Achieving citywide building energy reduction goals require extensive understanding of energy use at scale, which is challenging due to scarce and disparate data. Despite attention to urban building energy models (UBEM), unexplored aspects and missing details in this emerging field have remained, including further exploring non-homogenous commercial buildings, providing a detailed structure to create UBEMs for replication purposes, and developing methods to mitigate data scarcity and dependency. In this study, a structure is proposed using commercial buildings in Pittsburgh, Pennsylvania. We provide a description of an archetype library with relevant sources to improve reproducibility and describe a novel framework to create a database focusing on façade reconstruction through photogrammetry and image processing. For our UBEM, twenty archetypes that comprised eight commercial use types were identified. The average annual energy use intensity was estimated between 74 and 1302 kWh/m² for different use types. The simulation results also showed discrepancy in energy use of the buildings with similar use types. Validating the results utilizing actual data revealed an overall 7% error. Employing the model to evaluate energy conservation strategies showed energy use reduction of 2–5% for the entire stock. Outcomes of this research can aid policy makers in instituting energy goals and efficiency regulations.

实现全市建筑节能目标需要对大规模的能源使用有广泛的了解，由于数据稀缺和不同，这具有挑战性。尽管关注城市建筑能源模型 (UBEM)，但这个新兴领域中未探索的方面和缺失的细节仍然存在，包括进一步探索非同质商业建筑，提供详细结构以创建用于复制目的的 UBEM，以及开发减轻数据稀缺性的方法和依赖。在这项研究中，提出了一种使用宾夕法尼亚州匹兹堡商业建筑的结构。我们提供了具有相关来源的原型库的描述，以提高可重复性，并描述了一个新颖的框架，以创建一个专注于通过摄影测量和图像处理进行立面重建的数据库。对于我们的 UBEM，确定了包含八种商业用途类型的二十种原型。年平均能源使用强度估计在 74 至 1302 kWh/m 之间²用于不同的使用类型。模拟结果还显示了具有相似使用类型的建筑物在能源使用方面的差异。使用实际数据验证结果显示总体误差为 7%。使用该模型来评估节能策略表明，整个库存的能源使用量减少了 2-5%。这项研究的结果可以帮助政策制定者制定能源目标和效率法规。