Sign in Subscribe
Housing

Machine Learning Enhances Rural Housing Energy Retrofits in China

Strategic Insights

— 2 min read

Researchers at Anhui Jianzhu University in Hefei, People's Republic of China, have developed a novel framework that integrates uncertainty analysis and machine learning to increase the accuracy and efficiency of rural housing envelope retrofits. The framework, which focuses on key envelope components such as exterior walls, roofs, and windows, has been shown to enhance the reliability and generalizability of energy consumption predictions. By incorporating uncertainty-informed datasets and machine learning, the researchers aim to provide a scalable approach for optimizing rural housing energy retrofits.

Key Takeaways:

  • The researchers developed a framework that integrates uncertainty analysis and machine learning to improve the accuracy and efficiency of rural housing envelope retrofits.
  • The framework focuses on key envelope components such as exterior walls, roofs, and windows.
  • The study used reference models established via field surveys and monitoring data, and multiple energy datasets were generated via uncertainty analysis.
  • Sensitivity analysis identified key factors such as cooling and heating setpoint temperatures, infiltration, and exterior wall construction and roof construction.
  • These factors were used as inputs for machine learning models built with neural networks and random forests.
  • The coverage ratio and evenness were used to assess residual distributions, and results indicated that a more uniform residual distribution within the 95% interval balances data volume and prediction accuracy.
  • The study demonstrated that uncertainty-informed datasets and machine learning enhance the reliability and generalizability of energy consumption predictions.
  • The researchers successfully implemented the framework in a case study in Jiaxian, China, and the results show a scalable approach for optimizing rural housing energy retrofits.

Statistics:

  • 75%, 90%, 95%, and 100% intervals were used to train machine learning models.
  • Sensitivity analysis employed standardized regression coefficients, random forests, and the treed gaussian process.
  • The coverage ratio and evenness were used to assess residual distributions. Results indicate that a more uniform residual distribution within the 95% interval balances data volume and prediction accuracy.
  • The study demonstrates that uncertainty-informed datasets and machine learning enhance the reliability and generalizability of energy consumption predictions.

Sources:

  • "Machine Learning-based Energy Consumption Models for Rural Housing Envelope Retrofits Incorporating Uncertainty: a Case Study In Jiaxian, China." Case Studies in Thermal Engineering, 2025;72.
  • Zao Li, Taoyuan Zhang, Zihuan Zhang, Xia Sun, Yulu Chen. "Machine Learning-based Energy Consumption Models for Rural Housing Envelope Retrofits Incorporating Uncertainty: a Case Study In Jiaxian, China." Case Studies in Thermal Engineering, 2025;72.
  • National Natural Science Foundation of China (NSFC)
  • Anhui Province University Collaborative Innovation Project
  • Elsevier, Radarweg 29, 1043 Nx Amsterdam, Netherlands
  • Zao Li, Anhui Jianzhu University, Hefei 230601, People's Republic of China

Read more