Search

• A pipeline is formulated to extract window to wall ratios from building façades. • A grammar based and learning based method are compared for performance efficacy. • The learning-based method shows higher accuracy than the grammar-based method. • A hybrid method combining the strengths of the methods achieves higher accuracy. • An algorithm utilizing the learning-based method extracts detailed building heights. Urban feature extraction methods have presented multiple opportunities in the field of environmental building performance studies, and urban assessments. With reference to façade features, window size and position are key parameters that influence occupant perception and environmental performance of indoor spaces. A myriad of methods have been proposed to automatically detect façade window layouts from street view images yet it is still unclear how to assess the strengths and limitations of these methods or how they can be combined for higher detection performance. This paper aims to add clarity for those aiming to use computer vision techniques for façade layout extraction for urban level sustainability assessments. An automated pipeline to enable the extraction and detection of WWRs (Window to Wall ratios) is introduced that is based on two fundamentally different computational approaches; a grammar-based edge detection framework (Method 1) and a learning-based method (Method 2) that utilizes CNNs (Convolutional Neural Networks). The paper then compares the detection efficacy of both methods, focusing on WWR accuracy, in New York and Lisbon, including their ability to extract more detailed façade properties such as floor-to-floor heights. The study finds that the learning-based method shows lower error scores across the two cities. In Lisbon, 69 % of conditions were detected within the 10 % error range and 91 % were within the 20 % range under Method 1. Under Method 2, 82.5 % of conditions were within the 10 % error range and 95 % were within the 20 % range. In New York, 66 % of conditions were within the 10 % error range and 90 % were within the 20 % range under Method 1 while 77.5 % of conditions were within the 10 % error range and 93 % were within the 20 % range under Method 2. Finally, a hybrid method is proposed to leverage the strengths of the two models, and higher accuracies are obtained in both the New York and Lisbon dataset. In New York, 96.5 % are now detected within the 20 % error range and 81.5 % within the 10 % error range. In Lisbon, 96 % are now detected within the 20 % error range and 83.5 % within the 10 % error range. With reference to the total building height extraction formulated under Method 2, the results show a relative error of 3.5 % in height estimation in the sample set. [ABSTRACT FROM AUTHOR]

Urban building energy modeling (UBEM) and its associated tools have proliferated in recent years, leading to various model development, simulation, calibration approaches, and use cases. UBEM is becoming a valid policy support tool to guide planning and intervention efforts at the neighborhood and city scale. However, current UBEM workflows focus primarily on the physical properties of buildings and geospatial geometry data without consideration for socioeconomic factors or demographic characteristics of the modeled/studied areas. This limitation impedes UBEM's effectiveness as a policy support tool. This paper presents a novel method – using a combination of supervised and unsupervised learning techniques on smart meter and census data – to develop hourly usage schedules for different socioeconomic personas. The schedules can be used to refine building archetypes for UBEMs. The method is piloted in two cities with similar building stocks but different socioeconomic compositions. It was found that socioeconomic indicators are important in classifying building archetypes. Considering these indicators changes the predicted city-wide energy use for residential buildings by up to 10%. The method presented is scalable and applicable to cities and municipalities worldwide (large or small) and elucidates the importance of accounting for demographic and socioeconomic indicators to reflect lived realities accurately. • A novel method to incorporate socioeconomic indicators in urban building energy modeling was developed. • Persona archetypes for UBEMs were created using smart meter data. • The method combined supervised and unsupervised learning techniques. • Simulation results were found to vary by approximately 10 %. • Feature importance could further identify important variables in urban building energy use. [ABSTRACT FROM AUTHOR]

The outdoor microclimate highly affects the quality of urban outdoor spaces which are essential to a city's socio-economic vitality. Due to the complexity associated with estimating outdoor thermal comfort (OTC), its study is currently only feasible at a high computational cost and time which makes it ineffective in any iterative design process. By coupling physics-based simulations and statistical modelling techniques, this paper presents a probabilistic, progressive, and accuracy-adaptive modelling approach for faster spatially-resolved OTC estimation that is scalable to large urban neighborhoods. This is achieved through three interrelated strategies: (1) Throughout the different simulation stages, spatiotemporal OTC categories are displayed with successively rising confidence levels to support instant design decision-making. (2) Confidence levels are based on probability distributions of partially known environmental variables such as wind or mean radiant temperature. (3) Wind distributions across an urban area are initially based on a spatially-informed set of rules and later replaced with explicit simulations of wind flow fields. The approach is tested against state-of-the-art computational fluid dynamic simulations for a 3 km2 sample area of San Francisco's financial district. Results show that scalable and actionable predictions are achievable at all simulation stages with the percentage of misclassified hourly OTC ranges during occupied hours falling from 36% for instant climate-based results to 8% and 7% for spatially clustered wind and building aerodynamics informed predictions which take minutes to calculate for the investigated urban area. The building aerodynamics informed simulations accurately predicts diurnal and seasonal OTC ranges for, on average, 97% of outdoor points. • A novel probabilistic, progressive, and accuracy-adaptive framework is presented. • Faster, scalable, and actionable outdoor thermal comfort predictions are achievable. • Integrating spatially informed wind distributions enhance accuracy. • Fit-for-purpose modelling enables abstractions at reduced error trade-off. • Spatially resolved wind flow insights are captured through flow aggregation strategy. [ABSTRACT FROM AUTHOR]

• A web-based framework to rapidly generate urban building energy models was proposed. • Key stakeholders and the workflow to translate open datasets to urban building energy models and actionable insights for carbon reduction were described. • The framework is scalable, able to construct urban building geometries, assign build simulation templates, and incorporate modern open-source machine learning libraries. • The framework was first tested in a pilot study, and subsequently validated with policymakers and researchers in eight cities around the world. Policymakers are struggling to understand what specific mixes of building retrofitting upgrades are necessary to achieve carbon emission targets. Urban building energy modeling (UBEM) is a bottom-up simulation method to develop policy measures for building stocks. However, the use of UBEM tools requires hard-to-find individuals with training in multiple domains and significant setup time exacerbated through a lack of standardized building use and construction databases. To address these challenges, this paper presents UBEM.io, a novel web-based framework to rapidly generate UBEMs in an automated and scalable fashion with clear handover points. UBEM.io is designed to help any city or municipality conduct low-cost, rapid energy and carbon emissions scenario studies. UBEM.io contributes to practice and research in multiple ways: automated generation of urban-scale building geometries based on widely available inputs; assignment of building simulation templates from a pre-populated library; matching of templates to individual buildings via archetypes; and visualization of simulation results for various carbon emissions reduction pathways. The framework was piloted in Evanston, IL (USA) to build an UBEM comprising 1,363 buildings. It was then successfully deployed with representatives from eight municipalities: Braga (Portugal), Cairo (Egypt), Dublin (Ireland), Florianopolis (Brazil), Kiel (Germany), Middlebury, VT (USA), Montreal (Canada), and Singapore. [ABSTRACT FROM AUTHOR]

The Universal Thermal Climate Index (UTCI) has been linked to outdoor activity patterns and used to evaluate the effectiveness of urban interventions to improve thermal comfort. This study investigates how simulating the urban environment at increasing levels of physical accuracy impacts UTCI values along three cycling routes in Cambridge, Massachusetts. Baseline UTCI values are estimated using a local weather file, and the following increments in physical accuracy are considered: wind-scaling, shading from buildings, shading and cooling from trees, computational fluid dynamics simulations for wind speeds, and simulated surface temperatures. With bike ridership data from Bluebikes, Boston's bike-sharing program, the relationship between bike ridership patterns and UTCI values along each route is studied. Supervised machine learning models are applied to predict bike ridership based on UTCI and other predictors. UTCI simulation results show that incorporating the various increments of accuracy influences hourly UTCI values at urban areas and exposed areas differently. Incorporating local wind speeds is especially impactful for urban areas. The statistical models trained to predict hourly bike trip counts based on UTCI and other demand and weather predictors achieved a root-mean-squared error of 1.06 trips. 47% of predictions were correct, and an additional 42% of predictions were off by 1 trip. This study demonstrates the importance of spatial refinement in simulating UTCI, and motivates future research into efficient simulation methods or rules-of-thumb for deriving spatial-temporal UTCI values. Future work into building a robust predictive model would motivate the design of thermally comfortable environments for human-powered transportation in cities. [Display omitted] [ABSTRACT FROM AUTHOR]

Urban building energy and daylight modeling are bottom-up, physics-based approaches to simulate the thermal and daylight performance of neighborhoods and cities. The field has flourished in recent years due to a wider accessibility of urban data sets which contain the required information regarding building geometry and program. However, key building-level parameters, most notably window-to-wall ratio (WWR), are generally unavailable at the urban scale and tedious to collect manually. To resolve this challenge, this paper proposes a methodology to automatically extract façade opening layouts for each building adjacent to a Google Street View route. A comparison between auto-generated and manually determined WWRs for 1057 buildings in Manhattan yielded identical results (less than 10% difference) for 66% of all investigated façades. Manual and automated methods were within a 20% error in 90% of all cases. The validated method is applied to daylighting and building energy models of 2014 buildings in downtown Chicago to quantify the impact of building-by-building WWRs versus a uniform, industry-standard WWR of 40% for all buildings. The results reveal that while the total energy use predictions are within 0.2% difference, the total daylit area increases by 9.5% when the WWRs are detected. Furthermore, when individual buildings are ranked in terms of their daylight autonomy or suitability for employing different retrofitting strategies, they are oftentimes misplaced when 40% WWR assumption is used. For example, in the downtown Chicago model, 46 buildings were misclassified as belonging to the top 100 buildings with the greatest percentage-wise savings potential resulting from glazing retrofitting. • A window detection workflow valid for energy and daylight modeling is presented. • Relevance of using detected window-to-wall ratios for urban modeling is assessed. • Using window detection for developing renovation strategies is proven important. • Assigning representative window-to-wall ratios to building types is demonstrated. • Window detection is crucial in applications using rankings of individual buildings. [ABSTRACT FROM AUTHOR]

• Exoskeletons are proposed as combined structural and solar shading elements. • We combine structural sizing with solar gains and operational energy simulations. • Reduced operational emissions of 24–48% through exterior shading are achieved. • Embodied carbon savings of up to 36% (steel) and 80% (timber) in the lateral system. We propose the use of solar exoskeletons, an integrated building system that combines material efficiency in structural load transfer with passive solar gain control. This offers an impactful way to respond to the UN climate goals, as the architecture and engineering disciplines face the challenge of delivering low carbon buildings. While reducing operational and embodied emissions is often considered independently, we can show how approaching them in tandem, through a novel building system, can offer significant savings. With large spans for maximum spatial flexibility and full glazing maximizing daylight, high-rise buildings are often suboptimal in terms of their material usage from steel frame construction and cooling demand from uncontrolled solar gains. We view solar exoskeletons as a sustainable pathway for future high-rise structures – combining solar gain control through external shading with a highly efficient structural system optimized for lateral loads in tall buildings. We present an automated workflow that combines parametric modeling of architectural elements and structural simulation with Radiance-based annual radiation simulations and an operational energy model in EnergyPlus. Evaluating embodied carbon and energy use intensity of midrise and tower buildings in timber and steel, we compare hundreds of iterations for a prototypical building in Phoenix, USA. Our results show that exoskeletons can lead to embodied and operational carbon reductions in the lateral load-resisting structural system of 37–80% and 24–48%, respectively, vis-à-vis conventional construction techniques. Adding photovoltaic modules to the external shading system can lead to net zero building solutions for the buildings investigated in this case study. [ABSTRACT FROM AUTHOR]

As the building stock is projected to double before the end of the half-century and the power grid is transitions to low-carbon resources, planning new construction hand in hand with the grid and its capacity is essential. This paper presents a method that combines urban building energy modeling and local planning of renewable energy sources (RES) using an optimization framework. The objective of this model is to minimize the investment and operational cost of meeting the energy needs of a group of buildings. The framework considers two urban-scale RES technologies, photovoltaic (PV) panels and small-scale wind turbines, alongside energy storage system (ESS) units that complement building demand in case of RES unavailability. The urban buildings are modeled abstractly as "shoeboxes" using the Urban Modeling Interface (umi) software. We tested the proposed framework on a real case study in a neighborhood in Chicago, Illinois, USA. The results include estimated building energy consumption, optimal capacity of the installed power supply resources, hourly operations, and corresponding energy costs for 2030. We also imposed different levels of CO2 emissions cuts. The results demonstrate that solar PV has the most prominent role in supplying local renewables to the neighborhood, with wind power making only a small contribution. Moreover, as we imposed different CO2 emissions caps, we found that ESS plays an increasingly important role at lower CO2 emissions levels. We can achieve a significant reduction in CO2 emissions with a limited increase in cost (75% emissions reduction at a 15% increase in overall energy costs). Overall, the results highlight the importance of modeling the interactions between building energy use and electricity system capacity expansion planning. • Optimization framework combines urban building energy modeling and local planning of renewables. • Urban neighborhood buildings are modeled in an abstracted manner as "shoeboxes". • Considering local wind and solar resources along with energy storage and grid supply. • PV has the most prominent role in supplying local renewables to the neighborhood. • Substantial reduction in CO2 emissions (75%) can be achieved with limited increase in energy costs (15%). [ABSTRACT FROM AUTHOR]