No abstract

This work investigates the efficiency of the process of load disaggregation, considering only the values of active power. To perform the task, we use data collected from the NILM (Non-Intrusive Load Monitoring) measurement method, presented in the Rainforest Automation Energy Dataset (RAE) and Reference Energy Disagreggation Dataset (REDD) database. A strategy of assigning labels using combinations of equipment in use, by status ON/OFF, and also by choosing an appropriate temporal data window is discussed. Also, the performance of very well-known machine learning algorithms such as k-Nearest Neighbor (kNN), Decision Tree, and Random Forest are evaluated. The results show a very efficient and low computer complexity strategy presenting values of F1-Score above 95%, for RAE and REDD database. As presented in table 1, the proposed approach presents the highest F1-Score, compared to other methods in the literature, considering all appliances in the REDD database. The greatest benefit of the approach consists in the possibility of applying the disaggregation process in a household without smart outlets, under the restriction that the training and test houses hold identical or similar appliances.

In this Letter, the authors present a study on linear channel estimators and their respective mean square error expressions acknowledging spatially correlated channels and pilot contamination. They also investigate the impact of imperfect channel covariance matrix knowledge.

We are thankful to Companhia Paranaense de Energia (COPEL) for funding support in the current R&D project, under project ANEEL PD-02866-0508/2019.