Advanced nonintrusive load monitor for equipment diagnostics on fan coil unit

Fand Coil Units (FCUs) are widely used in commercial buildings to provide heating and cooling; and are therefore important for equipment and energy diagnostics in buildings. Currently, there is no low-cost and high-performance diagnostics method for FCUs. This hinders the deployment of equipment and energy diagnostics for building automation systems and smart building services. In this paper, we propose a novel approach which integrates two independent systems - an advanced nonintrusive load monitoring (NILM+) system and a statistical model-based diagnostics system into a single FCU diagnostic concept. NILM+ provides critical information enabling FCU equipment diagnostic models to detect fan and coil valve faults within 5 minutes. The diagnostic algorithms yields false alarm rates less than 8% and missed detect rates less than 10%.