Neural Discriminating Analysis for a Nonintrusive Electrical Load Monitoring System

  • J. M. Seixas UFRJ
  • L. P. Calôba UFRJ
  • C. B. Prado UFRJ
  • J. C. R. Aguiar CEPEL

Resumo


A nonintrusive electrical load monitoring system for household appliances is developed using neural networks. Appliances are characterized by features extracted from their transient and steady-state responses obtained from sampling information from the AC power line. A discriminating analysis is applied as an efficient way to achieve a compact neural discriminator which identifies seven classes of equipment. Over 100 different pieces of equipment studied, the system classifies correctly more than 90% of the sample. The system is implemented on a 16 node transputer based parallel machine to support massive application. A processing time smaller than 2µs is achieved for each pattern.

Referências

Nonintrusive Identification of Residential Electrical Loads Using Neural Networks. J.C.R. Aguiar. Master Thesis, UFRJ, Brazil (1996).

J. Hertz, A. Krogh and R.G. Palmer. Introduction to the Theory of Neural Computation. Addison-Wesley (1991).

L. Lönnblad et al. Comp. Phys. Comm. 70, 167 (1992).

Neural Discriminating Analysis for a Second-Level Trigger System. L.P. Calôba et al. International Conference on Computing for High Energy Physics, Rio de Janeiro, Brazil (1995).

TN 310 System Manual and Training Set. Telmat Multinode, France (1995).

DSP/MSP Products Reference Manual. Analog Devices, Inc. (1995).

A Neural Nonintrusive Electrical Load ldentification Using a Parallel Machine. J.M. Seixas et al. Third International Conference on Engineering Applications of Neural Networks, Stockholm, Sweden (1997).
Publicado
07/10/1997
SEIXAS, J. M.; CALÔBA, L. P.; PRADO, C. B.; AGUIAR, J. C. R.. Neural Discriminating Analysis for a Nonintrusive Electrical Load Monitoring System. In: INTERNATIONAL SYMPOSIUM ON COMPUTER ARCHITECTURE AND HIGH PERFORMANCE COMPUTING (SBAC-PAD), 9. , 1997, Campos do Jordão/SP. Anais [...]. Porto Alegre: Sociedade Brasileira de Computação, 1997 . p. 201-210. DOI: https://doi.org/10.5753/sbac-pad.1997.22625.