Komparativna analiza dubokih neuralnih mreža i algoritama gradijentnog pojačanja u dugoročnom predviđanju snage vetra

  • Luka Ivanovic Electrical Engineering Institute Nikola Tesla
  • Saša Milić Electrical Engineering Institute Nikola Tesla
  • Živko Sokolović Electrical Engineering Institute Nikola Tesla
  • Aleksandar Rakić The department of Signals and Systems School of Electrical Engineering
Ključne reči: Mašinsko učenje (ML), Rekurentne neuralne mreže (RNN), LSTM jedinica, GRU jedinica, GBM algoritam, XGBoost algoritam, Vetropark, Proizvodnja električne energije

Sažetak


Ključni korak ka održivoj budućnosti je integracija obnovljivih izvora energije u elektroenergetsku mrežu. Energija vetra je značajna zbog svoje široke dostupnosti i minimalnog uticaja na životnu sredinu. Ovaj rad predstavlja komparativnu analizu algoritama rekurentnih neuralnih mreža (RNN) i algoritama gradijentnog pojačanja (GBM) primenjenih na vremenske serije podataka za regresioni problem procene aktivne snage koju proizvodi vetropark (WF). GBM algoritmi kombinuju prednosti nekoliko modela mašinskog učenja (stabla odlučivanja, slučajne šume, itd.) kako bi proizveli moćan model predikcije. Pored postojećih konvencionalnih RNN, članak se bavi dugoročnim memorijskim jedinicama (LSTM) i rekurentnim jedinicama sa kapijama (GRU) kao najmodernijim modelima za predikciju vremenskih serija. Sveobuhvatna analiza je sprovedena na velikom skupu podataka o proizvodnji energije vetra.

Reference

Z. Tian, "A State-Of-The-Art Review on Wind Power Deterministic Predictio," Wind Engineering, vol. 45, pp. 1374–1392, 2021.

I. H. Sarker, "Deep Learning: A Comprehensive Overview on Techniques, Taxonomy, Applications and Research Directions," SN Computer Science, vol. 2, 420, 2021.

A. Zhang, Z. C. Lipton, M. Li, A. J. Smola, Dive into Deep Learning. 2023. [Online]. Available: https://d2l.ai/

P. J. Werbos, "Backpropagation through time: what it does and how to do it," Proceedings of the IEEE, vol. 78, no. 10, pp. 1550-1560, 1990.

I. Danish & Abbas, Asad. (2024). A Deep Dive into Neural Networks: Architectures, Training Techniques, and Practical Implementations. Journal of Environmental Sciences and Technology, vol. 02, pp. 61−71, 2023. https://doi.org/10.13140/RG.2.2.14866.84162

G. Van Houdt, C. Mosquera, G. Nápoles, "A Review on the Long Short-Term Memory Model," Artificial Intelligence Review, vol. 53, pp. 5929−5955, 2020.

Y. Freund, R. E. Schapire, "A Short Introduction to Boosting," Journal of Japanese Society for Artificial Intelligence, vol. 14, pp. 771−780, 1999.

J. H. Friedman, "Greedy function approximation: a Gradient Boosting machine," The Annals of Statistics, vol. 29, pp. 1189–1232, 2001.

T. Chen, C. Guestrin. "Xgboost: A scalable tree boosting system," in Proc. 22Nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD ’16, (New York, NY, USA, 13 Aug. 2016), p.p. 785–794.

A. Mobarak, A. Y. Owda, M. Owda, "Electrical Load Forecasting Using LSTM, GRU, and RNN Algorithms," Energies, vol. 16, 2283, 2023.

M.Chen, et al., "XGBoost-Based Algorithm Interpretation and Application on Post-Fault Transient Stability Status Prediction of Power System", IEEE Access, vol. 7, pp. 13149−13158, 2019.

Data available: https://www.kaggle.com/datasets/mubashirrahim/wind-power-generation-data-forecasting/data

J. Chung, et al., "Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling," 2015. [Online]. Available: https://doi.org/10.48550/arXiv.1412.3555

Objavljeno
2024/10/17
Rubrika
Originalni naučni članak