Sign in Subscribe
Energy management

Improving Sustainability through Advanced Chiller Power Prediction

Strategic Insights

— 1 min read

Researchers from the University of Malaysia Pahang have developed a novel approach to predicting chiller power consumption in commercial buildings. This breakthrough has significant implications for global sustainability efforts, as chillers are major energy consumers in commercial buildings. The team developed an Evolutionary Mating Algorithm (EMA) hybridized with Artificial Neural Networks (ANN) to optimize feature selection and enhance prediction accuracy. Their model demonstrated superior performance compared to other metaheuristic-ANN hybrid models, with a 38.3% reduction in Root Mean Square Error (RMSE) and a 6.0% improvement in coefficient of determination (R2).

Key Takeaways:

  • The research developed a novel hybrid model, EMA-ANN, to predict chiller power consumption, which demonstrated superior performance compared to other metaheuristic-ANN hybrid models.
  • The EMA-ANN approach identified seven optimal features primarily comprising temperature and humidity parameters, contributing to a 38.3% reduction in RMSE and a 6.0% improvement in R2.
  • The algorithm's unique evolutionary mating mechanism with adaptive crossover rate (Cr = 0.85) enabled effective feature space exploration.
  • The study concluded that the EMA-ANN approach can be utilized as a benchmark for evaluating feature selection algorithms in building energy applications.
  • The research was financially supported by the Ministry of Education, Malaysia.
  • The study was peer-reviewed and published in the Journal of Building Engineering.
  • The EMA-ANN approach can be applied to various building energy applications, including commercial and residential buildings.

Statistics:

  • RMSE reduction: 38.3%
  • R2 improvement: 6.0%
  • Mean Absolute Error (MAE): 0.2235
  • Root Mean Square Error (RMSE): 0.4150
  • Coefficient of determination (R2): 0.9689

Sources:

  • Feature Optimization With Metaheuristics for Artificial Neural Network-based Chiller Power Prediction. Journal of Building Engineering, 2025;105.
  • University of Malaysia Pahang, Department of Electrical and Electronic Engineering Technology
  • Ministry of Education, Malaysia
  • Journal of Building Engineering, Elsevier, Radarweg 29, 1043 Nx Amsterdam, Netherlands

Read more