An Empirical Investigation of V-I Trajectory based Load Signatures for Non-Intrusive Load Monitoring

摘要

Choice of load signature or feature space is one of the most fundamentaldesign choices for non-intrusive load monitoring or energy disaggregationproblem. Electrical power quantities, harmonic load characteristics, canonicaltransient and steady-state waveforms are some of the typical choices of loadsignature or load signature basis for current research addressing applianceclassification and prediction. This paper expands and evaluates appliance loadsignatures based on V-I trajectory - the mutual locus of instantaneous voltageand current waveforms - for precision and robustness of prediction inclassification algorithms used to disaggregate residential overall energy useand predict constituent appliance profiles. We also demonstrate the use ofvariants of differential evolution as a novel strategy for selection of optimalload models in context of energy disaggregation. A publicly available benchmarkdataset REDD is employed for evaluation purposes. Our experimental evaluationsindicate that these load signatures, in conjunction with a number of popularclassification algorithms, offer better or generally comparable overallprecision of prediction, robustness and reliability against dynamic, noisy andhighly similar load signatures with reference to electrical power quantitiesand harmonic content. Herein, wave-shape features are found to be an effectivenew basis of classification and prediction for semi-automated energydisaggregation and monitoring.