A non-intrusive load monitoring system (NILM) was installed in three commercial buildings, and
evaluated for its ability to disaggregate electric loads. The system largely failed to identify specific
loads, leading the authors to identify three key factors that make non-intrusive load identification
systematically difficult in mid-size commercial buildings, including the number and complexity of loads, difficulty in interpreting small changes in power consumption, and inability to identify continuously operating loads. Additionally, obtaining data sets for the evaluation of NILM technologies in actual buildings is hampered by disruptions to occupants, misidentification errors, measurement errors, and expense. Enhancements to basic NILM approaches include tagging key devices to facilitate
identification, hybrid or supplemental metering, and applying insights from engineering knowledge and audits.
A non-intrusive load monitoring system (NILM) was installed in three commercial buildings, and
evaluated for its ability to disaggregate electric loads. The system largely failed to identify specific
loads, leading the authors to identify three key factors that make non-intrusive load identification
systematically difficult in mid-size commercial buildings, including the number and complexity of loads, difficulty in interpreting small changes in power consumption, and inability to identify continuously operating loads. Additionally, obtaining data sets for the evaluation of NILM technologies in actual buildings is hampered by disruptions to occupants, misidentification errors, measurement errors, and expense. Enhancements to basic NILM approaches include tagging key devices to facilitate
identification, hybrid or supplemental metering, and applying insights from engineering knowledge and audits.