Scholarlink

<p>Reliable climate classification is essential for evaluating building components that strongly impact heating and cooling energy demand. Among building components, windows play a significant role in thermal exchange and are commonly assessed through climate-dependent energy labeling programs. In Iran, the current classification system based on ISO 18292:2011 divides the country into ten climatic zones relying on temperature and solar radiation. However, this classification does not fully reflect Iran’s climatic diversity, potentially misrepresenting window performance and energy consumption. This study proposes an adaptive approach incorporating local climate data to define suitable window properties, improve energy efficiency, and reduce greenhouse gas emissions while supporting the United Nations Sustainable Development Goal (SDG) 11 for resilient urban development. To develop and validate the framework, ten years of historical observed meteorological data from 158 stations of the Iran Meteorological Organization were analyzed based on degree-days and solar irradiance conversion to vertical surfaces. Two unsupervised clustering algorithms, K-means and hierarchical clustering, were adopted to define climatic zones. Ward’s hierarchical method identified 11 clusters validated by the Silhouette Coefficient, while EnergyPlus simulations showed improved annual window heating and cooling performance under the proposed zoning. Kernel interpolation with barriers was utilized to generate a thorough geographical representation of climatic diversity across 468 counties. The proposed approach demonstrates how machine learning and spatial analysis can be integrated to improve climate zoning. The resulting climate map provides a more reliable basis for climate-responsive window design and future updates of national energy performance standards.</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

Adaptive climate zoning for sustainable energy performance of window systems: A comparative study of unsupervised clustering algorithms

  • 期刊论文

  • 发表日期：

    2026年3月28日

查看全文
  • Mojgan Moradi

摘要

Reliable climate classification is essential for evaluating building components that strongly impact heating and cooling energy demand. Among building components, windows play a significant role in thermal exchange and are commonly assessed through climate-dependent energy labeling programs. In Iran, the current classification system based on ISO 18292:2011 divides the country into ten climatic zones relying on temperature and solar radiation. However, this classification does not fully reflect Iran’s climatic diversity, potentially misrepresenting window performance and energy consumption. This study proposes an adaptive approach incorporating local climate data to define suitable window properties, improve energy efficiency, and reduce greenhouse gas emissions while supporting the United Nations Sustainable Development Goal (SDG) 11 for resilient urban development. To develop and validate the framework, ten years of historical observed meteorological data from 158 stations of the Iran Meteorological Organization were analyzed based on degree-days and solar irradiance conversion to vertical surfaces. Two unsupervised clustering algorithms, K-means and hierarchical clustering, were adopted to define climatic zones. Ward’s hierarchical method identified 11 clusters validated by the Silhouette Coefficient, while EnergyPlus simulations showed improved annual window heating and cooling performance under the proposed zoning. Kernel interpolation with barriers was utilized to generate a thorough geographical representation of climatic diversity across 468 counties. The proposed approach demonstrates how machine learning and spatial analysis can be integrated to improve climate zoning. The resulting climate map provides a more reliable basis for climate-responsive window design and future updates of national energy performance standards.