Global climate challenges often manifest as localized discomfort within buildings, including overheating, glare, and excessive energy consumption. Conventional solar control systems—such as large awnings or mechanical louvers—tend to be rigid, visually dominant, and disconnected from occupants’ experiences. This project proposes an alternative approach through a distributed system of small-scale shading devices composed of knitted textiles, 3D-printed scissor mechanisms, and Arduino-controlled servos.

Rather than relying on large monolithic interventions, the system operates through multiple modular shading units positioned across a window surface. Each unit combines the elasticity of knitted textiles with lightweight scissor mechanisms fabricated through additive manufacturing. The knitted membranes expand and contract seamlessly, while micro-controlled servos actuate the mechanisms in response to environmental sensors or user input. Together, these components enable dynamic adjustments to solar angles, local microclimates, and occupant preferences throughout the day.

Parametric design tools guide the development of textile patterns with localized elasticity, allowing the membranes to form volumetric geometries when deployed and collapse softly when retracted. This patterning produces shading elements that move beyond flat surfaces, introducing tactile and decorative qualities that contribute to the spatial experience of interior environments.

Within broader discussions of architecture’s role in climate action, this research explores how softness, adaptability, and distributed small-scale interventions can support environmental performance. By reducing reliance on mechanical cooling and encouraging passive solar control, the system promotes both thermal comfort and occupant agency. Its modular structure and accessible materials allow adaptation across diverse climates and building types.

By integrating textile craft, digital fabrication, and responsive technologies, the project demonstrates how small architectural devices can contribute to broader conversations on sustainability and climate-responsive design.

This interdisciplinary research investigates the spatial relationship between Urban Heat Island (UHI) effects and the potential for photovoltaic (PV) solar energy generation on the North Carolina State University (NC State) campus in Raleigh, North Carolina. As urban campuses face rising energy demands and climate challenges, understanding where to integrate renewable infrastructure is critical (Santamouris,2015). This project combines geospatial analysis, remote sensing, and energy simulation to identify locations with elevated surface temperatures and high solar potential,key indicators for PV deployment.

Satellite-derived thermal imagery and land surface temperature (LST) data were used in ArcGIS Pro to map UHI intensity and locate urban "hotspots" (Voogt&Oke, 2003). These were overlaid with solar irradiance maps generated using NREL’s PVWatts Calculator (Dobos,2014) to model PV energy potential on rooftops and open spaces. Cooling degree days and building-level energy consumption were integrated to assess the link between thermal stress and energy demands, especially for buildings in high-UHI zones.

Results show strong spatial correlations between high surface temperature areas and high solar irradiance, highlighting optimal zones for PV installation. Buildings most affected by UHI also exhibited elevated cooling loads during peak summer months, emphasizing the need for renewable energy offsets. Comparative analysis between denser, developed areas and more vegetated or open zones demonstrated that urban buildings experience stronger UHI effects while offering greater solar energy capture potential due to reduced shading and higher irradiance.

This study contributes to campus sustainability by providing a replicable, data-driven framework for evaluating UHI intensity and identifying solar-ready zones through integrated geospatial and energy modeling (Zhou et al.,2011). Findings support NC State University’s sustainability and climate resilience initiatives by guiding strategic investment in renewable infrastructure and demonstrating how co-analysis of environmental stressors and energy potential can inform equitable, integrated campus planning.

This research was supported by the Sustainable Futures Fellowship at NC StateUniversity.

Solar photovoltaic (PV) generation now lies at the heart of the most decarbonization plans. Policy and decision makers need more than energy forecasts, like clear estimates of the annual economic value of each plant in particular market context. This study examined a harmonized solar-to-grid dataset from the U.S. Department of Energy for distributed photovoltaic systems. The study applies machine learning methods to model and interpret the economic value of distributed photovoltaic systems using two value-based performance metrics: Energy_Value_MWh and Total_Value_MWh. By framing PV systems as socio-technical infrastructure, the analysis integrates physical generation characteristics with temporal and regional market context. Three regression models: Support Vector Regression (SVR), Random Forest, and Extreme Gradient Boosting (XGBoost) are evaluated using operational, temporal, and spatial features, including annual generation, installed capacity, year, balancing authority, and geographic identifiers. Model performance is assessed through coefficient of determination (R²), root mean square error (RMSE) and predicted-versus-actual visual analysis.

Results indicate that ensemble tree-based models substantially outperform SVR across both targets. Random Forest achieves the strongest overall performance, with R² values exceeding 0.88 and consistently lower prediction errors, while XGBoost demonstrates comparable accuracy with slightly higher dispersion at extreme values. Feature importance analysis reveals that temporal evolution and regional grid structures dominate PV economic outcomes, whereas physical generation alone plays a secondary role. These findings highlight the importance of contextual and system-level factors in PV valuation and demonstrate the potential of machine learning as a planning-oriented tool for evaluating renewable energy performance within architectural and urban energy systems. The proposed framework supports more informed decision-making in PV design, deployment, and policy evaluation across diverse built environments.