An Analysis of Customer Power Profile Events for Non-invasive Energy Meter Error Monitoring

In this paper, a statistical analysis of low voltage customers' load power profiles is carried out aiming to suggest a remote non-invasive energy meter error monitoring technique. The synthesized open access 1 s resolution load power profiles are used for the analysis. A detection procedure for power steps in a sampled load power profile is defined. Power steps are classified into overlapping and non-overlapping. It is shown that the probability distribution function of the differences of power step magnitudes acquired by a meter under test and a reference sum meter is sensitive to the errors of the meter despite the presence of overlapping power steps at other meters in the grid. A noninvasive error monitoring technique is found possible by acquiring, comparing and statistically processing simultaneous power steps at the energy meter under monitoring and reference sum meter.