As sustainability becomes a priority across the building industry, innovative tools like 3D modeling and Building Information Modelling (BIM) allow architects to experiment with various design configurations and materials that reduce environmental impact. These new tools empower designers to work collaboratively in a virtual environment, eliminating the need for physical prototypes and increasing accuracy and efficiency throughout the process.
However, incorporating sustainability into the design process requires more than just changing materials and configurations. It also involves understanding a building’s operational energy performance and life cycle, as well as evaluating trade-offs between expected energy savings and carbon emissions.
3D Modeling for Energy Audits: Visualizing Efficiency Before You Build
To address these challenges, Georgia Insulation explores a new way to characterize and represent as-built conditions of existing buildings for retrofit assessment purposes using 3D spatially-registered thermal and digital imagery. This method automatically analyzes and interprets a collection of inexpensive and often already-existing 2D images and combines them into a single virtual space with geometrical and thermal information. This approach provides a low-cost, reliable, and automated solution for as-built modeling that can support proactive building diagnostics, identify deviations between as-built and intended energy performance data, create much-needed feedback loops with design predictions, and enable more efficient energy retrofits.
This system utilizes machine learning techniques on novel data sources, such as drone images, to automate the building 3D modeling procedure, and to improve the effectiveness of the identification and quantification of energy savings opportunities. It identifies the geometry of the building by a combination of line extraction, polygonization, and 3D reconstruction algorithms; detects windows and wall-to-window ratios with state-of-the-art deep neural network semantic segmentation; and identifies rooftop energy equipment with an unsupervised clustering algorithm. Energy demand is directly compared to reference consumption either from similar buildings or generated by simulation tools, allowing for the direct identification of excessive and unacceptable energy demands.
Georgia Insulation
2092 Crow Rd, Gainesville, GA 30501
(770)758-4459
