Trend Prediction of Hydraulic Liquid Leakage Based on ES-LSSVM

Abstract:

A novel prediction method which combined evolutionary strategy with least-square support vector machine is presented and applied to the trend prediction of hydraulic liquid leakage in this paper. In order to improve the prediction performance, the evolutionary strategy is employed to optimize the internal parameters of least-square support vector machine. Through the experiment study, the result validated the effectiveness of the prediction method, and it is also demonstrated that the method is able to do the short-term fault prediction for the hydraulic system.

• Periodical: Applied Mechanics and Materials (Volume 137)
• Pages: 440-444
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.137.440
• Online since: October 2011
• Authors: Zhi Yong Wu, Zeng Bing Xu, Ming Gao
• Keywords: Evolutionary Strategy, Hydraulic Liquid Leakage, Least Squares Support Vector Machines (Ls-SVM), Trend Prediction

References

[1] SUYKENS JAK, VANDEWALLE J. Least squares support vectors machine classifiers, Neural Processing Letters, 9(1999) 293-300.

[2] LEE Y J, HSIEH W F, HUANG C M. ε-SSVR: A smooth support vector machine forε-insensitive regression, IEEE Transactions on Knowledge and Data Engineering, 17(2005): 678-685.

DOI: 10.1109/tkde.2005.77

[3] Wang Keqi, Yang Shaochun, Method of optimizing parameter of least squares support vector machines by genetic algorithm. Computer Applications and Software, 26(2009) 109-111.

[4] Li Dazhong, Han Pu and Wang Zhen, The biomass gasification process modeling and optimization based on SVM and PSO. Journal of North China Electric Power University, 36(2009) 74-79.

[5] T. Bck. Evolutionary Algorithms in Theory and Practice. Oxford University Press, New York, (1996).