Optimization of Processing Parameters for Micro Arc Oxidation Based on Orthogonal Design and Support Vector Machine Regression Analysis

Abstract:

During processing, the interaction among the multi-parameters which influences the coating surface roughness is very complex. In order to gain rapidly the best parameters, this paper raises the method to optimize parameters of the micro arc oxidation based on the orthogonal design and the support vector machine regression analysis. The experiments were performed for the coating surface roughness according to the parameters designed by orthogonal on LD10. And then, the support vector machines regression was used to obtain the model between the surface roughness and the parameters according to these data. Further, the model optimized the parameters and predicted the corresponding coating with Ra1.025μm. At last, the model was verified by the experiments of single factor method under the same condition as the orthogonal experiments. The results, comparative analysis of the surface roughness of predicting and actual values generated by the same parameters, shows that the square error and the ratio of the average error influenced by the parameters expect for the temperature is less than 0.1 and 10% respectively, and the actual coating with Ra1.199μm was obtained that the parameters optimized by the model treated.

• Periodical: Advanced Materials Research (Volumes 156-157)
• Pages: 307-310
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.156-157.307
• Online since: October 2010
• Authors: Ming Qiang Pan
• Keywords: Micro-Arc Oxidation (MAO), Orthogonal Design, Support Vector Machine Regression Analysis

References

[1] Zhang Yichen, Xu Hui, Yang Yingfu. Proceedings of the 8th Vacuum Metallurgy and Surface Engineering Conference. P. 33.

[2] Li Wenfang, Huang Jinghao, et al. Material Engineering. 2(2006), P. 51(In Chinese).

[3] Guo Ping, Li Wenfang, et al. New Technology and New Process. 2(2006), P. 51(In Chinese).

[4] Yan Fenyun, Lin Hua, Wang Sheng. New Technology & New Process. 5(2008), P. 107.

[5] Shawkat Ali, Kate A. Amith-Miles, Neurocomputing, 70(2006), P. 173.

[6] D. S. Broomhead, D. Lowe, Complex Systems, 2(1988), P. 321.

[7] Roman Rosipal, Leonard J. Trejo, Journal of Machine Learning Research, 2(2001), P. 97.

[8] Wenjian Wang, Zongben Xuan, Weizhen Lu, et al, Neurocomputing, 55(2003),P. 643.