Power Transformer Fault Diagnosis Based on Improved Bat Algorithms to Optimize RNN

  • Chun Yan
  • Meixuan Li
  • Wei LiuEmail author
Conference paper
Part of the Advances in Intelligent Systems and Computing book series (AISC, volume 1074)


In order to improve the accuracy of transformer fault diagnosis, this paper presents a fault diagnosis method based on Bat algorithm to optimize randomized neural network. Because the initial connection weights and hidden neuron biases of the randomized neural network are generated randomly, the network results are easily unstable. Therefore, based on the excellent performance of at algorithm in function optimization, this paper uses the improved Bat algorithm to optimize the weight of randomized neural network. In order to overcome the shortcomings of early convergence and poor results of Bat algorithm, the inertia weight of particle swarm optimization (PSO) is introduced into Bat algorithm, and the velocity update formula is improved. In order to improve the diversity of population and the local search ability of the Bat algorithm, the chaotic local search method is added to balance the global search ability of the Bat algorithm in the early stage and the local search ability in the later stage. In order to control the range of bat’s later position, the empirical factor is introduced into the position change formula to accelerate the convergence speed of the algorithm. Finally, the empirical results show that the proposed fault identification model can better diagnose various types of transformer faults.


Bat algorithm Randomized neural network Transformer fault diagnosis 


  1. 1.
    Liang, Y., Li, K., Zhao, J., et al.: Study on dynamic adjustment strategy of transformer oil chromatography on-line monitoring cycle. Proc. Chin. J. Electr. Eng. 34(09), 1446–1453 (2014)Google Scholar
  2. 2.
    Zheng, R., Zhao, J., Zhao, T., et al.: Fault diagnosis of power transformer based on genetic support vector machine and grey artificial immune algorithm. In: CSEE, vol. 31, no. 07, pp. 56–63 (2011)Google Scholar
  3. 3.
    Seifeddine, S., Khmais, B., Abdelkader, C.: Power transformer fault diagnosis based on dissolved gas analysis by artificial neural network. In: First International Conference on Renewable Energies & Vehicular Technology, no. 03, pp. 230–236. IEEE (2012)Google Scholar
  4. 4.
    Siddique, M.A.A., Mehfuz, S.: Artificial neural neworks based incipient fault diagnosis for power transformers. In: India Conference, pp. 1–6. IEEE (2016)Google Scholar
  5. 5.
    Long, Q., Guo, S., Li, Q., et al.: Research of converter transformer fault diagnosis based on improved PSO-BP algorithm, vol. 231, p. 012015 (2017)Google Scholar
  6. 6.
    Setiawan, N.A., Sarjiya, Adhiarga, Z.: Power transformer incipient faults diagnosis using dissolved gas analysis and rough set. In: International Conference on Condition Monitoring and Diagnosis, Bali, Indonesia, pp. 950–953. IEEE (2012)Google Scholar
  7. 7.
    Aizpurua, J.I., Catterson, V.M., Stewart, B.G., et al.: Power transformer dissolved gas analysis through Bayesian networks and hypothesis testing. IEEE Trans. Dielectr. Electr. Insul. 25(2), 494–506 (2018)CrossRefGoogle Scholar
  8. 8.
    Wang, L., Zhang, L., Pu, J., et al.: Application of improved BP neural network in object recognitio. Electron. Opt. Control 19(4), 68–71 (2012)Google Scholar
  9. 9.
    Wang, K., Li, J., Zhang, S., et al.: New characteristic parameters of dissolved gas in oil for transformer fault diagnosis. Proc. Chin. J. Electr. Eng. 36(23), 6570–6578 + 6625 (2016)Google Scholar
  10. 10.
    Yang, X.: A new metaheuristic bat-inspired algorithm. In: Computer Knowledge & Technology, pp. 65–74 (2010)Google Scholar
  11. 11.
    Song, Y.: Research on logistics center location problem based on improved bat algorithm, Henan University (2017)Google Scholar
  12. 12.
    Zhang, X., Hu, Z., Tang, T.: Application of improved bat algorithm in Fuzzy Analytic hierarchy process. Appl. Comput. Syst. 27(03), 143–148 (2018)Google Scholar
  13. 13.
    He, Z.: Improvement and application of bat algorithm, Guangdong University of Technology (2016)Google Scholar
  14. 14.
    Wang, W.: Research on the improvement of cuckoo search algorithm and bat optimization algorithm, Guangxi University for Nationalities (2014)Google Scholar
  15. 15.
    Schmidt, W.F., Kraaijveld, M.A., Duin, R.P.W.: Feedforward neural networks with random weights. In: Proceedings of 11th IAPR International Conference on Pattern Recognition, Conference B: Pattern Recognition Methodology and Systems, vol. 2, pp. 1–4 (1992)Google Scholar
  16. 16.
    Pao, Y.H., Park, G.H., Sobajic, D.J.: Learning and generalization characteristics of random vector functional-link net. Neurocomputing 6, 163–180 (1994)CrossRefGoogle Scholar
  17. 17.
    Suganthan, P.N.: On non-iterative learning algorithms with closed-form solution. Appl. Soft Comput. 70, 1078–1082 (2018)CrossRefGoogle Scholar
  18. 18.
    Cheng, J.T., Duan, Z.M., Xiong, Y., et al.: Research on QPSO-BP model transformer fault diagnosis method based on DGA. High Volt. Appl. 52(02), 57–61 (2016)Google Scholar

Copyright information

© Springer Nature Switzerland AG 2020

Authors and Affiliations

  1. 1.College of Mathematics and System ScienceShandong University of Science and TechnologyQingdaoChina
  2. 2.College of Computer Science and EngineeringShandong University of Science and TechnologyQingdaoChina

Personalised recommendations