Accurate point-wise classification of structural objects on building surfaces is essential for high-fidelity 3D modeling and digital twin applications. However, maintaining spatial coherence in complex façade environments remains a significant challenge due to geometric heterogeneity. This study proposes an explainable machine learning framework for classifying façade-mounted structural elements from terrestrial laser scanning (TLS) point clouds using multiscale geometric features. The dataset, comprising over 10 million labeled points across nine classes, represents dense infrastructural components attached to building surfaces. To capture both fine- and coarse-scale structural details, nine geometric descriptors were evaluated across multiple neighborhood radii ranging from 0.02 m to 1.0 m. Feature relevance was assessed using three filter-based selection methods—Fisher score, Gini importance, and Chi-square—under an 80% cumulative contribution threshold. The resulting subsets were used to train Random Forest, XGBoost, and LightGBM classifiers, with macro-F1-based hyperparameter optimization to mitigate class imbalance. XGBoost combined with Fisher score-selected features yielded the best performance, achieving a 0.920 overall accuracy and 0.811 macro-F1. Class-wise SHAP analysis revealed that vertical positioning (Z) and scale-dependent surface descriptors are the primary drivers of class discrimination. These findings enable diagnosing systematic misclassifications, proving explainable frameworks offer transparent alternatives for enriching complex building façades. Practical Application This explainable framework provides AEC and digital twin professionals with a transparent, efficient tool to process massive TLS data. By identifying scale-dependent geometric features, engineering practitioners can accurately automate the semantic enrichment of complex façade elements (e.g., pipelines and structural components). Ultimately, these diagnostic insights minimize manual annotation costs, mitigate systematic misclassifications, and directly enhance Scan-to-BIM modeling and structural asset monitoring.
Prior to the current study, the role played by the distribution of room heating on hybrid ventilation was not known. Here, we establish this role across a spectrum of heating distributions. By increasing the number n of localised heat sources, we quantify and compare hybrid–ventilation behaviour ranging from a single heat input ( n = 1 ) through to the heating of the entire floor plan ( n → ∞ ) . For a given heat load, our analysis reveals that the choice of heating distribution n has a pronounced effect on airflow rates. Indeed, increasing n can even reverse the direction of airflow. This sensitivity highlights the central role of heating distribution in hybrid ventilation. Contrary to common perception, the natural ventilation component of the hybrid system is reduced to zero as the strength of the mechanical component is increased from zero, revealing that hybrid ventilation provides only weak mechanical leverage; modest increases in hybrid-ventilation flow rate requiring large increases in mechanical supply. Key implications for design of this interaction between the mechanical component and the natural component (the latter indelibly linked to the heating distribution n ) are highlighted and discussed. Practical application The simplified mathematical model and analytical solutions developed offer practitioners a rapid predictive capability and clearer insight into hybrid ventilation. Spanning a range of room-heating distributions, the model provides a first-order tool for assessing and refining preliminary system designs. The resulting design implications show how this predictive framework supports effective decision-making. Strengthening fundamental understanding in this way should improve design guidance and build practitioner confidence when seeking to implement efficient hybrid systems. This work demonstrates how to achieve desired ventilation airflow rates and flow directions while highlighting design pitfalls to avoid.
Destination Control System (DCS) is getting popular in the elevator industry where a significant portion of new high-rise buildings are equipped with elevators controlled under DCS rather than the conventional collective control system (CCS). However, both ISO 8100-32 and CIBSE Guide D are still advising calculation as the first step of conceptual design using formulae associated with CCS under an up-peak condition. The effect of the control system is checked by simulation. Formulae for DCS in uniformly populated buildings during up-peak conditions have been suggested since the first DCS elevators were installed in the early 1990s, evolving steadily ever since. However, not even one formula is globally accepted up to this moment. This paper aims at revisiting the suggested formulae for DCS and extending these formulae to include non-uniformly populated buildings. They are studied by calculation and cross checked by computer simulation. The goal is to determine a set of formulae that safely covers DCS performance under an up-peak condition for both uniformly and non-uniformly populated buildings. Such selected best formulae must see a close resemblance between results of calculation and of simulation. In elevator planning, this new set of formulae is recommended for DCS design when simulation tools are unavailable. Practical application: CIBSE Guide D and ISO 8100-32 have been two standards widely used by the lift industry to design a lift system. Both emphasize on the importance of conceptual design by calculation as the first stage, followed by computer simulation to get the details. But only equations applicable to uniformly populated buildings for systems under the conventional collective control are available in the two publications. In this paper, existing equations for destination control systems applicable to uniformly populated buildings are re-examined and improved, and they are extended to non-uniformly populated buildings under DCS, verified by both calculation and computer simulation. With this new set of equations, it is hoped that traffic design could be more comprehensive.
Green roofs (GR) are critical for reducing building energy demands, yet thermal interactions between specific vegetation architectures and substrate layers remain difficult to quantify for building services design. This study applies the amplitude method over an annual cycle to calculate the apparent thermal diffusivity of GR simulators. Using distinct species as models for creeping ( Modiola caroliniana ) and bunchgrass ( Nassella tenuis ) architectures, we demonstrate how plant morphology fundamentally alters heat transfer mechanics. Results revealed creeping vegetation caused significantly higher diffusivity in upper substrate layers (0–5 cm), accelerating surface heat transfer. Conversely, bunchgrass restricted surface diffusivity but increased it in the deeper 5–12 cm profile. Crucially for building load calculations, this divergence yielded a substantial difference in thermal inertia: the bunchgrass system provided a 3.5 h thermal lag, significantly outperforming the 2.7 h lag in the creeping system. These findings establish that plant architectural traits are critical parameters modulating heat flux into the building envelope. This study provides actionable data to optimise GR thermal performance and peak load attenuation in variable climates.
The increasingly stringent energy and emission targets have made thermal bridging a critical consideration for building construction. Thermal bridging is excessive, localized heat flow through discontinuities in the insulation layer, such as at window and door assemblies, building envelope junctions, or balconies, that significantly increases space-heating and cooling demand, depresses interior surface temperatures, and promotes moisture-related durability problems. This paper presents a state-of-the-art critical review of thermal bridging in buildings, with emphasis on its mechanisms, quantification, and mitigation. The classifications, causes, and locations of thermal bridges within the building envelope are synthesized, and methods for evaluating them, including the Equivalent Wall Method (EWM), the Equivalent U-Value Method (EUVM), and 2D/3D numerical modelling, are compared. Element-specific mitigation strategies for openings, exposed structural components, walls, junctions, and balconies are reviewed, highlighting their influence on effective envelope R-values, building energy use, and condensation risk. The literature shows that thermal bridges account for a substantial fraction of heat loss in highly insulated buildings and become proportionally more important as envelope requirements tighten. Key knowledge gaps are identified, including the scarcity of long-term experimental validation and the need for holistic, constructible solutions that integrate structural, architectural, and thermal performance. Practical Application Thermal bridging can materially affect heating and cooling loads, interior surface temperatures, and moisture risk in high-performance envelopes. This review provides building professionals with a comprehensive foundation to identify high-impact details (openings, slab edges, balconies, and junctions), select appropriate quantification methods (Ψ/χ values, EUVM, and 2D/3D modelling) for design and compliance, and prioritize constructible mitigation strategies, such as, but not limited to, continuous insulation, thermally broken connections, and improved detailing and quality control. The review supports a more holistic, durability-focused decision-making process through efficient HVAC system sizing and smart envelope design.
This study investigates the impact of occupancy diversity on air distribution efficiency and energy use in office buildings located in hot, arid regions, focusing on Air Change Effectiveness (ACE) and the Air Diffusion Performance Index (ADPI). This is relevant in countries as Kuwait, where outdoor temperatures exceed 50°C. Air conditioning accounts for a substantial portion of energy consumption in both residential and commercial buildings. Enhancing HVAC performance while maintaining occupant comfort can reduce energy demand and support sustainability goals. Over one-year, environmental and performance data were collected from three office buildings, including temperature, relative humidity, carbon dioxide (CO 2 ), particulate matter, ACE, and ADPI. Measurements were obtained using power meters, indoor air quality monitors, and real-time sensors, supplemented by occupant comfort surveys. Building 1 maintained an average indoor temperature of 22.4°C, CO 2 concentration of 439.5 ppm, and AQI of 97.5. Building 2 recorded higher temperature (24.5°C) and CO 2 levels (493.9 ppm), with ADPI of 94.4%, compared to 100% in Buildings 1 and 3. Building 3, which includes both ground and first floors, was analyzed to assess vertical variations in occupancy distribution and HVAC performance. The corresponding results are presented separately to enable detailed comparison across spatial configurations. PMV analysis indicated increased discomfort with rising humidity. Therefore, integrating occupancy diversity with optimized HVAC control enhances thermal comfort and energy efficiency in extreme climates. In Kuwait the building sector is a major energy consumer with residential and commercial buildings requiring a large amount of energy, mainly due to the high demand for air conditioning. Improving HVAC performance and optimizing building energy management are therefore critical strategies to reduce peak demand, lower operational costs, and support national decarbonization objectives.
Rapid global urbanisation necessitates multi-family housing designs that enhance residents’ physical, social, and psychological well-being. However, prevailing assessment tools often prioritise environmental and economic aspects while underrepresenting architect-controlled design attributes. This oversight has significant yet under-studied effect: poor architectural design can increase operational energy loads through reliance on mechanical systems. This study develops and validates a framework for assessing architectural design quality in relation to residents’ well-being in multi-family residential buildings, addressing gaps in existing assessment tools. Using a PRISMA - guided systematic scoping review, 2536 indicators were identified. Unsupervised machine learning identified thematic clusters, which were refined through expert review based on the study’s theoretical foundation and scope. The criteria and indicators were weighted using the Analytic Hierarchy Process (n = 51 architects) and refined through resident survey analysis (n = 411), resulting in a framework comprising seven criteria and 21 indicators. From the architects’ perspective, natural ventilation (25.3%) emerged as the highest-weighted criterion, followed by daylighting (21.7%) and functionality (21.6%). The findings highlight the role of passive architectural design in supporting well-being. The study presents a novel, design-focused, decision-support framework that complements existing assessment tools by introducing design quality as a foundational layer for early-stage housing design decision-making. Practical application The study validates a framework comprising seven criteria and 21 indicators aligned with the tripartite well-being model, keeping architectural design quality at the forefront of residential assessment. Weighted criteria enable architects to prioritise passive design strategies, particularly natural ventilation and daylighting, that enhance resident satisfaction while reducing reliance on mechanical systems. During concept design and massing iterations, the framework can assess spatial configurations, test alternatives, and develop client briefs against evidence-based well-being criteria. The scoring supports benchmarking among alternatives within a project and across residential developments. Apartment-level well-being scores can guide prospective residents towards options that best support their well-being.
Heat networks (HN) play a key role in strategies proposed by the UK’s Committee on Climate Change, and evidence-based standards are recognised as essential for good design and operation. Collaboration across industry has culminated in technical standards, however, evidence has shown that methods currently recommended for estimating demand diversity, widely used in the UK, tend to oversize networks and lead to avoidable thermal losses. This study uses high-frequency data from a recent UK case study HN to investigate the limitations of these methods and to explore how data from existing networks can be used to inform the design of new networks. Results show that peak demands are adequately captured at a resolution of 10 min, a useful benchmark for data-informed network sizing. An assessment of oversizing relative to current sizing methods shows that the case study may potentially be oversized by 52% to 140%. The agreement between measured total peak demand and domestic hot water (DHW) demand design estimates suggests that inclusion of the space heating (SH) component in the design estimates may be unnecessary. The empirical demand curves developed in this work may be a useful reference for designers, as part of a larger empirical evidence base.
Controlling smoke propagation in asymmetrical V-shaped tunnels remains a formidable challenge, particularly at slope transitions where buoyancy-driven stack effects often compromise traditional ventilation strategies. This study proposes and optimizes a synergistic smoke control system that integrates side-wall extraction with an active air curtain barrier. Using Fire Dynamics Simulator (FDS), a comprehensive parametric analysis was conducted under varying fire loads (5-15 MW) and geometric configurations, specifically targeting the critical scenario of a 1% fire-side and 3% non-fire-side gradient. The results demonstrate that the proposed system significantly outperforms standalone extraction, effectively confining smoke to the fire source side and reducing peak ceiling temperatures. Through sensitivity analysis, jet velocity and injection angle were identified as the dominant control parameters, with an optimized configuration (2 m/s at 15 degrees) achieving maximum confinement efficiency. Crucially, based on dimensionless analysis, predictive correlations are proposed for the ceiling-level temperature rise as functions of the lateral extraction velocity and air-curtain parameters, showing good agreement with simulations. These predictive models provide a robust theoretical tool for fire safety engineering, offering practical guidance for designing resilient smoke control systems in complex V-shaped underground infrastructures.Practical application This study provides building services engineers and fire safety consultants with a validated smoke control strategy for complex asymmetrical V-shaped tunnels. By integrating side-wall extraction with optimized air curtains, professionals can effectively prevent smoke backlayering at critical slope transitions. The research defines optimal design parameters-specifically identifying jet velocity and angle as primary control factors-offering a more space-efficient and cost-effective alternative to traditional oversized longitudinal ventilation. The proposed dimensionless correlations serve as a direct technical reference for sizing extraction systems, ensuring enhanced life safety and structural protection in modern underground infrastructure design.
To accurately analyze student evacuation patterns and fire spread dynamics in the event of a fire, this study developed a 1:1 building information model (BIM) of a university's No. 2 teaching building. The Agent-based modeling approach was adopted to account for the complexity of individual evacuation behaviour, with particular attention to factors influencing evacuees' self-perception. Three evacuation scenarios were simulated using Pathfinder: a partial first-floor evacuation (1/4 area), a full first-floor evacuation, and a whole-building evacuation. The results indicate that during evacuation, individuals tend to choose exits based on familiarity and reasonable proximity rather than strictly opting for the shortest path. Additionally, PyroSim was employed to simulate the spread of smoke and temperature changes within the building. Simulations revealed that smoke accumulation near Stairs1 significantly affects egress capacity; when Exit 1 becomes impassable at 76.5 s, the total evacuation time reaches 256.8 s, which is 23.03 s longer than that under fire-free conditions. This study provides a comprehensive BIM-based framework for simulating evacuation processes in university teaching buildings, offering a scientific basis for the design of targeted and effective fire safety strategies in public educational buildings.Practical application This study provides a critical tool for enhancing fire safety in educational buildings. By integrating occupant behaviour with realistic fire dynamics, our model delivers accurate evacuation predictions, overcoming the limitations of traditional methods. The findings directly benefit architects and safety engineers by enabling the design of more effective evacuation routes and safer building layouts. Furthermore, university administrators and safety officers can utilize this approach to develop data-driven emergency plans, optimise alarm systems, and conduct targeted drills. Ultimately, this research offers a practical framework for improving evacuation efficiency, informing safety regulations, and safeguarding lives in complex public structures like teaching buildings.
The mass balance equation is utilized for theoretical analysis and experimental verification of the indoor PM2.5 concentration within residential buildings equipped with air cleaners. This study delves into the impact of various factors such as the positioning of air cleaners, Clean Air Delivery Rate (CADR), and the volume of the room on the non-uniform mixing model of air purification systems through meticulously designed experiments. Furthermore, an exponential function relationship between Effective Air Cleaning Ratio (EACR) and the air exchange rate of the air cleaner's circulation is derived. When the air cleaner is positioned at the central position, the calculated EACR is 0.8. When placed against the wall, the calculation result is 0.76. In contrast, when situated in the corner, the indoor PM2.5 distribution is unbalanced due to the air flow delivered by the air cleaner, the EACR is only 0.74. To obtain the purification capacity of air cleaner in actual environments, an empirical correlation between EACR, room volume, and the nominal CADR is established based on experimental results.Practical application The placement of an air cleaner is a critical factor that directly influences its Clean Air Delivery Rate (CADR), resulting in a notable discrepancy between its laboratory-rated and actual performance for controlling indoor PM2.5 concentrations. The rated CADR value is insufficient for predicting actual removal effectiveness, the actual purification performance must account for practical room conditions, including room volume and airflow patterns. To obtain the purification capacity of air cleaner in actual environments, an empirical correlation between EACR, room volume, and the rated CADR is established based on experimental results. This study can provide a basis for effectively controlling indoor PM2.5 levels.
With the progression of urban renewal, the functional transformation of numerous old industrial heritage buildings has imposed new demands on their indoor physical environments. This paper focuses on the adaptive renovation of thermal environments in old industrial buildings, using two case studies: Welding Workshop (Before Renovation) and the Cylinder Casting Workshop (After Renovation) of Hefei Motor Factory and Diesel Engine Factory. By integrating on-site thermal environment measurements and subjective thermal sensation questionnaires, we employs statistical regression methods to analyze the relationship between operative temperature and actual thermal sensation (MTS) and subjective thermal discomfort. The study identifies the acceptable temperature range and duration proportion in old industrial buildings, and further compares objective and subjective differences in human thermal comfort between summer and transitional seasons in the same workshop. Based on the acceptable duration proportion, a quantitative relationship between subjective sensations and operative temperature is established. These findings offer theoretical and empirical support for green renovation strategies of existing industrial buildings and design optimization of new constructions. Practical application This study provides empirical, decision-support evidence for the green renovation of industrial heritage. At its core, it establishes operative temperature as a critical design parameter and adopts the acceptable duration proportion of thermal comfort as a quantifiable target-thereby translating comfort needs into actionable design language. The data support a practical approach combining enhanced building envelope performance with flexible indoor environmental adjustments to balance heritage preservation and thermal comfort improvements. This research framework can be integrated into the design justification, scheme comparison, and post-occupancy evaluation processes of similar projects, offering a scientific and operational reference for enhancing environmental performance in the adaptive reuse of industrial heritage.
Accurate estimation of water demand in buildings is essential for designing safe, efficient, and sustainable water supply systems. Conventional design approaches often lead to significant overestimations of water demand, resulting in water supply systems that are frequently oversized. This study introduces the Water Demand Estimation Model (WDEM), a novel stochastic model developed specifically for application to non-residential buildings. The model integrates statistical data on sanitary appliances and user behaviour using extensive Monte Carlo simulations to generate realistic scenarios of simultaneous appliance usage. In addition to appliance properties, WDEM accounts for building occupancy, an essential factor in design flow rate estimation. It provides a set of user-friendly design equations, as an essential step towards future application in practice. Application of WDEM to three diverse case study buildings revealed substantial reductions in estimated design flow rates, ranging from 63% to 73%, compared to current UK design guides. These findings demonstrate WDEM's effectiveness in estimating water demand and thereby avoiding system oversizing, which is crucial for designing water supply systems that are cost-effective and have improved water quality.Practical application The WDEM provides a new approach to estimate water supply design flow rates by combining occupancy-based usage with appliance-efficiency ratings to derive explicit design equations. The outcomes include preliminary design equations which, once validated, should support improved estimation of design flows without running the simulation. In practice, WDEM may support faster early-stage decisions, potentially smaller pipework and storage volumes, reduced capital costs and pumping energy, and lower stagnation and water quality risks.
Indoor carbon dioxide (CO2) accumulation in university classrooms is associated with fluctuations in students' cognitive performance. This study investigated the preliminary associations between short-term CO2 exposure, heart rate (HR), and cognitive accuracy (ACC), leveraging a small-N intensive longitudinal design with 54 synchronized observation sets over 6 days. Results indicated that HR responded to CO2 in two stages under the observed conditions: an initial sensitive response with relatively stable HR when CO2 was below the 1000 ppm reference level, followed by a gradual adaptive decline as concentrations increased. Standardized mediation analysis confirmed a global indirect effect (beta = -0.106, p < 0.05). Notably, segmented analysis revealed that this physiological-cognitive coupling was primarily driven by the intensified impact of HR on cognitive accuracy in Phase II (beta = 0.389, p = 0.022), whereas the mechanism remained exploratory in Phase I. These preliminary findings suggest that the relationship between CO2 and cognition may be mediated by autonomic regulation (reflected by HR). Under the observed classroom conditions, 1000 ppm may serve as a guideline-aligned environmental reference associated with physiological-cognitive shifts. Practical applications: This pilot study suggests that heart rate (HR) mediates the association between indoor CO2 and cognitive performance in university classrooms. Specifically, CO2 levels exceeding 1000 ppm were associated with distinct physiological changes and reduced cognitive accuracy, highlighting this value as a critical reference for ventilation control. Practically, these results support implementing occupancy-sensitive ventilation strategies to limit CO2 accumulation. Additionally, the observed CO2-HR coupling indicates that aggregate HR trends derived from wearable devices could serve as non-invasive, supplementary indicators of indoor environmental conditions. These exploratory findings inform future human-centric approaches to indoor environmental quality (IEQ) management in educational settings.
CIBSE weather files are currently used by the building industry as the standard input data for building performance assessment for the purpose of regulatory compliance in the UK. In this study, the state-of-the-art CIBSE weather files are created with four major improvements incorporated, namely, (1) the enhanced representation of the UK climate through the creation of discriminative climate zones; (2) the latest climate change signals from the UK Climate Projection 2018 (UKCP18); (3) the satellite based solar radiation data from CAMS (Copernicus Atmosphere Monitoring Service) data repository; (4) the up-to-date observation record from 1994 to 2023. The methodology for creating the latest CIBSE weather files is elaborated in detail to enhance the transparency of the new weather data. Evaluated using a simulation case study, the new weather files demonstrate spatial and temporal coherency. The new future weather files enable robust building performance assessment against future climate conditions under different scenarios and will play an important role in designing climate-resilient buildings and delivering a net zero built environment.Practical applications As per the principle of "garbage in, garbage out", weather data plays an instrumental role in streamlining building design to achieve both energy efficiency and thermal comfort. In this study, we present the methodology for the creation of the state-of-the-art CIBSE weather files. The new CIBSE weather files not only employ the update-to-date observation and projection data, but are also grounded on a total of 28 granular climate zones to account for diverse climate characteristics and eliminate the ambiguity with weather data selection. The new files will lay a solid data foundation for future-proofing building design in the UK.
Dry bulb air temperatures are increasing especially in the UK. On top of this there is the urban heat island intensity (UHII), the temperature difference between rural and hotter urban temperatures. This is important as most new and existing buildings are in urban areas subject to the urban heat island (UHI)., In this paper is an analysis of weather data in Manchester city centre is compared to the CIBSE Design Summer Year type 3 (DSY 3) weather data. The latter, based on weather data from the Meteorological Office at a non-urban contains little if any UHII data. The DSY3 is based on a very warm year but the lack of the UHII data severely underestimates city centre night time temperatures and overestimates the winter minima temperatures. This underestimates the heating effect for net-zero design for urban buildings. A simple algorithm, which has been derived from Manchester and London data, is shown to give a useful method of adding in the UHII.Practical application For the weather data, including the urban heat island effect, for the design of buildings and plant.
The outbreak of cold waves often causes a rapid decrease in temperature and a sharp increase in the heating load demand of urban buildings, imposing enormous pressure on the energy and power systems of megacities. In this study, the spatiotemporal variations in cold waves were analyzed over a 30-year period (1991-2020) in Tianjin, China. Based on the above study, three typical cold wave (CW) events were selected, the rural weather stations were selected using satellite-based methodology, and the impacts of UHI effect on heating loads of residential buildings during CW periods were evaluated by simulating the hourly heating loads during CW and non-CW periods. The results show that the UHI intensity (UHII) was lower during CW periods than non-CW periods. The UHI effect reduced urban heating loads by 8.72% compared to rural areas during CW periods. During high-load periods (18:00 (Beijing time) to 07:00 the following day), urban heating loads were 10.24% lower than rural areas, while urban heating loads were 11.65-16.04% lower than rural areas during non-CW periods. Therefore, during CW periods, the UHII weakens, and its impacts on residential building loads are reduced.Practical application This study provides the variations of UHII during typical CW events in cold climates, and evaluates the impacts of UHI effect on residential heating loads. The response of the heating load variations in residential building to UHI effect during CW periods should be comprehensively considered to improve the fine-level of heating operation regulations, especially in urban areas in cold climates, to reduce the heating energy consumption and emission of buildings.
Accurately predicting Energy Use Intensity (EUI) has become increasingly important in efforts to enhance energy efficiency and sustainability in buildings. This study aims at comparing the performance of three machine learning approaches, namely, Baseline Ensemble, Auto Hyperparameter Optimized Ensemble and Bayesian Optimized Ensemble, using real world sensor data collected from four zones of a university building in Thailand via a Building Energy Management System. The models are tested on both un-normalized and min-max normalized datasets to examine how data preprocessing influences prediction accuracy, error reduction, and training efficiency. The results of the study demonstrate that normalization improves prediction precision by significantly reducing mean absolute error and mean squared error values, although it has a limited effect on R2 values. Among the three approaches, the Bayesian Optimized model trained on normalized data provides the most accurate and stable results while maintaining reasonable training times. These results highlight the practical value of integrating normalization and automated tuning when designing building energy models. The proposed Least Squares Boosting (LSBoost) - Bayesian Optimization model in the study offers a reliable and adaptable tool for forecasting Energy Use Intensity, with potential applications in real-time control, diagnostics, and long-term energy planning.Practical application This study presents a robust, data-driven framework for accurately forecasting Energy Use Intensity (EUI) in real-world building operations using Bayesian-optimized LSBoost models. By integrating indoor sensor data, external weather variables, and advanced machine learning, the proposed method supports energy managers, building operators, and HVAC control engineers in enhancing predictive maintenance, operational efficiency, and real-time energy management. The approach is particularly suited for smart building systems and retrofitting strategies that require scalable, accurate, and resource-efficient energy modeling under variable occupancy and environmental conditions.
Excessive use of fossil fuel-based energy has led to significant environmental problems, primarily due to greenhouse gas emissions. In regions with hot and dry climates, natural ventilation strategies, such as wind catchers, can reduce the demand for cooling and ventilation, thereby improving thermal comfort and indoor air quality. This study evaluates the performance of a wind catcher in the hot, dry climate of Mexicali, Mexico, using Computational Fluid Dynamics (CFD) to analyze ventilation efficiency and thermal comfort across nine different air inlet and outlet configurations. The results indicate that configurations promoting cross-ventilation, particularly air inlets and outlets positioned at the bottom and top of the room, respectively, achieve the most efficient air renewal (with an average 18.32% reduction in stagnation time compared to the others) and improved thermal comfort, resulting in a predicted dissatisfaction percentage of 35.1%. These findings highlight the potential of wind catchers as a passive cooling strategy to reduce electricity consumption in hot and dry climates.Practical application This study provides clear design guidance for architects and engineers aiming to optimize natural ventilation in hot-dry climates. By identifying the most effective inlet and outlet positions in wind-catcher systems, the findings help reduce air stagnation and improve indoor thermal comfort without relying on mechanical cooling. The proposed configurations support energy-saving strategies in building design, particularly in regions with high cooling loads. These insights are applicable not only in Mexican cities such as Mexicali, but also in similar arid environments globally, contributing to the advancement of passive cooling solutions in climate-responsive architecture.
Occupant window interaction is a critical component in optimizing energy consumption and indoor environmental quality (IEQ). Understanding the influence of environmental and behavioral factors on window state decisions remains a significant challenge in building management systems (BMS). We present a hybrid probabilistic model to assess thermal comfort and predict the probability of the occupant opening or closing the window. The data was acquired from an open-source platform that provided yearly university dormitory window interactions. Bayesian networks (BNs) and logistic regression (LR) models were applied to predict the window-opening behavior of the occupants. An average accuracy of 92% for Bayesian and 94% for LR were obtained. The results were further enhanced by combining these models through weighted methods, with weights extrapolated through generative recursive iterations generating an average accuracy of 95% and Area Under the Curve (AUC) of 98%. The proposed hybrid approach significantly improves over existing predictive models in thermal comfort and window state prediction.Practical Application This research provides a practical tool for building engineers, facility managers, and smart system developers to significantly improve energy efficiency and occupant comfort. The developed hybrid model predicts window-opening behavior with high accuracy (95%). This enables the creation of next generation BMS that can anticipate occupant needs, proactively adjust heating, ventilation, and air conditioning (HVAC) operations, and reduce unnecessary energy consumption. For building designers, the model offers data-driven understandings into realistic occupant behavior (OB), leading to better-performing natural ventilation approaches.
