Autonomous Heat Loss Detection in Buildings for Efficient Façade Control and Energy Management

Insulation defects in building envelopes can lead to heat loss regions, resulting in energy wastage, moisture accumulation, and mold formation within the building fabric. In cold climates, where heating demand is high, timely detection of these defective areas is crucial for enhancing energy efficiency and improving occupants' well-being. This study introduces a method that integrates infrared thermography with YOLOv7 deep learning model to automatically identify and measure thermal anomalies in the building envelope. The method was tested on a multi-unit residential building and achieved a mean average precision ([email protected]) of 0.81 pinpointing significant heat loss areas. This approach optimizes real-time building inspections, enabling targeted energy-saving interventions while reducing the need for labor-intensive and expert-driven analysis. The study identified 28 thermal anomaly regions on the examined building façades, with the highest heat loss in the south façade, which accounted for 35% of the total detected anomalies. The proposed method offers a practical solution for improving energy management and façade control in buildings.