Adaptive training of a vibration-based anomaly detector for wind turbine condition monitoring system (CMS) is carried out to achieve high-performance detection from the early stages of monitoring. Machine learning-based wind turbine CMSs are required to collect large-scale data to yield reli-able predictions. Existing studies in this area have postulated that both data for training a monitoring system and those during the operation of the system are obtained from identical devices. In addition, constant monitoring of data is desirable, but in practice, the data can be observed periodically (e.g., several tens of seconds of data are observed every two hours). In this case, collecting sufficient data is time consuming, making it difficult to conduct accurate predictions at the early stage of the CMS operation. To address this problem, a small amount of vibration data observed at a target wind turbine is utilized to adapt the anomaly detector that is trained on relatively large-scale vibration signals obtained from other wind turbines. In the present study, maximum a posteriori (MAP) adaptation is applied to a Gaussian mixture model (GMM)-based anomaly detector. Experimental comparisons using vibration data from the gearbox in the ex- perimental environment and those used in the wind turbine demonstrated that MAP-based GMM adaptation yielded an improvement in anomaly detection accuracy even when only a small amount of data is observed at the target gearbox.
anomaly detection, machine learning, Wind Turbine, Condition monitoring system, Gaussian Mixture Model, adaptive training
Bechhoefer, E. (2014). High speed gear dataset. (http://data-acoustics.com/measurements/gear-faults/gear-1/)
Camci, F., & Chinnam, R. B. (2010). Health-state estimation and prognostics in machining processes. IEEE Trans. on Automation Science and Engineering, 7(3), 581-597.
Dempster, A. P., Laird, N. M., & Rubin, D. B. (1977). Maximum likelihood from incomplete data via the EM algorithm. Journal of the Royal Statistical Society, Series B: Methodological, 39, 1–22.
Deutsch, J., & He, D. (2016, Oct.). Using deep learning based approaches for bearing remaining useful life prediction. In Proceedings of Annual Conference of the Prognostics and Health Management Society 2016.
Gauvain, J.-L., & Lee, C.-H. (1994). Maximum a posteriori estimation for multivariate gaussian mixture observations of Markov chains. IEEE Trans. on Speech and Audio Processing, 2(2), 291-298.
Gebraeel, N. Z., & Lawley, M. A. (2008). A neural network degradation model for computing and updating residual life distributions. IEEE Trans. on Automation Science and Engineering, 5(1), 154-163.
Goodenough, D. J., Rossmann, K., & Lusted, L. B. (1974). Radiographic applications of receiver operating characteristic (ROC) curves. Radiology, 110(1), 89-95.
Guclu, A., Yilboga, H., Eker, O. F., Camci, F., & Jennions, I. (2010). Prognostics with autoregressive moving average for railway turnouts. In Proceedings of Annual Conference of the Prognostics and Health Management Society 2010.
Hanley, A. J., & Mcneil, J. B. (1982). The meaning and use of the area under a receiver operating characteristic curve. Radiology, 143(1), 29-36.
Jolliffe, I. (2002). Principal component analysis. Wiley Online Library.
Kim, H. ., Tan, A. C. C., Mathew, J., & Choi, B. . (2012). Bearing fault prognosis based on health state probability estimation. Expert Systems with Applications, 39(5), 5200-5213.
Lee, C.-H., & Gauvain, J.-L. (1993, April). Speaker adaptation based on MAP estimation of HMM parameters. In Proceedings of International Conference on Acoustics, Speech and Signal Processing (Vol. II, p. 558-561).
Lee, C.-H., Juang, B.-H., & Lin, C.-H. (1991). A study on speaker adaptation of the parameters of continuous density hidden Markov models. IEEE Trans. on Signal Processing, 39(4), 806-814.
Liu, Z., Zuo, M. J., & Zhang, L. (2014). Remaining useful life prediction of rolling element bearings based on health state assessment. In Proceedings of the Annual Conference of the Prognostics and Health Management Society 2014 (p. 86-91).
Lusted, L. B. (1971). Decision-making studies in patient management. New England Journal of Medicine, 284(8), 416-424.
Medjaher, K., Tobon-Mejia, D. A., & Zerhouni, N. (2012). Remaining useful life estimation of critical components with application to bearings. IEEE Trans. on Reliability, 61(2), 292-302.
Metz, C. E. (1978). Basic principles of ROC analysis. Seminars in nuclear medicine, 8(4), 283-298.
Ogata, J., & Murakawa, M. (2016, Aug.). Vibration-based anomaly detection using FLAC features for wind turbine condition monitoring. In Proceedings of 8th European Workshop on Structural Health Monitoring (EWSHM) (Vol. 3, p. 1952-1961).
Sheng, S. (2014). Wind turbine gearbox condition monitoring vibration analysis benchmarking datasets.(http://en.openei.org/datasets/dataset/wind-turbine-gearbox
Stander, C. J. (2002). Using vibration monitoring for local fault detection on gears operating under fluctuating load conditions. Mechanical Systems and Signal Processing, 16(6), 1005-1024.
Wang, T. (2012, Sept.). Bearing life prediction based on vibration signals: A case study and lessons learned. In Proceedings of IEEE Conference on Prognostics and Health Management (PHM).
Wold, S., Esbensen, K., & Geladi, P. (1987). Principal component analysis. Chemometrics and intelligent laboratory systems, 2(1-3), 37–52.
Ye, J., Kobayashi, T., & Higuchi, T. (2010a, Sept.). Audiobased sports highlight detection by Fourier local autocorrelations. In Proceedings of the 11th Annual Conference of the International Speech Communication Association (INTERSPEECH 2010) (p. 2198-2201).
Ye, J., Kobayashi, T., & Higuchi, T. (2010b, Oct.). Vowel recognition based on FLAC acoustic features and subspace classifier. In Proceedings of International Conference on Signal Processing (ICSP) (p. 530-533).
Ye, J., Kobayashi, T., Murakawa, M., Higuchi, T., & Otsu, N. (2012). Anomaly detection using multi-channel FLAC for supporting diagnosis of ECG. IEEJ Trans. on Electronics, Information and Systems, 132(1), 111-119.