Crack Damage Assessment Based on Gaussian Process Model

Abstract:

The material dispersion of structure led to a lager error in crack length evaluating, the assessment method utilizing Gaussian Process (GP) model was proposed to solve the problem. The fatigue crack was monitored by active Lamb monitoring technology, and the four damage indices were extracted from the measured sensor signal, and then inputted to the GP model to realize the online evaluating of crack length. The fatigue test of hole-edge crack was made in LY12-CZ Aluminum specimen, which was used in aerospace structures frequently, the results shows that the method could efficiently decrease the evaluation error of crack length, which caused by material dispersion of structure.

• Periodical: Applied Mechanics and Materials (Volume 487)
• Pages: 247-254
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.487.247
• Online since: January 2014
• Authors: Jin Xin Shao*, Zi Xue Qiu, Jiang Yuan
• Keywords: Fatigue Crack, Gaussian Process, Lamb Wave, Length Evaluating, SHM

References

[1] SF. Yuan. Structural health monitoring and damage control, National defense press, Beijing, 2007.

[2] Y. Lu, L. Ye, Z. Q. Su, et al. Artificial neural network (ANN)-based crack identification in aluminum plates with lamb wave Signals, J. Journal of Intelligent Material Systems and Structures. 20 (2008) 39–49.

DOI: 10.1177/1045389x08088782

[3] A. Widodo, B. Yang. Support vector machine in machine condition monitoring and fault diagnosis, J. Mechanical Systems and Signal Processing. 21 (2007) 2560–2574.

DOI: 10.1016/j.ymssp.2006.12.007

[4] M. P. Prashant, N. J. Sung. Support vector machine based online composite helicopter rotor blade damage detection system, J. Journal of Intelligent Material Systems and Structures, 19 (200) 1217–1228.

DOI: 10.1177/1045389x07084713

[5] J. B. Ihn, F. K. Chang. Detection and monitoring of hidden fatigue crack growth using a built-in piezoelectric sensor/actuator network: I. Diagnostics, J. Smart Materials and Structures, 13 (2004) 609–620.

DOI: 10.1088/0964-1726/13/3/020

[6] J. B. Ihn, F. K. Chang. Pitch-catch active sensing methods in structural health monitoring, J. Structural Health Monitoring, 7 (2008) 5–19.

DOI: 10.1177/1475921707081979

[7] P. S. Tua, S. T. Quek, Q. Wang. Detection of crack in thin cylindrical pipes using Piezo-actuated Lamb waves, J. Smart Structures and Materials, 10 (2004) 820–831.

DOI: 10.1117/12.596074

[8] M. Gibbs. Bayesian Gaussian processes for classification and regression, University of Cambridge, Cambridge, (1997).

[9] C. E. Rasmussen, C. K. I. Williams. Gaussian processes for machine learning, The MIT Press, Cambridge, (2006).

[10] J. Cao, SF. Yuan, J. Cai, et al. Real-time structural health monitoring for fatigue crack growth, J. Piezoelectectrics & Acoustooptics, 30 (2008) 776–778.

[11] S. Mohanty, S. Das, P. Peralta. Adaptive residual useful life estimation of a structural hotspot, J. Journal of Intelligent Material Systems and Structures, 21 (2010) 321–335.

DOI: 10.1177/1045389x09357972

[12] Y. H. Lu, J. E. Michaels. Feature extraction and sensor fusion for ultrasonic structural health monitoring under changing environmental conditions, J. IEEE Sensors Journal, 9 (2009) 1462–1471.

DOI: 10.1109/jsen.2009.2019339

[13] J. E. Michaels, T. E. Michaels. Detection of structural damage from the local temporal coherence of diffuse ultrasonic signals, J. IEEE transactions on ultrasonics, ferroelectrics, and frequency control, 52 (2005) 1769–1782.

DOI: 10.1109/tuffc.2005.1561631

[14] Z. Q. Su, L. Ye. Identification of damage using lamb waves, The Springer Press, London, (2009).