Scholarlink

<p>The growing emphasis on sustainable building operation and the rising cost of energy have made the improvement of heating, ventilation, and air conditioning system efficiency a central challenge in modern engineering. Despite considerable progress in data-driven prediction methods, existing studies often lack hybrid modeling frameworks capable of accurately capturing the complex and nonlinear interactions among environmental variables, operational conditions, and energy usage. To address this knowledge gap, the present study employs a machine learning approach based on Adaptive Boosting Regression integrated with two advanced evolutionary optimization algorithms, Giant Armadillo Optimization and Gradient-Based Optimizer, to develop high-accuracy models for forecasting the energy performance of heating, ventilation, and air conditioning systems. The study utilizes a comprehensive dataset containing environmental, operational, and energy-related variables collected from a real building environment. The results demonstrate a substantial improvement in predictive accuracy compared with traditional regression approaches. The best-performing hybrid model achieved a determination coefficient of 0.997 and reduced root mean square prediction error by more than 75% relative to the baseline method. The model also maintained consistently high accuracy in training, validation, and testing phases, confirming its strong generalizability. Sensitivity analyses further revealed that humidity-related variables exert the strongest influence on system behavior.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Enhancing heating, ventilation, and air conditioning system efficiency through renewable energy integration and smart feature analysis

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年1月16日

查看全文
  • Huijiao Nie

摘要

The growing emphasis on sustainable building operation and the rising cost of energy have made the improvement of heating, ventilation, and air conditioning system efficiency a central challenge in modern engineering. Despite considerable progress in data-driven prediction methods, existing studies often lack hybrid modeling frameworks capable of accurately capturing the complex and nonlinear interactions among environmental variables, operational conditions, and energy usage. To address this knowledge gap, the present study employs a machine learning approach based on Adaptive Boosting Regression integrated with two advanced evolutionary optimization algorithms, Giant Armadillo Optimization and Gradient-Based Optimizer, to develop high-accuracy models for forecasting the energy performance of heating, ventilation, and air conditioning systems. The study utilizes a comprehensive dataset containing environmental, operational, and energy-related variables collected from a real building environment. The results demonstrate a substantial improvement in predictive accuracy compared with traditional regression approaches. The best-performing hybrid model achieved a determination coefficient of 0.997 and reduced root mean square prediction error by more than 75% relative to the baseline method. The model also maintained consistently high accuracy in training, validation, and testing phases, confirming its strong generalizability. Sensitivity analyses further revealed that humidity-related variables exert the strongest influence on system behavior.