食品零售建筑的能源需求约占英国能源消耗和由此产生的碳排放的3%。之前的研究（Spyrou等人，2014年，Tassou等人，2011年）表明，食品零售建筑的电力需求中最大的组成部分是食品制冷系统的冷却需求（从30%到50%不等）。因此，更好地了解制冷用电需求将有助于开发有效的能源管理工具，包括评估服务和维护干预措施，以减少运营用电需求。过去，为了快速识别商用制冷系统运行期间的故障，已经开发并采用了各种方法。传统上，这些方法的重点是车间食物的温度。 这项工作的目的是通过识别导致制冷系统电力需求增加的事件，增强全球多渠道零售组织采用的现有故障查找方法。本文提出了一种分析制冷系统数据的方法，使故障识别更加直观。这包括耗电量、压缩机运行时间、接收器中制冷剂的百分比、蒸发器上下空气的温度、排放和吸入压力等数据。还收集和分析了每个站点的控制策略和维护计划以及气象数据。采用数据挖掘方法去除已知的运行模式（例如除霜周期）和季节变化。突出显示了对系统耗电量产生影响的事件，并过滤掉了现有方法识别的故障。 然后进一步分析生成的数据集，以了解增加系统电力需求的事件，从而创建自动识别方法。引用:ASHRAE论文:2015年ASHRAE年会，伊利诺伊州芝加哥

The energy demands of food retail buildings account for approximately 3% of the UK’s energy consumption and resultant carbon emissions. Previous studies (Spyrou et al. 2014, Tassou et al. 2011) demonstrate that the greatest component of the electricity demand of food retail buildings is the cooling demand of the food refrigeration systems (ranging from 30 to 50%). Therefore a better understanding of the electricity demand for refrigeration would enable the development of effective energy management tools, including the evaluation of service and maintenance interventions to reduce operational electricity demand. Various methodologies have been developed and employed in the past for the quick identification of faults during the operation of commercial refrigeration systems. The focus of these methodologies has traditionally been on the temperature of food on the shop floor. The aim of this work is to enhance the existing fault-finding methodologies employed by a global multichannel retail organization, by enabling the identification of events that cause an increase in electricity demand of the refrigeration systems. This paper presents a methodology that analyzes data from refrigeration systems and enables a more straightforward identification of faults. This includes data for electricity consumption, compressor run times, percentage of refrigerant in the receiver, temperature of air on and off the evaporator, discharge and suction pressures, etc. Control strategies and maintenance schedules as well as meteorological data for each site were also collected and analyzed. Data mining methods were employed to remove known operational patterns (e.g. defrost cycles) and seasonal variations. Events that have had an effect on the electricity consumption of the system were highlighted and faults that have been identified by the existing methodology were filtered out. The resulting data set was then analyzed further to understand the events that increase the electricity demand of the systems in order to create an automatic identification method.