Traditional data-driven models for building cooling load prediction often fail during the initial operational stage due to insufficient training data, leading to unreliable forecasts. To address this challenge, a knowledge-data hybrid forecasting framework was proposed, it combines simplified heat-transfer-based load calculations with deep learning networks, where physics-based load estimates are embedded as auxiliary inputs to guide the data-driven predictor. Four models were evaluated, including three knowledge-embedded variants (using theoretical fresh-air load, envelope heat-transfer load, and their combination) and a purely data-driven CNN–LSTM baseline. Validated on an actual building under varying training-data richness, models based on the proposed framework reduce prediction errors by 39% to 69% and decrease error variance by nearly an order of magnitude compared with the baseline while effectively mitigating overfitting in small-sample scenarios. The computational simplicity and reliance only on easily accessible data underscore its practical value for early-stage engineering deployment.

Amid global warming and urbanization,building energy systems face the dual challenge of balancing growth in energy demand with environmental sustainability and resistance to future climate change.This study proposes a predictive framework that integrates the effects of future climate change and urban microclimate into energy consumption prediction and energy system optimization for typical office buildings in Hangzhou,China.First,optimal general circulation models(GCMs)from CMIP6 are selected through a performance evaluation,and statistical downscaling is employed to generate future typical meteorological year(TMY)data.Next,the urban weather generator(UWG)is used to simulate urban heat island(UHI)effects.Empirical formulas are applied to calculate urban wind speeds,while DesignBuilder is used to model solar radiation and hourly energy consumption.These data are then used to optimize the building energy system.The results reveal that future climate change significantly increases cooling demand(28.9%-103.0%)and reduces heating demand(19.7%-52.6%),with urban microclimates further amplifying these trends.The energy system optimization demonstrates that the net present value(NPV)of future climate and urban microclimate scenarios is 5.1%-16.7%higher than that of historical climate scenarios.Additionally,future climate scenarios result in higher peak energy demand and thus necessitate larger system capacities to ensure reliability.While the initial required investment is higher,buildings optimized to account for global warming are more reliable and carry lower operational costs.We comprehensively quantify the effect of future urban microclimate on building energy systems,emphasizing its critical role in energy system planning and providing insights for addressing the challenges of climate change and urbanization.

Residential building energy consumption is a complex interplay of geography, architecture, and human behavior. However, current models fail to adequately integrate these factors, leading to generic strategies that lack regional specificity. This study introduces a physics-based End-Use Energy Predictive Model (EUE-PM) that explicitly quantifies the sensitivity of energy use to regional climate conditions, main facade orientation, and HVAC operational behaviors. The scenario-based comparative simulation addresses these factors across diverse climates and represents a significant advancement over previous approaches. We developed a database of climatic conditions and building archetypes to simulate 80 energy models across 40 cities in Japan and Vietnam, spanning a wide range of latitudes and climate zones. Our analysis reveals two key findings. First, sensitivity to building orientation is highly latitude-dependent, with high-latitude Japanese cities showing up to 16

Natural ventilation (NV) is an effective passive strategy for reducing building cooling demand, yet its performance varies substantially with climate and may not increase proportionally with larger openings. This study examines the nonlinear cooling benefits of NV in detached residential buildings under different climatic conditions in Japan using dynamic building energy simulation. Eight representative cities were selected to evaluate how regional temperature and humidity influence NV availability, indoor thermal response, and cooling-energy demand under multiple window-opening scenarios. The results reveal strong climate dependence. In temperate regions, NV provides the greatest cooling benefits by enhancing nighttime heat dissipation and reducing multi-day indoor heat accumulation. In cold regions, cooling demand is relatively low, resulting in limited or negligible net cooling-energy benefits from NV. In warm-humid regions, high moisture conditions significantly weaken the cooling effect of larger openings. Across all climates, the relationship between window opening area and cooling-energy savings is nonlinear and not universally monotonic. In seven of the eight cities, the 5% opening scenario provides the best annual cooling-energy performance among the examined non-zero opening conditions. These findings indicate that NV design should be climate-responsive rather than based on uniform opening assumptions and provide practical guidance for region-specific residential cooling strategies.

In modern urban environments, limited space and the drive for building efficiency have led to the increasing prevalence of windowless offices. Because lighting is essential for mood, alertness, cognitive performance, health, and productivity, the lack of natural daylight and outdoor views can adversely affect well-being and work outcomes. Consequently, artificial windows (AW), designed to replicate natural light and external views, have been proposed. This study compared real windows (RW), AW, and no windows (NW) to investigate their effects on cognitive performance and productivity among office workers during working phases. A controlled experiment was conducted, combining heart rate variability (HRV) data and cognitive tests. In addition, questionnaires were used to evaluate participants′ subjective perceptions of the thermal and lighting environments. The results showed that AW led to better attention, short-term memory, and creativity. In particular, the time for attention tests was about 14.4 % shorter than under NW, and the time for short-term memory tests was reduced by 6.7 %. Focus, measured by the HRV (nLF/nHF) ratio, increased by approximately 22.2 %, while subjective work efficiency demonstrated an even more substantial increase. Moreover, AW provided superior thermal and lighting perceptions compared to NW and even outperformed RW in lighting comfort and acceptability. Overall, these findings indicate that AW holds considerable potential for improving attention, productivity, and environmental comfort during working phases, offering an effective alternative for windowless offices.

This study examines the contributions of thermal, acoustic, and luminous environments to students’ comprehensive comfort in university activity rooms. Field measurements and surveys were conducted over six days with 198 students at the University of Kitakyushu, Japan. Correlation and regression analyses indicate that acoustic comfort is the dominant determinant of comprehensive comfort under non optimal conditions, while enhancements in thermal and luminous conditions beyond basic levels yield minimal additional benefit. These results underscore the priority of acoustic interventions—such as sound-absorbing materials and acoustic zoning—over further improvements in temperature or lighting. Despite limitations including a short study duration, single-site sample, and limited environmental variables, the findings offer practical guidance for optimizing multifunctional academic spaces. Future research should expand to diverse building types, climates, and broader environmental dimensions to validate and generalize these conclusions.

Rural public spaces are crucial to the daily activities of older adults; however, limited research has examined the effects of their environmental characteristics on older adults’ spatiotemporal behavior and perception from a multisensory perspective. This study hypothesizes that composite sensory environments have significant nonlinear predictive effects on older adults’ behavior types and satisfaction. In this study, 10 sample spaces were selected in Qingdao, China. Multi-source data were collected through a two-week period of unobtrusive observation and subjective questionnaire surveys (N = 241). Multiple logistic regression was used to analyze the main effects of environmental characteristics, and an MLP model with a single hidden layer of 100 units was constructed to predict dwell time and satisfaction. The results show that, in the investigated rural context, older adults’ dominant behavior was social activity (81.12%), which mainly occurred in built spaces such as squares. Multiple logistic regression indicated that, among the various environmental factors, visual aesthetics had a statistically significant effect on behavior types (p = 0.013). The MLP model achieved prediction accuracies of 85.3% for dwell time and 93.1% for satisfaction. The key predictive variables were volume perception (100% importance), the Natural Sound Index (NSI) (92.1%), and visual aesthetics (89.3%). Subgroup heterogeneity analysis further showed that older-old adults and those with poorer health conditions were more sensitive to pavement quality and physical comfort, whereas older adults living alone or with limited household companionship were more strongly influenced by visual aesthetics and natural soundscape quality. The theoretical significance of this study lies in proposing quantitative measures of natural sound and odor indices and revealing that, in the specific northern rural built environment, the coordinated design of visual and auditory environments plays an important role in improving spatial quality. The findings provide empirical support for the age-friendly micro-renewal of rural public spaces in specific regions. However, due to the limitations of single-season data and a relatively small sample size, their generalizability needs to be further verified across regions.

Building envelope systems in coastal environments are continuously exposed to intense solar radiation and ultraviolet irradiation, high humidity, temperature fluctuations, and salt deposition. Under such conditions, durability degradation is rarely governed by a single factor, but is instead driven by coupled multi-physical actions. This review focuses on polymeric and organic coating-substrate systems for coastal building envelopes and systematically examines their degradation behavior and durability evolution under the combined effects of ultraviolet radiation, hygrothermal cycling, and salt input, with particular attention to how substrate type - including metals, cementitious materials, wood-based materials, and FRP/composites - fundamentally determines interfacial failure mechanisms and coating durability. A unified analytical framework is established based on the chain of environmental loads-transport/reaction-damage evolution-performance degradation. Particular attention is given to the destabilization of organic protective layers, polymer-modified regions, interfacial adhesion, and barrier integrity under coupled fields. Ultraviolet radiation can induce surface photooxidation, chain scission, embrittlement, and chalking. Fluctuations in temperature and humidity promote the initiation and propagation of microcracks through repeated expansion and contraction, molecular transport, and changes in moisture state, while also accelerating oxidation- and hydrolysis-related reactions. Salt input further intensifies underfilm corrosion, salt-crystallization-induced stress, and interfacial deterioration under persistently wet or wet-dry cycling conditions. These degradation processes generally exhibit pronounced characteristics of surface initiation, gradient propagation, and interface amplification, especially in coating systems and polymer-modified protective systems. Metals, cementitious materials, wood-based materials, and FRP/composites are discussed as representative substrates whose interfacial properties - including electrochemical activity, porosity, moisture retention, and defect sensitivity - fundamentally determine coating durability, rather than as independent parallel material categories. On this basis, this review establishes a unified analytical framework linking environmental loads, transport/ reaction processes, damage evolution, and performance degradation for polymeric and organic coating-substrate systems in coastal building envelopes. By reorganizing degradation mechanisms around coating integrity, substrate dependence, moisture/salt transport, interfacial stability, and barrier-function retention, the review clarifies how coupled coastal exposure governs coating failure and long-term durability. The proposed framework provides guidance for coating-system design, material selection, protection optimization, degradationindicator prioritization, service-life prediction, and maintenance planning under coastal service conditions. It also highlights future needs for field-validated accelerated aging, hygrothermal and electrochemical model validation, active/smart coating strategies, and durability-based lifecycle management of coastal protective coating systems.

Human exploration of Mars will require habitat morphologies that reduce long-term resupply and maintenance burdens under extreme environmental forcing, yet transferable geometric design laws remain poorly constrained. An explainable reverse-engineering framework is developed using refurbishment-phase embodied-carbon intensity as an Earth-analog proxy for maintenance-driven material replacement and logistics load. Across 59192 extreme-environment built habitats sampled from 384 cities, prefabrication-enabled benefits are quantified by absolute savings Δ and relative improvement η. Δ spans 0.211–3.942 kgCO2e m⁻³ (mean 1.874) and η spans 0.026–0.427 (mean 0.114), with weak coupling between national means (r = −0.154), indicating distinct drivers for absolute versus proportional gains. SHAP interaction topology reveals non-additive plateau/ridge/saddle optima, defining a narrow, mutually reinforcing proportion band (Shape Factor ≈ 1.01–1.06; Footprint Ratio ≈ 0.92–0.94; Aspect Ratio ≈ 0.93–0.96). Dual-objective screening yields a sparse non-dominated frontier (n = 7) and a best-balanced solution (Δ = 4.02 kgCO2e m⁻³; η = 0.186), supporting a constrained, testable morphology-threshold law. A synthesis of 530 algorithm-driven extra-terrestrial habitat morphology studies (1981–2025) shows rapid post-2015 expansion and a shift toward deep/generative approaches, while a scan of 74 Mars-habitat R&D platforms across 19 countries indicates that most real-world efforts cluster at TRL 4–6, with high-maturity field sites dominating TRL 7–9. Together, these results connect a transferable geometric threshold corridor to the evolving algorithmic toolkit and current technology-maturity pathways for Mars habitat development.

Abstract Most architects rely on aesthetic intuition, often neglecting the mathematical foundations that underlie aesthetic judgement in design. This study addresses this gap by proposing a mathematical framework that supports both aesthetic excellence and low-carbon performance in architecture. The first contribution is to identify the essence of modular aesthetics as rooted in classical mathematical constructs—ratio, sequence, and spiral. The second contribution is an empirical investigation of 100 award-winning architectural projects using multivariate algorithms to examine the relationships among ratio–sequence–spiral features, award recognition, and decarbonization. The results indicate statistically significant associations between ratio-derived sequences and spiral configurations and both aesthetic recognition and carbon reduction, suggesting a viable pathway for integrating these principles into low-carbon architectural design. Overall, the proposed framework advances a mathematically grounded approach to architectural design that aligns aesthetic quality with environmental goals, while also offering foundational mathematical logic that may inform future robotic 3D-printing applications in extreme environments (e.g., lunar and Martian habitats) and the emerging AI-driven transformation of architectural design and construction.

Advancements in sensing technologies have propelled the integration of physiological sensors into environmental assessment approaches, providing a nuanced understanding of the complex interplay between the environment and human well-being. This research explores the incorporation of physiological sensors, such as those monitoring skin temperature (ST) and heart rate (HR), into traditional environmental sensing frameworks. The study investigates the potential of physiological data to enhance the comprehensiveness of environmental assessments, shedding light on the direct impact of environmental conditions on individuals. The research aims to unveil new insights into how physiological sensor data integration can refine real-time monitoring data, contribute to personalized environmental assessments, and ultimately influence decision-making for enhanced environmental quality and human health. Challenges related to technology adoption, privacy considerations, and user acceptance are also explored, providing a comprehensive overview of the opportunities and obstacles in the integration of physiological sensors for environmental assessment.

With the advancing implementation of the 'carbon peak and carbon neutrality' strategy, reducing building energy consumption and improving indoor thermal comfort have become increasingly critical. Targeting regions with high summer temperatures and significant solar radiation, the spray-cooled green roof system developed in this study offers a viable solution for enhancing passive cooling in buildings and supporting climate-responsive design. This study proposes a novel spray-cooled green roof system to enhance thermal regulation efficiency while minimizing resource consumption. Through experimental measurements, we systematically analyzed the effects of four key air parameters-temperature, relative humidity, moisture content, and enthalpy-on the system's cooling performance. The results demonstrate a maximum indoor temperature reduction of 7.12 degrees C and effective suppression of heat transfer, primarily attributable to the synergistic interplay of spray evaporation, vegetation insulation, and plant transpiration. We developed predictive models quantifying the relationships between spraying operations and thermal-humidity parameters, with polynomial regression achieving excellent correlation (R-2 > 0.8). Furthermore, we established a coupled-factor model that comprehensively evaluates the thermal regulation effectiveness of spray-cooled green roofs. Based on these models, we optimized the spray control strategy, maintaining equivalent cooling performance while reducing water consumption by 40%. This research provides valuable insights for developing costeffective, high-efficiency green infrastructure supporting urban sustainability under the dual carbon strategy.

In the context of societal digital transformation, intelligent control of buildings and energy systems has become a crucial approach for improving energy efficiency and reducing carbon emissions. However, most existing studies on ice thermal storage systems (ITSS) still focus on single indicators such as building cooling load or overall energy consumption, and there is a lack of systematic comparison of different machine learning (ML) and hybrid models for multi-objective parameters (MOPs) prediction of ITSS under real meteorological conditions. This gap limits the practical deployment of data-driven strategies for the refined operation and optimization of ITSS. In this study, operational data of MOPs for an ITSS were collected via a Building Automation System (BAS) platform and combined with meteorological parameters to construct a data-driven prediction framework. Various ML and hybrid models were applied to predict the operational data of ITSS MOPs, and their performances were comparatively analyzed. The results indicate that ensemble learning models generally achieve strong predictive performance for predicting the cooling capacity of the ITSS, while among deep learning models, TimesNet and recurrent neural network (RNN) exhibit superior results. When combined with a convolutional neural network (CNN), long short-term memory (LSTM), gated recurrent unit (GRU), and TimesNet exhibit significant improvements in predictive performance. Especially, CNN-GRU shows the most significant improvement compared with GRU alone, with the R value increasing from 0.7279 to 0.9042 (an enhancement of 24.22%, and the R2 value improving by 54.32%, from 0.5298 to 0.8176), whereas the improvement from the CNN-RNN combination is not significant. CNN-TimesNet achieves the best predictive performance, with an R value of 0.9078, representing an 11.17% improvement over TimesNet alone, while the R value of CNN-LSTM improves by 12.99%. Furthermore, CNN-TimesNet is employed to conduct predictive analysis on the MOPs of the ITSS. The results reveal that the total ice storage capacity parameter achieves the highest predictive accuracy, with an R value of 0.9725 (R2 of 0.9457). This study provides theoretical support and practical guidance for applying ML in the ITSS domain and for optimizing ITSS operational strategies in low-carbon buildings.

Employing timely short-term rainfall time series can alleviate the limitations of outdated meteorological information embedded in long-term datasets when assessing the performance of rainwater harvesting systems (RWHs) under climate change. However, indicators derived from short-term series often lack statistical stability across large samples, leading to uncertainty in projected outcomes. This study examines the influence of varying lengths of short-term series on RWH performance across 14 cities in Japan under climate change scenarios. A Bayesian network is then developed to capture the probabilistic features of these impacts, thereby identifying optimal rainfall conditions for RWH design under climate change. Results reveal that rainfall series longer than 16 years are unsuitable for RWH designing in cool temperate regions. In general, rainfall series with higher wet-period frequencies should be prioritized in inland cities, whereas the opposite trend is evident in coastal areas. Moreover, in northern regions the optimal series are characterized by longer dry periods, greater annual rainfall, and higher seasonal indices, while southern regions exhibit the reverse pattern. Validation confirms that the proposed Bayesian network reliably infers optimal short-term rainfall time series from statistical indictors, providing a robust framework for climate-adaptive RWH planning across diverse regional climates.