Skip to main content
Log in

Influence of ANN parameters on the performance of a refined procedure to solve the load-flow problem

Einfluss von ANN-Parametern auf die Eigenschaften eines verfeinerten Verfahrens zur Lösung des Leistflusses

  • Reviewed Original Papers
  • Published:
e&i Elektrotechnik und Informationstechnik Aims and scope Submit manuscript

Abstract

In recent years, interest in the application of Artificial Neural Networks (ANN) to electrical power systems has grown rapidly. In particular the use of ANN in the solution of the load-flow problem in wide electrical networks is an interesting research topic, because it constitutes a good alternative to the classical numerical algorithms. In this paper a refined solution strategy based on statistical methods, on a particular Grouping Genetic Algorithm (GGA) and on Progressive Learning Network (PLN) is presented. Tests on the solution of load-flow equations of the standard IEEE 118 bus network confirm the good potential of this approach; in particular the search for optimal values of the PLN’s parameters, with the aim at reducing the number of processing units, reducing further the calculation time, it’s possible to make the procedure presented extremely competitive.

Zusammenfassung

In den letzten Jahren ist das Interesse an künstlichen neuralen Netzwerken (Artificial Neural Networks — ANN) rasch gewachsen. Insbesondere ist die Verwendung von ANN bei der Lösung des Flussproblems bei weiterreichenden elektrischen Netzwerken ein interessantes Thema, da es eine gute Alternative zu klassischen nummerischen Algorithmen darstellt.

Dieser Beitrag stellt eine auf statistischen Methoden beruhende erweiterte Lösung dar sowie eine auf einem besonderen gruppierenden genetischen Algorithmus (Grouping Genetic Algorithms — GGA) und einem progressiven Lern-Netzwerk (Progressive Learning Network — PLN). Prüfungen auf Grund der Lösung von Leistflussgleichungen des Standard-IEEE-118-Bus-Netzwerks bestätigen die guten Möglichkeiten dieser Verfahrensweise; insbesondere ermöglicht die Suche nach optimalen Werten von PLN-Parametern, die auf die Verringerung von Verfahrenseinheiten und damit der Rechenzeit abzielen, eine extreme Konkurrenzfähigkeit des vorgelegten Verfahrens.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Similar content being viewed by others

References

  1. Stott, B.: Review of load-flow calculation methods. IEEE Proceedings Power Systems 62 (1974).

  2. Silvestri, A.: Analisi della sicurezza statica dei sistemi elettrici di potenza. L’Energia Elettrica 70 (1993) No. 11, pp. 5–16.

    Google Scholar 

  3. Stott, B., Alsac, O.: Fast decoupled load-flow. IEEE Trans. Power Appl. System Pas-93, (May–June 1974).

  4. Zecevic, A. I., Siljak, D. D.: A block-parallel Newton method via overlapping epsilon decompositions. SIAM Journal on Matrix Algebra and Application (1994).

  5. Amano, M., Zecevic, A. I., Siljak, D. D.: An improved block parallel Newton method via epsilon decompositions for load flow calculations. IEEE Transactions on Power Systems 11 (1996), No. 3.

  6. Wassermann, P.: Neural Computing. New York: Van Nostrand Reinhold. 1989.

    Google Scholar 

  7. Piglione, F., Cirrincione, G., Cirrincione, M.: Load flow approssimato mediante reti neurali. Rendiconti 95a Riunione AEI 5 (1994), pp. 131–139.

    Google Scholar 

  8. Cirrincione, G., Cirrincione, M., Piglione, F.: Applicazione di una rete neurale ad apprendimento progressivo allo studio dei sistemi dipotenza. Rendiconti 96a Riunione AEI 1995, pp. 93–100.

  9. Cuccia, P., Piglione, F.: Impiego di una rete neurale nella stima delle grandezze nodali di una rete AT in condizioni di avaria del sistema di acquisizione dati. Rendiconti 96a Riunione AEI 1995, pp. 101–108.

  10. Cataliotti, V., Filizzola, V.: Applicazioni delle reti neurali per la valutazione del load-flow in un sistema di potenza. AEI 83 (1996), No. 12, pp. 43–49.

    Google Scholar 

  11. Augugliaro, A., Cataliotti, V., Dusonchet, L., Favuzza, S., Scaccianoce, G.: Progressive learning networks applied to solve wide electrical power systems. Proceedings of the International ICSC/IFAC Symposium on neural Computation NC’98 — Technical University of Vienna. 1998, p. 335–341.

  12. Falkenauer, E.: A new representation and operators for gas applied to grouping problems. Evolutionary Computation 2 (1994).

Download references

Author information

Authors and Affiliations

Authors

Rights and permissions

Reprints and permissions

About this article

Cite this article

Augugliaro, A., Cataliotti, V., Dusonchet, L. et al. Influence of ANN parameters on the performance of a refined procedure to solve the load-flow problem. Elektrotech. Inftech. 116, 348–353 (1999). https://doi.org/10.1007/BF03159194

Download citation

  • Issue Date:

  • DOI: https://doi.org/10.1007/BF03159194

Keywords

Schlüsselwörter

Navigation