Non-intrusive Load Identification Algorithm Based on Stacking Modeling

Abstract The non-intrusive power load identification relies on the smart meter’s voltage, current and signals, which can realize classification monitoring of domestic electric load. algorithm is core system. Non-intrusive intelligent systems are generally based a single model for aggregation, commonly used models include hidden Markov models, graphical deep learning models. Usually, only effective in identifying certain types electrical loads, its universality generality need to be improved. Improving decomposition accuracy main purpose implementing powerload monitoring. stacking integration optimize model. divided into two layers. base initially decomposes realizes through meta-model. Experiment results show that Stacking modeling proposed this paper accurately decompose minute-level sampled data. Compared model, method improved by 8.2%.