Abstract
A promising approach toward efficient energy management is non-intrusive loadmonitoring (NILM), that is to extract the consumption profiles of applianceswithin a residence by analyzing the aggregated consumption signal. Amongefficient NILM methods are event-based algorithms in which events of theaggregated signal are detected and classified in accordance with the appliancescausing them. The large number of appliances and the presence of applianceswith close consumption values are known to limit the performance of event-basedNILM methods. To tackle these challenges, one could enhance the feature spacewhich in turn results in extra hardware costs, installation complexity, andconcerns regarding the consumer's comfort and privacy. This has led to theemergence of an alternative approach, namely semi-intrusive load monitoring(SILM), where appliances are partitioned into blocks and the consumption ofeach block is monitored via separate power meters. While a greater number of meters can result in more accurate disaggregation,it increases the monetary cost of load monitoring, indicating a trade-off thatrepresents an important gap in this field. In this paper, we take acomprehensive approach to close this gap by establishing a so-called notion of"disaggregation difficulty metric (DDM)," which quantifies how difficult it isto monitor the events of any given group of appliances based on both theirpower values and the consumer's usage behavior. Thus, DDM in essence quantifieshow much is expected to be gained in terms of disaggregation accuracy of ageneric event-based algorithm by installing meters on the blocks of anypartition of the appliances. Experimental results based on the REDD datasetillustrate the practicality of the proposed approach in addressing theaforementioned trade-off.