The application discloses an iterative unsupervised non-intrusive load identification method and
system, and the method comprises the following steps: obtaining a load power
time sequence; pre-
processing the load power
time sequence; detecting key points of the load power
time sequence; extracting a load power subsequence and matching the load power subsequence with a template; judging whether the number of unknown load power subsequences accumulates to 30; performing theme discovery on the load power subsequence; judging whether the number of subsequences in each group in a non-directional graph is not less than 10; performing electrical appliance power waveform theme sequence mode mining; constructing corresponding electrical appliance load mark templates, and updating a load mark
template library of a non-intrusive load
monitoring system; and the
system comprises corresponding
processing modules. The application can autonomously discover simple and complex
power consumption modes of different types of electrical appliances in various unfamiliar scenes, and is suitable for the diversity of
power consumption modes of the same type of electrical appliances and the difference of
power consumption modes of the same electrical appliance.