A Novel PCA-GRNN Flow Pattern Identification Algorithm for Electrical Resistance Tomography System
Abstract
The two-phase flow measurement plays an increasingly important role in the real-time, on-line control of industrial processes including fault detection and system malfunction. Many experimental and theoretical researches have done in the field of tomography image reconstruction. However, the reconstruction process cost quite long time so that there are number of challenges in the real applications. An alternative approach to monitor two-phase flow inside a pipe/vessel is to take advantage of identification of flow regimes. This paper proposes a new identification method for common two phase flow using PCA feature extraction and GRNN classification based on electrical resistance system measurement. Simulation was carried out for typical flow regimes using the method. The results show its feasibility, and the results indicate that this method is fast in speed and can identify these flow regimes correctly.
Keywords
Electrical Resistance Tomography Flow regime identification Principal component analysis GRNNPreview
Unable to display preview. Download preview PDF.
References
- 1.Yang, X.D., Ning, X.B., Yin, Y.L.: Fingerprint image preprocessing technique and feature extraction algorithm. Journal of Nanjing University Natural Sciences 4(42), 351–361 (2006)Google Scholar
- 2.Areekul, V., Watchareeruetai, U.: Separable gabor filter realizationfor fast fingerprint enhancement. In: IEEE International Conference, vol. 3(1), pp. 5–10 (2005)Google Scholar
- 3.Karhumen, J., Joutsensalo, J.: Represention and Separation of Signals Using Nonlinear PCA Type Learning. Helsinki University of Technology 8, 1083–1091 (2001)Google Scholar
- 4.Ham, F.M., Kim, I.: Extension of the Generalized Hebbian Algorithm for Principal Component Extraction. Appliction and Science of Neural Networks 3455, 85–274 (1998)Google Scholar
- 5.da Rocha Gesualdi, A., Manoel de Seixas, J.: Characterrecognition in car license plates based on principal components and neural processing. Neural Networks 14, 206–211 (2002)Google Scholar
- 6.Yang, Z., Shi, Z.: Fast approach for optimal brain surgeon. Journal of System Simulation 17(1), 162–165 (2005)Google Scholar
- 7.Martinez, A.M., Kak, A.C.: PCA versus LDA. IEEE Trans., Pattern Anal., Machine Intelligent 23(2), 228–233 (2008)CrossRefGoogle Scholar
- 8.Wang, L., Wang, X., Feng, J.: In: On image matrix based feature extraction algorithms. IEEE Transactions on Systems, Man, and Cybernetics-Part B: Cybernetics: Accepted for Future Publication 99, 194–197 (2005)Google Scholar
- 9.Tao, J., Jiang, W.: Improved two-dimensional principal component analysis based on the palmprint identification. Optical Technology 33(2), 283–286 (2005)Google Scholar
- 10.Xing, X., Luo, W.: State estimation in power system based on GRNN algorithm. Heilongjiang Electric Power 33(1), 50–54 (2011)Google Scholar
- 11.Liu, Z., Zhou, T.: Application of GRNN Model Based on SSA in Financial Time Series 28(2), 29–33 (2011)Google Scholar
- 12.Wang, H., Lin, J.: Sales Predictio n Based on Improved GRNN 32(1), 153–157 (2010)Google Scholar