Separation Study on Spectral Wear Signal for Engine


Article Preview

The monitoring signal of oil analysis is usually not from a single friction pair but from the whole friction system, which leads to insufficient pertinence of oil analysis. On the basis of proportional relationship between metal constituents of single friction pair namely wear source, constrained linear separation model was put forward to separate spectral wear detection signal to obtain single wear source signal. The application of the model to engine spectral wear detection information shows its availability, and the pertinence of oil analysis has been improved.



Edited by:

Han Zhao






Z. Y. Xu and Q. S. Xu, "Separation Study on Spectral Wear Signal for Engine", Applied Mechanics and Materials, Vols. 130-134, pp. 809-813, 2012

Online since:

October 2011




[1] YU Feng; YANG Cheng-wu. Application of Material-Units-Analysis Method to the Monitoring of Wearing Conditions in Engines Based on Oil Spectrum, Transactions of Csice, Volume 24, Issue 3, pp.281-283.

[2] Velten K, Reinicke R, Friedrich K. Wear Volume Prediction with Artificial Neural Networks , Tribology International, Volume 33, Issue 10, pp.731-736.

DOI: 10.1016/s0301-679x(00)00115-8

[3] ZHENG Chang-song; MA Biao. Improved Euler Algorithm for Trend Forecast Model and Its Application to Oil Spectrum Analysis, SPECTROSCOPY AND SPECTRAL ANALYSIS, Volume 29, Issue 4, pp.1078-1082.

[4] HyvarinenA, KarhunenJ, OjaE. Independent Component Analysis [M]. John Wiley & Sons, Inc., (2001).

[5] Xu Qisheng, Li Zhuguo. Recognition of wear mode using multivariable synthesis approach based on Wavelet packet and improved three-line method, Mechanical Systems and Signal Processing, Volume 21 Issue 8, pp.3146-3166.

DOI: 10.1016/j.ymssp.2007.04.008

[6] Xu Qisheng, Xu Zeyin, Xu Houchang. Study on Multi-feature Analysis of Engine Wear Based on Oil Denoised Data , China Mechanical Engineering,volume 21 Issue 12, pp.1405-1409.

DOI: 10.1109/icsps.2010.5555643

In order to see related information, you need to Login.