A Statistical Approach to Detect Abnormal Building Energy Consumption with ABCAT

The persistence of savings generated from commissioning is a significant topic of concern. The Automated Building Commissioning Analysis Tool (ABCAT) is a prototype tool that continuously monitors building energy consumption after commissioning, alerts operations personnel to significant increases in consumption, and assists them in identifying the problem. This paper introduces a statistical method using a Days Exceeding Threshold plot to detect abnormal whole-building energy consumption with ABCAT. The paper reviews the methodology of ABCAT, describes the application of the Days Exceeding Threshold plot in ABCAT, and then presents the results of a simulation test and a retrospective field test. In the simulation test, the Days Exceeding Threshold plot successfully identified eight abnormal energy consumption faults caused by five synthetic control changes. In the retrospective field test, 30 abnormal energy consumption faults were detected in 15 years of building energy consumption data. The causes of some of the detected faults are verified with historical documentation.