Non-Intrusive Identification of Loads by Random Forest and Fireworks Optimization

The control of expenses related to electricity has been showing significant growth, especially in residential environments. Monitoring of electrical loads that are turning on and off from home are often performed using smart plugs, providing to the consumers' information about operation intervals and power consumed by each device. Despite a practical solution to control and reduce electricity costs, it has a high cost due to the number of meters required. The high-cost problem can be worked around by using a non-intrusive load monitoring proposal (NILM), where voltage and current measurements are taken at the home entrance, in counterpart demand an extra processing step. In this extra step, various actions like computation of powers, identification of the occurrence of events, and identification of the status of which equipment (on/ off) must be made. The purpose of this work was the elaboration of a heuristic type event detector using floating analysis windows for locating stability zones on power signals after indicating a power change above a predetermined value. For that, tests of the best arrangement of event identifier data to identify which load has been added or removed from the monitored circuit are made. The proposed hybrid approach optimizes the processes using the Fireworks Algorithm (FWA) were used in the Random Forest classifier to improve classification performance. The proposed event identifier and classifier tests were performed on the dataset BLUED, which contains data collected at a North-American residence over one week. The event identifier results were compared with other publications that used different approaches, and the results of the classifications were compared to each other, using various data entry forms, and as an ideal classifier.