美国建筑业目前消耗了美国一次能源供应的40%以上。据估计，由于照明和舒适调节系统的病态故障，任何建筑的年能耗中有5%到30%在不知不觉中被浪费掉。本文介绍了包含在有用的软件工具中的分析方法，以快速识别和评估导致大型建筑能源效率低下的选定建筑系统故障。这项工作的技术贡献包括适用于暖通空调设备规模和操作的专家规则，以及根据用户定义的兴趣（如能源低效的年度成本）对故障信号进行排序的方法。 这些贡献在处理模型和在故障评估中仔细考虑用户兴趣方面尤其独特。作为开发此故障检测通用框架的第一步，针对一阶故障，例如多个大型空气处理装置内的同时加热和冷却以及不平衡气流。当一台空气处理机在一个多月内同时加热和冷却的故障发生时，预计可节省约22500美元。这是一个大型医院节能潜力的例子。通过在空气中修正循环泵的操作，预计每年可节省约24000美元- 处理器热回收回路。用户测试和实验表明，在暖通空调故障检测和评估中包含不确定性不仅对明智的故障推断至关重要，而且对于获得系统用户的信任和认可也至关重要，这些用户最终可以应用故障检测信息来实际修复和改进建筑操作。引用:2018年年度会议，德克萨斯州休斯顿，会议论文

The building sector of the U.S. currently consumes over 40% of the U.S. primary energy supply. Estimates suggest that between 5 and 30% ofany building's annual energy consumption is unknowingly wasted due to pathologically malfunctioning lighting and comfort conditioning systems.This paper presents analytical methods embodied within useful software tools to quickly identify and evaluate selected building system faults thatcause large building energy inefficiencies. The technical contributions of this work include expert rules that adapt to HVAC equipment scale andoperation and methods for sorting fault signals according to user- defined interests such as annual cost of energy inefficiencies. These contributionsare particularly unique in their treatment of models and the careful consideration of user interests in fault evaluation. As a first step to developingthis general framework for fault detection, first-order faults such as simultaneous heating and cooling and imbalanced airflows within several largeair-handling units were targeted. Savings of around $22,500 were predicted when the fault of simultaneous heating and cooling occurred over anentire month in an air handler. An example of the potential energy savings in a large hospital that has been monitored. Yearly savings of around$24,000 were predicted by correcting the operation of the circulation pump in an air-handler heat recovery loop. User testing and experimentsshow that embracing uncertainty within HVAC fault detection and evaluation is not only paramount to judicious fault inference but it is alsocentral to gaining the trust and buy-in of system users who ultimately can apply fault detection information to actually fix and improve buildingoperations.