Publication

Non-intrusive load monitoring based on event detection and unsupervised learning for airport baggage handling systems

Conference Article

Conference

International Conference on Soft Computing Models in Industrial and Environmental Applications (SOCO)

Edition

16th

Pages

567-577

Doc link

https://doi.org/10.1007/978-3-030-87869-6_54

File

Download the digital copy of the doc pdf document

Authors

Projects associated

Abstract

With a non-intrusive load monitoring paradigm, this paper poses the first steps to monitor the health of airport baggage handling systems. This goal is reached by measuring the energy consumption of the electrical cabinets that power a set of conveyor belt systems. Therefore, using an energy disaggregation approach, each motor in the conveyor system can be monitored. The proposed methodology consists of an algorithm to detect power-on/off events and how those events can be clustered by characterizing transient states, employing unsupervised clustering algorithms. Energy measurements were filtered to remove noise, and the on-off events were detected and characterized. The power-on and -off events were clustered, using k-means and Gaussian mixture model (GMM), which the latter can properly discriminate events into different groups. Thus, the results obtained with GMM are presented and analyzed. From each cluster important insight are extracted in terms of energy consumption.

Categories

pattern clustering, pattern recognition.

Scientific reference

M.A. Bermeo, David A. Cruz R., J. Diaz and C. Ocampo-Martínez. Non-intrusive load monitoring based on event detection and unsupervised learning for airport baggage handling systems, 16th International Conference on Soft Computing Models in Industrial and Environmental Applications, 2021, Bilbao, Spain, pp. 567-577, Springer.