高斯混合模型在風機預兆式健康管理上的應用研究

Study on the Application of Gaussian Mixture Model in the Prognostic and Health Management of Wind Turbine

10.6342/NTU201602401

楊其昌

風力發電機 ； 高斯混合模型 ； 預兆式健康管理 ； wind turbine ； Gaussian mixture model ； Prognostic and health management

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

本研究以高斯混合模型(Gaussian Mixture Model)為核心建立一套風機預兆式健康診斷與預測的方法，此方法包含：採用DBSCAN(Density-Based Spatial Clustering of Applications with Noise)對風機原始參數進行過濾，高斯混合模型建立風機營運性能模型，以信心值來表示風機的健康狀態，再利用迴歸分析來預測風機未來營運的健康信心值。 本研究利用所建立的預兆式診斷方法對台電林口四號風機的資料進行分析，分析的結果顯示，此風機在一般正常運作的情況下，健康狀況信心值約在0.4至0.8之間，但風機營運出現了異常狀況時，信心值大多低於0.4。但若就長時間的信心值變化而言，此風機在2013年至2015年這三年間是呈現穩定的狀態，即代表此風機在這三年間並無太大的性能衰退。另外，透過迴歸分析計算，可預測出此風機於2033年8月8日後，整體的健康狀況信心值將會下降到2013年平均值的兩個標準差以下，代表風機的性能可能在該時間衰退至不健康的狀態。

A prognostic and health management method based on the Gaussian mixture model is proposed in this study to analyze and predict the performance of wind turbine. The proposed method includes preprocessing the raw data of wind turbines by DBCSAN (Density-Based Spatial Clustering of Applications with Noise), building the model on operating performance of wind turbines by GMM (Gaussian Mixture Model), indicating the operating performance by the CV (Confidence Value), and predicting the CV in the future by regression analysis. The proposed method was applied to analyze the performance data of the Wind Turbine No.4 of Taiwan Power Company at Linkou District. The analysis showed that the CV is between 0.4 and 0.8 in the normal condition and is smaller than 0.4 in the abnormal condition. The CV of the wind turbine is stable between 2013 and 2015. That is, the performance of this wind turbine was not degrading obviously. Furthermore, by regression analysis, the trend of CV will reduce to 0.66 which is out of 2 standard deviations below the mean of 2013 on August 8, 2033. It means that the wind turbine may be probably unhealthy at that moment.

1. [3] J. Lee, M. Ghaffari, and S. Elmeligy, "Self-maintenance and engineering immune systems: Towards smarter machines and manufacturing systems," Elsevier Ltd., 2011.
   連結：
2. [4] B. L. Song and J. Lee, "Framework of Designing an Adaptive and MultiRegime Prognostics and Health Management for Wind Turbine Reliability and Efficiency Improvement," International Journal of Advanced Computer Science and Applications, Vol. 4, No. 2, 2013.
   連結：
3. [6] E. Lapira, D. Brisset, H. D. Ardakani, D. Siegel, and J. Lee, "Wind turbine performance assessment using multi-regime modeling approach," Renewable Energy, 2012.
   連結：
4. [7] Y. Feng, Y. Qiu, C. J. Crabtree, H. Long, and P. J. Tavner, "Use of SCADA and CMS Signals for Failure Detection and Diagnosis of a Wind Turbine Gearbox," Proceedings of European wind energy conference 2011.
   連結：
5. [13] D. Reynolds, "Gaussian Mixture Models," 2009.
   連結：
6. [14] D. Reynolds, T. F. Quatieri, and R. B. Dunn, "Speaker Verification Using Adapted Gaussian Mixture Models," 2000.
   連結：
7. [15] M. H. Yang and N. Ahuja, "Gaussian Mixture Model for Human Skin Color and Its Applications in Image and Video Databases," Beckman Institute and Department of Electrical and Computer Engineering, University of Illinois at Urbana-Champaign, Urbana, 1999.
   連結：
8. [16] C. B. Do and S. Batzoglou, "What is the expectation maximization algorithm?," Nat Biotechnol. , 2008.
   連結：
9. [17] P. A. Bromiley, "Products and Convolutions of Gaussian Probability Density Functions," Tina Memo Report, Tech. Rep. , 2003.
   連結：
10. [20] C. C. Holt, "Forecasting seasonals and trends by exponentially weighted moving averages," Elsevier B.V. on behalf of International Institute of Forecasters, 2004.
    連結：
11. [1] 經濟部能源局, "2014年能源產業技術白皮書 第肆章 新及再生能源," 2014.
12. [2] 經濟部能源局, "2014年能源產業技術白皮書 第壹章 緒論," 2014.
13. [5] S. A. Niknam and D. R. Sawhney, "A Framework for Prognostic-based Real-time Life Extension," The University of Tennessee, Knoxville. http://web.utk.edu/~rmc/documents/sawhney_paper.pdf
14. [8] K. Kim, G. Parthasarathy, O. Uluyol, W. Foslien, S. Sheng, and P. Fleming, "Use of SCADA Data for Failure Detection in Wind Turbines," the 2011 Energy Sustainability Conference and Fuel Cell Conference, 2011.
15. [9] C. S. Gray, F. Langmayr, N. Haselgruber, and S. J. Watson, "A Practical Approach to the Use of SCADA Data for Optimized Wind Turbine Condition Based Maintenance," EWEA conference, 2011.
16. [10] O. Uluyol, G. Parthasarathy, W. Foslien, and K. Kim, "Power Curve Analytic for Wind Turbine Performance Monitoring and Prognostics," Annual conference of the prognostics and health management society 2011.
17. [11] L. A. Osadciw, Y. Yan, X. Ye, G. Benson, and E. White, "Wind Turbine Diagnostics based on Power Curve Using Particle Swarm Optimization," Syracuse University, Department of Electrical Engineering and Computer Science, Syracuse, NY, 2010.
18. [12] G. A. Skrimpas, C. W. Sweeney, K. S. Marhadi, B. B. Jensen, N. Mijatovic, and J. Holb?ll, "Detection of Wind Turbine Power Performance Abnormalities Using Eigenvalue Analysis," Proceedings of the 2014 Annual Conference of the Prognostics and Health Management Society, 2014.
19. [18] M. Ester, H. P. Kriegel, J. Sander, and X. Xu, "A Density-Based Algorithm for Discovering Clusters in Large Spatial Databases with Noise," Proceedings of 2nd International Conference on Knowledge Discovery and Data Mining, 1996.
20. [19] E. W. Weisstein, "Moore-Penrose Matrix Inverse," 2002.

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。