類神經網路在風機預兆式健康管理上的應用研究

Study on the Application of Neural Network in the Prognostic and Health Management of Wind Turbine

10.6342/NTU201603156

詹勳智

風力發電機 ； 類神經網路 ； 預兆式診斷及健康管理 ； 健康評估 ； 健康剩餘使用壽命 ； Wind Turbine ； Neural Network ； Prognostics and Health Management ； Health Assessment ； Health Remaining Useful Life

國立臺灣大學工程科學及海洋工程學系學位論文

2016年

碩士

蔡進發

繁體中文

本研究提出以倒傳遞類神經網路建立風機功率預測模型，並以預測值與實際值之誤差，定義風機健康評估指標，建立風機預兆式診斷及健康管理系統。本研究以林口發電廠4號風機為研究對象，利用4號風機於2012年至2015年完整4年之SACDA資料，實現風機健康之診斷。再利用Elman類神經網路進行風機健康評估指標衰退趨勢之預測，評估風機健康剩餘使用壽命。研究結果指出，定義健康評估指標值為0.15時，林口4號風機之健康剩餘使用壽命約為15年。 本研究藉由功率預測演算法及健康指標預測演算法，達成健康診斷及健康剩餘壽命預測兩部分，建立風機預兆式診斷及健康管理系統，做為風機維修之決策依據。

This research proposed a back-propagation neural network algorithm to establish Wind Power Forecasting model and implement the Health Assessment of wind turbine by Health Index which is defined using the error between forecast power and actually power. Based on wind turbine NO.4 in Linkou of Taipower, the system of prognostics and health management was set up with the data collected by the supervisory control and data acquisition(SCADA) system from 2012 to 2015. Then the Elman neural network was used to get the degradation of health index. Finally, health remaining useful life time of the wind turbine was predicted from the SCADA data. The analysis shows that the health remaining useful life time of the wind turbine NO.4 in Linkou of Taipower is about 15 years if the health index is defined as 0.15. The health assessment and health remaining useful life time of the wind turbine can be forecasted by the proposed neural network prognostics model and the criteria of health index. The developed prognostics and health management model can be used for wind turbine maintenance.

1. [3] B. Song and J. Lee, "Framework of designing an adaptive and multi-regime prognostics and health management for wind turbine reliability and efficiency improvement," Framework, vol. 4, pp. 142-149, 2013.
   連結：
2. [4] E. Lapira, D. Brisset, H. Davari, D. Siegel, and J. Lee, "Wind turbine performance assessment using multi-regime modeling approach," Renewable Energy, vol. 45, pp. 86-95, 2012.
   連結：
3. [5] M. Schlechtingen and I. Ferreira Santos, "Comparative analysis of neural network and regression based condition monitoring approaches for wind turbine fault detection," Mechanical Systems and Signal Processing, vol. 25, pp. 1849-1875, 2011.
   連結：
4. [7] R. Hecht-Nielsen, "Theory of the backpropagation neural network," in Neural Networks, 1989. IJCNN., International Joint Conference on, pp. 593-605, 1989.
   連結：
5. [8] Y. Yan, J. Li, and D. Gao, "Condition Parameter Modeling for Anomaly Detection in Wind Turbines," Energies, vol. 7, pp. 3104-3120, 2014.
   連結：
6. [9] A. Zaher, S. D. J. McArthur, D. G. Infield, and Y. Patel, "Online wind turbine fault detection through automated SCADA data analysis," Wind Energy, vol. 12, pp. 574-593, 2009.
   連結：
7. [10] K. Fukushima, "Neocognitron: A self-organizing neural network model for a mechanism of pattern recognition unaffected by shift in position," Biological Cybernetics vol. 36, pp. 193-202, 1980.
   連結：
8. [11] N. Kalchbrenner, E. Grefenstette, and P. Blunsom, "A convolutional neural network for modelling sentences," arXiv preprint arXiv:1404.2188, 2014.
   連結：
9. [12] K. Y. Chan, T. S. Dillon, J. Singh, and E. Chang, "Neural-network-based models for short-term traffic flow forecasting using a hybrid exponential smoothing and Levenberg–Marquardt algorithm," IEEE Transactions on Intelligent Transportation Systems, vol. 13, pp. 644-654, 2012.
   連結：
10. [13] M. D. Odom and R. Sharda, "A neural network model for bankruptcy prediction," in IJCNN International Joint Conference on neural networks, pp. 163-168,1990.
    連結：
11. [17] M. F. M?ller, "A scaled conjugate gradient algorithm for fast supervised learning," Neural networks, vol. 6, pp. 525-533, 1993.
    連結：
12. [18] A. Goh, "Back-propagation neural networks for modeling complex systems," Artificial Intelligence in Engineering, vol. 9, pp. 143-151, 1995.
    連結：
13. [19] M. Riedmiller and H. Braun, "A direct adaptive method for faster backpropagation learning: The RPROP algorithm," in Neural Networks, 1993., IEEE International Conference On, pp. 586-591, 1993.
    連結：
14. [22] X. Gao, X. Gao, and S. Ovaska, "A modified Elman neural network model with application to dynamical systems identification," in Systems, Man, and Cybernetics, IEEE International Conference on, pp. 1376-1381,1996.
    連結：
15. [23] F. O. Heimes, "Recurrent neural networks for remaining useful life estimation," in Prognostics and Health Management, PHM 2008. International Conference on, pp. 1-6, 2008.
    連結：
16. [24] M. T. Hagan and M. B. Menhaj, "Training feedforward networks with the Marquardt algorithm," IEEE transactions on Neural Networks, vol. 5, pp. 989-993, 1994.
    連結：
17. [1] 經濟部能源局. 千架海陸風力機風力資訊整合平台. Available: http://www.twtpo.org.tw/
18. [2] D. He, E. Bechhoefer, and A. Saxena, "Special Issue on Wind Turbine Prognostics and Health Management," International Journal of Prognostics and Health Management, vol. 4, 2013.
19. [6] O. Uluyol, G. Parthasarathy, W. Foslien, and K. Kim, "Power curve analytic for wind turbine performance monitoring and prognostics," in Annual conference of the prognostics and health management society, 2011.
20. [14] M. Lydia, S. S. Kumar, A. I. Selvakumar, and G. E. Prem Kumar, "A comprehensive review on wind turbine power curve modeling techniques," Renewable and Sustainable Energy Reviews, vol. 30, pp. 452-460, 2014.
21. [15] R. F. M. Brandão, J. A. B. Carvalho, and F. P. M. Barbosa, "Neural networks for condition monitoring of wind turbines gearbox," Journal of Energy and Power Engineering, vol. 6, 2012.
22. [16] M. T. Hagan, H. B. Demuth, M. H. Beale, and O. De Jesús, Neural network design vol. 20: PWS publishing company Boston, 1996.
23. [20] 羅華強 and 通信工程, 類神經網路: MATLAB 的應用: 高立, 2011.
24. [21] R. J. Williams, "Adaptive state representation and estimation using recurrent connectionist networks," Neural networks for control, vol. 30, pp. 97-114, 1990.

1. 謝佩鈞(2017)。相似分群方法在風場風機故障檢測的應用研究。國立臺灣大學工程科學及海洋工程學系學位論文。2017。1-124。