley Additive Explanations (SHAP) model to quantify the impact of input features on model outputs, verifying the effectiveness of the equivalent parameters. These findings confirm the feasibility of improving the performance of transfer learning models by enhancing domain adaptation in cooling load forecasting of central air-conditioning systems and provide a new technical approach to addressing data limitations in building energy system modeling.
{"title":"Domain adaptation-enhanced transfer learning framework for cross-building cooling load forecasting: Case studies in metro stations","authors":"Panlong Liu , Yinghuai Liang , Shuhong Li , Yanjun Li , Wei Sheng","doi":"10.1016/j.enbuild.2026.117059","DOIUrl":"10.1016/j.enbuild.2026.117059","url":null,"abstract":"<div><div>Cooling load forecasting of central air-conditioning systems is a critical prerequisite for improving building energy efficiency and optimizing system control. Although existing data-driven prediction methods do not require complex physical modeling, their performance is still limited by the quality and quantity of historical data. Based on transfer learning, this study proposed a cross-building domain-adaptive cooling load forecasting framework and developed two domain adaptation models for cooling load forecasting: the CORAL-LSTM model and the DANN-LSTM model. These two models are a <strong>L</strong>ong <strong>S</strong>hort-<strong>T</strong>erm <strong>M</strong>emory (LSTM) model coupled with <strong>Cor</strong>relation <strong>Al</strong>ignment (CORAL) and an LSTM model coupled with <strong>D</strong>omain <strong>A</strong>dversarial <strong>N</strong>eural <strong>N</strong>etworks (DANN), respectively. Through transfer learning, both models integrate domain adaptation enhancement capabilities to address the scarcity of historical data. The models were evaluated using real building data under simulated data scarcity conditions. Experimental results show that in data-scarce scenarios, when the number of source domain samples reaches 4 times or more that of the target domain, compared with traditional LSTM and Gated Recurrent Unit (GRU) models, the overall Performance Improvement Rate (PIR) of the CORAL-LSTM model ranges from 70.8% to 92.62%, and that of the DANN-LSTM model ranges from 23.8% to 68.57%. Meanwhile, the Mean Absolute Percentage Error (MAPE) of the CORAL-LSTM model ranges from 0.99% to 1.15%, and its Coefficient of Determination (R<sup>2</sup>) ranges from 0.97 to 0.99. Furthermore, this study creatively introduces equivalent parameters to solve the problem of inconsistent feature dimensions when applying transfer learning models to heterogeneous systems. It also uses the Shapley Additive Explanations (SHAP) model to quantify the impact of input features on model outputs, verifying the effectiveness of the equivalent parameters. These findings confirm the feasibility of improving the performance of transfer learning models by enhancing domain adaptation in cooling load forecasting of central air-conditioning systems and provide a new technical approach to addressing data limitations in building energy system modeling.</div></div>","PeriodicalId":11641,"journal":{"name":"Energy and Buildings","volume":"355 ","pages":"Article 117059"},"PeriodicalIF":7.1,"publicationDate":"2026-01-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146033331","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-22DOI: 10.1016/j.enbuild.2026.117035
Bilal Khan , Sahibzada Muhammad Ali , Zahid Ullah
The paper proposes a Model Predictive Control (MPC) energy model for environment-driven Multi-Dimensional Digital Twins (MDDTs) powered by consumer psychological behaviour accomplished via Deep Learning (DL) to minimise energy consumption. This real-time integration of environmental factors, temperature, humidity, and lighting, with consumer behaviour patterns and physiological responses, provides the basis for a new integrated model for the dynamic control of energy systems. The proposed model relies on IoT sensors and real-time data aggregation in making predictions and optimising energy consumption, considering the environmental impacts on consumer comfort. The use of DL models improves MPC by uncovering non-linear correlations in the data and having the ability to predict future energy demands. The MPC architecture operates under a closed-loop operating system and, therefore, enables adjustment of real-time feedback according to the space, environmental, and consumer behaviour changes. Due to its predictive nature, MPC can make anticipatory changes to energy systems, which will save energy without compromising comfort. The proposed model is validated using extensive simulation to respond to dynamic situations with optimal energy consumption while ensuring adequate user comfort. The real-time application of multi-dimensional heterogeneous data proves the applicability and robustness of the proposed system in real-world environments.
{"title":"Environment driven consumer psychological behavior based MPC energy model: a multi-dimensional digital twins framework using deep learning","authors":"Bilal Khan , Sahibzada Muhammad Ali , Zahid Ullah","doi":"10.1016/j.enbuild.2026.117035","DOIUrl":"10.1016/j.enbuild.2026.117035","url":null,"abstract":"<div><div>The paper proposes a Model Predictive Control (MPC) energy model for environment-driven Multi-Dimensional Digital Twins (MDDTs) powered by consumer psychological behaviour accomplished via Deep Learning (DL) to minimise energy consumption. This real-time integration of environmental factors, temperature, humidity, and lighting, with consumer behaviour patterns and physiological responses, provides the basis for a new integrated model for the dynamic control of energy systems. The proposed model relies on IoT sensors and real-time data aggregation in making predictions and optimising energy consumption, considering the environmental impacts on consumer comfort. The use of DL models improves MPC by uncovering non-linear correlations in the data and having the ability to predict future energy demands. The MPC architecture operates under a closed-loop operating system and, therefore, enables adjustment of real-time feedback according to the space, environmental, and consumer behaviour changes. Due to its predictive nature, MPC can make anticipatory changes to energy systems, which will save energy without compromising comfort. The proposed model is validated using extensive simulation to respond to dynamic situations with optimal energy consumption while ensuring adequate user comfort. The real-time application of multi-dimensional heterogeneous data proves the applicability and robustness of the proposed system in real-world environments.</div></div>","PeriodicalId":11641,"journal":{"name":"Energy and Buildings","volume":"355 ","pages":"Article 117035"},"PeriodicalIF":7.1,"publicationDate":"2026-01-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146033329","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The cleaning staff plays a crucial role in maintaining the good condition of buildings. In southern China, they were frequently exposed to non-heated cold environments in winter, where a personal comfort system is needed. However, the study about their thermal comfort is insufficient. This study seeks to examine the impact of cold stress and local heating on the thermal responses of cleaning staff and college students during cleaning work under cold winter conditions. Thirty-two participants (16 cleaning staff and 16 college students) were recruited to perform cleaning activities in a semi-open corridor (9.2 °C, 52.5% RH) in winter (outside temperature: 5.4 °C). Three heating cases (head, hands, and feet heating modes) using heating sheets with adjustable power levels (max to 5.5 W / 2 sheets) were tested to compare with the no heating case. Each case lasted for 60 min and included three 20-minute periods: windless cleaning activities (initial period), windless rest (second period), and wind-exposed cleaning activities (third period). The results indicated that the cleaning staff group felt satisfied with the cold environments under the no heating case and insensitive to local heating. Head heating was unwanted in short-term (within 20 min) cold exposure because of the higher sensitivity of the head, and overheating at the head causes thermoregulation disorder. For students, feet heating resulted in significantly lower thermal sensation vote (TSV) compared to head or hands heating during the wind-exposed cleaning activities period (p < 0.01). For cleaning staff, feet heating was found to have no significant influence on overall TSV, while hands heating maintained the lowest blood pressure throughout the experiment. Hands heating consistently resulted in the highest average thermal pleasure throughout the experiment, with mean values of 0.47 for students and 0.97 for cleaning staff. Thus, hands heating with gloves was recommended for local heating of cleaning staff in cold, windy environments.
清洁人员在保持建筑物的良好状态方面起着至关重要的作用。在中国南方，他们经常在冬天暴露在没有暖气的寒冷环境中，在那里需要个人舒适系统。然而，对其热舒适性的研究尚不充分。本研究旨在探讨冷应激和局部加热对清洁人员和大学生在寒冷冬季清洁工作中的热反应的影响。研究招募了32名参与者（16名清洁人员和16名大学生），在冬季（室外温度：5.4°C）在一个半开放的走廊（9.2°C, 52.5% RH）进行清洁活动。测试了三种加热情况（头，手和脚加热模式），使用可调节功率水平的加热片（最大到5.5 W / 2片），并与无加热情况进行比较。每个病例持续60分钟，包括3个20分钟的时间段：无风清洁活动（第一阶段）、无风休息（第二阶段）和有风暴露的清洁活动（第三阶段）。结果表明：清洁人员组对无暖气情况下的寒冷环境较为满意，对局部暖气不敏感；头部加热是不希望在短期内（在20分钟内）冷暴露，因为头部的灵敏度较高，在头部过热会导致体温调节紊乱。对于学生来说，在风暴露的清洁活动期间，脚加热导致的热感觉投票（TSV）显著低于头或手加热（p < 0.01）。对于清洁人员来说，脚部加热对整体TSV没有显著影响，而手部加热在整个实验过程中保持最低的血压。在整个实验过程中，手部加热的平均热愉悦度最高，学生的平均值为0.47，清洁人员的平均值为0.97。因此，在寒冷、多风的环境中，建议清洁人员使用手套进行局部加热。
{"title":"Effect of local heating on thermal responses of cleaning staff when working in an unheated space in winter and comparisons with the college students","authors":"Yicheng Ren , Yuxin Wu , Yujie Zhou , Yuting Li , Shuang Zheng , Yonghong Wu","doi":"10.1016/j.enbuild.2026.117039","DOIUrl":"10.1016/j.enbuild.2026.117039","url":null,"abstract":"<div><div>The cleaning staff plays a crucial role in maintaining the good condition of buildings. In southern China, they were frequently exposed to non-heated cold environments in winter, where a personal comfort system is needed. However, the study about their thermal comfort is insufficient. This study seeks to examine the impact of cold stress and local heating on the thermal responses of cleaning staff and college students during cleaning work under cold winter conditions. Thirty-two participants (16 cleaning staff and 16 college students) were recruited to perform cleaning activities in a semi-open corridor (9.2 °C, 52.5% RH) in winter (outside temperature: 5.4 °C). Three heating cases (head, hands, and feet heating modes) using heating sheets with adjustable power levels (max to 5.5 W / 2 sheets) were tested to compare with the no heating case. Each case lasted for 60 min and included three 20-minute periods: windless cleaning activities (initial period), windless rest (second period), and wind-exposed cleaning activities (third period). The results indicated that the cleaning staff group felt satisfied with the cold environments under the no heating case and insensitive to local heating. Head heating was unwanted in short-term (within 20 min) cold exposure because of the higher sensitivity of the head, and overheating at the head causes thermoregulation disorder. For students, feet heating resulted in significantly lower thermal sensation vote (TSV) compared to head or hands heating during the wind-exposed cleaning activities period (p < 0.01). For cleaning staff, feet heating was found to have no significant influence on overall TSV, while hands heating maintained the lowest blood pressure throughout the experiment. Hands heating consistently resulted in the highest average thermal pleasure throughout the experiment, with mean values of 0.47 for students and 0.97 for cleaning staff. Thus, hands heating with gloves was recommended for local heating of cleaning staff in cold, windy environments.</div></div>","PeriodicalId":11641,"journal":{"name":"Energy and Buildings","volume":"355 ","pages":"Article 117039"},"PeriodicalIF":7.1,"publicationDate":"2026-01-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146014894","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-21DOI: 10.1016/j.enbuild.2026.117044
Huiyu Yan, Jili Zhang, Liangdong Ma
Building energy consumption data quality plays a critical role in analytical accuracy, yet temporal accuracy remains underexplored compared to numerical accuracy in existing research. Our analysis of monitoring platform data reveals temporal deviations inducing up to ± 20% numerical deviations of hourly data in extreme circumstances. To address this, we develop the CCKF-SPI-EP methodology, a novel multi-sensor data fusion framework that achieves simultaneous time synchronization and constraint optimization through three key techniques: a Constrained Centralized Kalman Filter framework with normalization of cumulative energy sequences, a Shape-Preserving Interpolation for monotonic time registration, and an Estimation Projection technique for constraint incorporation. Experimental results demonstrate the method’s superiority with more than 42%–67% reduction in RMSE and 59%–76% reduction in MAE on the building’s main meter compared to the best conventional method. Furthermore, we provide practical recommendations for improving data acquisition protocols to incorporate temporal accuracy into building energy data quality assessment. This work not only presents an effective correction framework but also makes forward-looking contributions in problem awareness and data quality system development for building energy informatics.
{"title":"The time deviation of building energy consumption data and its synchronization based on the CCKF-SPI-EP framework","authors":"Huiyu Yan, Jili Zhang, Liangdong Ma","doi":"10.1016/j.enbuild.2026.117044","DOIUrl":"10.1016/j.enbuild.2026.117044","url":null,"abstract":"<div><div>Building energy consumption data quality plays a critical role in analytical accuracy, yet temporal accuracy remains underexplored compared to numerical accuracy in existing research. Our analysis of monitoring platform data reveals temporal deviations inducing up to ± 20% numerical deviations of hourly data in extreme circumstances. To address this, we develop the CCKF-SPI-EP methodology, a novel multi-sensor data fusion framework that achieves simultaneous time synchronization and constraint optimization through three key techniques: a Constrained Centralized Kalman Filter framework with normalization of cumulative energy sequences, a Shape-Preserving Interpolation for monotonic time registration, and an Estimation Projection technique for constraint incorporation. Experimental results demonstrate the method’s superiority with more than 42%–67% reduction in RMSE and 59%–76% reduction in MAE on the building’s main meter compared to the best conventional method. Furthermore, we provide practical recommendations for improving data acquisition protocols to incorporate temporal accuracy into building energy data quality assessment. This work not only presents an effective correction framework but also makes forward-looking contributions in problem awareness and data quality system development for building energy informatics.</div></div>","PeriodicalId":11641,"journal":{"name":"Energy and Buildings","volume":"355 ","pages":"Article 117044"},"PeriodicalIF":7.1,"publicationDate":"2026-01-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146033358","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The global phase-down of high-GWP refrigerant R410A has created an urgent need to identify environmentally sustainable alternatives for variable refrigerant flow (VRF) systems. This study develops a comprehensive multi-dimensional evaluation framework that integrates thermodynamic performance, environmental impact, safety constraints, and practical feasibility to assess five low-GWP candidates—R32, R452B, R454B, R454C, and R466A. The methodology combines steady-state cycle simulations for determining seasonal efficiency indices, life-cycle climate performance (LCCP) analysis for carbon footprint quantification, and standardized charge-safety evaluations in accordance with IEC 60335–2-40 and ISO 5149–1:2014/Amd 2:2021. In parallel, experimental investigations were performed to determine lubricant miscibility with POE68 and PVE68 oils. Results highlight distinct trade-offs among key indicators: R32 achieves the highest efficiency and lowest LCCP, yet its A2L flammability and limited lubricant compatibility necessitate enhanced safety engineering and material optimization. R454B demonstrates the best overall balance of environmental performance, miscibility, and retrofit compatibility, positioning it as the most practical near drop-in replacement. R466A provides non-flammable characteristics but suffers from poor miscibility, constraining its applicability. A semi-quantitative multi-criteria decision analysis identifies optimal replacement strategies under varying design priorities. The primary contribution of this work lies in integrating lubricant miscibility data and standardized charge-safety evaluation into a unified decision-support framework, delivering a transparent, scientifically grounded tool for selecting sustainable refrigerants and guiding the transition toward low-GWP VRF technologies.
{"title":"Comprehensive evaluation of low-GWP refrigerant alternatives for variable refrigerant flow air conditioning systems considering lubricant miscibility","authors":"Hongxia He, Zhao Yang, Yong Zhang, Shuping Zhang, Zixuan Bian, Cheng Liu","doi":"10.1016/j.enbuild.2026.117045","DOIUrl":"10.1016/j.enbuild.2026.117045","url":null,"abstract":"<div><div>The global phase-down of high-GWP refrigerant R410A has created an urgent need to identify environmentally sustainable alternatives for variable refrigerant flow (VRF) systems. This study develops a comprehensive multi-dimensional evaluation framework that integrates thermodynamic performance, environmental impact, safety constraints, and practical feasibility to assess five low-GWP candidates—R32, R452B, R454B, R454C, and R466A. The methodology combines steady-state cycle simulations for determining seasonal efficiency indices, life-cycle climate performance (LCCP) analysis for carbon footprint quantification, and standardized charge-safety evaluations in accordance with IEC 60335–2-40 and ISO 5149–1:2014/Amd 2:2021. In parallel, experimental investigations were performed to determine lubricant miscibility with POE68 and PVE68 oils. Results highlight distinct trade-offs among key indicators: R32 achieves the highest efficiency and lowest LCCP, yet its A2L flammability and limited lubricant compatibility necessitate enhanced safety engineering and material optimization. R454B demonstrates the best overall balance of environmental performance, miscibility, and retrofit compatibility, positioning it as the most practical near drop-in replacement. R466A provides non-flammable characteristics but suffers from poor miscibility, constraining its applicability. A semi-quantitative multi-criteria decision analysis identifies optimal replacement strategies under varying design priorities. The primary contribution of this work lies in integrating lubricant miscibility data and standardized charge-safety evaluation into a unified decision-support framework, delivering a transparent, scientifically grounded tool for selecting sustainable refrigerants and guiding the transition toward low-GWP VRF technologies.</div></div>","PeriodicalId":11641,"journal":{"name":"Energy and Buildings","volume":"356 ","pages":"Article 117045"},"PeriodicalIF":7.1,"publicationDate":"2026-01-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146033359","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-21DOI: 10.1016/j.enbuild.2026.117042
Chuanming Li , Xiangshen Gao , Rongshan Han , Nianping Li , Jibo Long , Minghao Ren , Fajin Xu , Qingqing Long
To enhance the heating performance and building load matching capability of finned evaporator heat pumps in winter, this study proposes a frost-suppression method using an air–water source finned evaporator integrated with a hot-water coil. A computational model for this combined heat transfer unit was established. Based on the air dew point temperature, an Artificial Neural Network prediction model with a coefficient of determination R2 of 0.9998 was developed, using inlet air temperature, humidity ratio, air velocity, and hot-water temperature as input variables and refrigerant heat gain as the output. This model was employed to simulate the maximum heat supply capacity and conduct load matching analysis under frost-free evaporator operation. Results indicate that a lower air humidity ratio corresponds to greater frost-free heating potential. For instance, at 5℃ air temperature, the maximum heat supplies for humidity ratios of 0.5 g/kg and 3.5 g/kg are 2.88 W and 0.38 W, respectively. Increasing the evaporator hot-water temperature significantly boosts the heat supply under frost-free operation: at −10℃ air temperature and 0.5 g/kg humidity ratio, the maximum heating capacities with 20℃ hot water and without hot water are 12.86 W and 5.54 W, respectively. Under typical winter conditions, raising the hot-water temperature effectively enhances exerts a more substantial influence on the matching rate between heat supply and building demand than varying the air velocity: in Xiangtan, increasing the temperature from 10℃ to 20℃ improves the matching rate of 11.87% (with 20℃ hot water meeting demand for 12.85% of the heating period), while in Xi’an, the corresponding improvement is 31.66% (with 20℃ hot water satisfying 50.87% of the demand). This research provides an effective methodology for frost suppression and load matching regulation in air-source heat pumps.
{"title":"Research on frost-resistant characteristics of air-water source finned evaporator based on air dew point temperature","authors":"Chuanming Li , Xiangshen Gao , Rongshan Han , Nianping Li , Jibo Long , Minghao Ren , Fajin Xu , Qingqing Long","doi":"10.1016/j.enbuild.2026.117042","DOIUrl":"10.1016/j.enbuild.2026.117042","url":null,"abstract":"<div><div>To enhance the heating performance and building load matching capability of finned evaporator heat pumps in winter, this study proposes a frost-suppression method using an air–water source finned evaporator integrated with a hot-water coil. A computational model for this combined heat transfer unit was established. Based on the air dew point temperature, an Artificial Neural Network prediction model with a coefficient of determination R2 of 0.9998 was developed, using inlet air temperature, humidity ratio, air velocity, and hot-water temperature as input variables and refrigerant heat gain as the output. This model was employed to simulate the maximum heat supply capacity and conduct load matching analysis under frost-free evaporator operation. Results indicate that a lower air humidity ratio corresponds to greater frost-free heating potential. For instance, at 5℃ air temperature, the maximum heat supplies for humidity ratios of 0.5 g/kg and 3.5 g/kg are 2.88 W and 0.38 W, respectively. Increasing the evaporator hot-water temperature significantly boosts the heat supply under frost-free operation: at −10℃ air temperature and 0.5 g/kg humidity ratio, the maximum heating capacities with 20℃ hot water and without hot water are 12.86 W and 5.54 W, respectively. Under typical winter conditions, raising the hot-water temperature effectively enhances exerts a more substantial influence on the matching rate between heat supply and building demand than varying the air velocity: in Xiangtan, increasing the temperature from 10℃ to 20℃ improves the matching rate of 11.87% (with 20℃ hot water meeting demand for 12.85% of the heating period), while in Xi’an, the corresponding improvement is 31.66% (with 20℃ hot water satisfying 50.87% of the demand). This research provides an effective methodology for frost suppression and load matching regulation in air-source heat pumps.</div></div>","PeriodicalId":11641,"journal":{"name":"Energy and Buildings","volume":"355 ","pages":"Article 117042"},"PeriodicalIF":7.1,"publicationDate":"2026-01-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146014908","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-21DOI: 10.1016/j.enbuild.2026.117038
Jihyeon Cho , Suyeon Bang , Hoseong Lee , Yeonsook Heo
Automated shading can curb building energy use; however, its performance depends on the control strategy design. This study quantified how state variable, dynamicity level, and adaptation horizon shape end-use savings and visual comfort. and formulated static and dynamic rule-based strategies for exterior slat-type blinds that track the solar horizontal profile angle (SHPA). Both schemes were tuned monthly or annually. The candidate state variables include direct solar irradiance, outdoor temperature, and indoor temperature. We simulated the total energy loads (cooling, heating, and lighting) and operational behavior (mode duration and switching frequency) for the perimeter zones across the orientations. Monthly optimized static control achieved 39–41% total-load savings vs. a no-control baseline and outperformed a fixed 90° reference with dimming (∼29%). Dynamic SHPA tracking offers marginal energy gains (<≈1 percentage point) but incurs orders-of-magnitude higher switching frequency. Hourly analysis showed that neither the slat modulation level nor state variable choice produced load savings because the shaded areas controlled by the strategies were not significantly different. Monthly tuning outperformed annual optimized cases by ∼ 12 percentage points, reflecting seasonal non-stationarity in sun geometry and weather. This indicates that the adaptation time interval is critical in rule-based shading control. Finally, visual comfort of static controls remained acceptable; hours with visual discomfort (daylight glare index > 22) were < 3% in every zone. Overall, these findings provide a practical guideline: use optimized static rules as the default approach and select the adaptation timescale according to local climate variability, while aligning shading system design with the proposed control framework.
{"title":"Exploration of key design questions in rule-based shading control for building energy load reduction","authors":"Jihyeon Cho , Suyeon Bang , Hoseong Lee , Yeonsook Heo","doi":"10.1016/j.enbuild.2026.117038","DOIUrl":"10.1016/j.enbuild.2026.117038","url":null,"abstract":"<div><div>Automated shading can curb building energy use; however, its performance depends on the control strategy design. This study quantified how state variable, dynamicity level, and adaptation horizon shape end-use savings and visual comfort. and formulated static and dynamic rule-based strategies for exterior slat-type blinds that track the solar horizontal profile angle (SHPA). Both schemes were tuned monthly or annually. The candidate state variables include direct solar irradiance, outdoor temperature, and indoor temperature. We simulated the total energy loads (cooling, heating, and lighting) and operational behavior (mode duration and switching frequency) for the perimeter zones across the orientations. Monthly optimized static control achieved 39–41% total-load savings vs. a no-control baseline and outperformed a fixed 90° reference with dimming (∼29%). Dynamic SHPA tracking offers marginal energy gains (<≈1 percentage point) but incurs orders-of-magnitude higher switching frequency. Hourly analysis showed that neither the slat modulation level nor state variable choice produced load savings because the shaded areas controlled by the strategies were not significantly different. Monthly tuning outperformed annual optimized cases by ∼ 12 percentage points, reflecting seasonal non-stationarity in sun geometry and weather. This indicates that the adaptation time interval is critical in rule-based shading control. Finally, visual comfort of static controls remained acceptable; hours with visual discomfort (daylight glare index > 22) were < 3% in every zone. Overall, these findings provide a practical guideline: use optimized static rules as the default approach and select the adaptation timescale according to local climate variability, while aligning shading system design with the proposed control framework.</div></div>","PeriodicalId":11641,"journal":{"name":"Energy and Buildings","volume":"356 ","pages":"Article 117038"},"PeriodicalIF":7.1,"publicationDate":"2026-01-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146033360","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号