Stream Turbine Vibration Fault Diagnosis

Abstract:

RBF neural networks possessed the excellent characteristics such as insensitive on the initial weights and parameters with artificial fish-swarm algorithm (AFSA) applied, which made it have abilities to get rid of the local extremum and obtain the global extremum, and called as AFSA-RBF neural networks. In this paper, a new stream turbine vibration fault diagnosis method was presented based on AFSA-RBF neural networks. After quantification and reduction of the diagnosis decision table, the simplified decision table served as the learning samples of AFSA-RBF neural network, and the well-trained neural network was then applied to diagnose stream turbine vibration faults. The diagnosis results show that the proposed method possesses higher convergence speed and diagnosis precision, and is a very effective turbine fault diagnosis method.

• Periodical: Applied Mechanics and Materials (Volume 340)
• Pages: 90-94
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.340.90
• Online since: July 2013
• Authors: Hong Sheng Su
• Keywords: Artificial Fish-Swarm Algorithm (AFSA), Fault Diagnosis, RBF Neural Network, Stream Turbine

References

[1] Li, Xiaolei, Shao, Zhijiang, Qian, Jixin: An Optimizing Method Based on Autonomous Animates: Fish-Swarm Algorithm. Systems Engineering-Theory & Practice. 22(4), 32-38 (2002).

[2] Su, H. S: Main Converter Fault Diagnosis for Power Locomotive Based on PSO-BP Neural Networks. Advanced Materials Research. 267(12), 271-276 (2011).

DOI: 10.4028/www.scientific.net/amr.267.271

[3] Zeng, X., Liu, W: Particle Swarm Optimization Algorithm and Its Application in Neural Networks. Electric Drive Automation. 5(42), 17-19 (2009).

[4] Zhang, Bide, Li, Ming, Zhang, Gao: The Multi-layer Fuzzy Model for Tubro-Generator Faulty Diagnosis. Turbine Technology. 45(21), 25-28 (2003).

[5] Zhang, Xiao, Miao, Changxin: A Trouble Diagnosis Expert System for Stream-Turbine Based on Fuzzy Theory and Synthesis Analysis. Coal Engineerin. 12(7), 28-30 (2002).