Proxy zone-level energy use estimation in a commercial building with a variable air volume system

Abstract The usefulness of zone level sensors is contentious in the context of commercial building controls. Few building owners are willing to increase capital costs with conflicting information about the most valuable sensors to owners and operators. Few studies have systematically explored this topic in a practical manner. This study compared zone-level sensors of 27 rooms of an institutional building for their ability to estimate energy use indirectly using a novel methodology. Notable sensors in this study include hydronic flow meters, room level lighting and plug load current transducers, window contacts, and ceiling illuminance at three interior depths. Data was collected at 5-min sample intervals over four months total. Sensors were scored and ranked by energy use using mean bias error to infer estimation accuracy over the validation period (half the measurement period) with artificial neural network models. A hierarchy of sensor value for each energy end use was generated. It was found that outdoor air temperature sensors provide the most usefulness for energy use estimation in the building studied (where heating and cooling loads dominated due to the high outdoor air supply ratio). Hydronic flow meter data was also useful for energy use estimation.