Genetic Algorithm-Based COD Calibration Method Research

Abstract:

For better realization of the function of Chemical oxygen demand (COD) online measuring instrument and improving its measurement accuracy , a good calibration and identification of signals collected is needed. During the process, the problem on parameter identification of undetermined function can be transformed into function optimization. Considering the characteristics of genetic algorithm It is introduced into the function identification of the measuring system and compare it with the radial basis function neural network. As for the premature of population evolutionary process, this article presents the method to select operators according to genetic fitness value of each individual and designs a set of system identifier based on Genetic Algorithm to identify the system. Finally, test the experimental data get from water bath in the lab dish. The relative error of output value does not exceed 8%.The experiment results show that genetic algorithm has a good effect in the system identifier on the calibration and identification of COD measuring system, better than radial basis function neural network.

• Periodical: Applied Mechanics and Materials (Volumes 341-342)
• Pages: 748-753
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.341-342.748
• Online since: July 2013
• Authors: Jia Ni Qian, Tian Wang, Xi Peng Lv, Yun Long Tang, Xiu Fen Ye
• Keywords: CODsystems Identification, Genetic Algorithm (GA), Online Surveys, Three-Electrode Method

References

[1] Tian Chunjun, Liu Pide, Study on Linear Fitting Algorithm Based on Matlab, Information Command Control System & Simulation Technology , Vol. 27. No. 6, Dec. (2005).

[2] Muhlenbein,H. and Schlierkamp Voosen,D. Predictive models for the breeder genetic algorithm: I. continuous parameter optimization . Evolutionary Computation, 1(1), 1993. pp: 25-49.

DOI: 10.1162/evco.1993.1.1.25

[3] Huang Xuming, A Kind of Improved Hereditary Algorithm, Journal of Changsha University . Vol. 19, No. 5, Sep. (2005).

[4] Mohammad M. Ahmadi, Graham A. Jullien, Current- Mirror Based on Potentiostats for Three-Electrode Amperometric Electrochemical Sensors[J]. Transactions on\ Circuits and Systems. IEEE. Vol. 56. No. 7, pp: 1339-1348.

DOI: 10.1109/tcsi.2008.2005927

[5] Wang Fulin, Wang Jiquan, Wu Changyou, The Improved Research on Actual Number Genetic Algorithms[J]. Journal of Biomathematics. Vol. 21. No. 1, 2006, pp: 153-158.

[6] Xiao J M , W ang X H. Study on Traffic Flow Prediction using RBF Neural Network[A]. Machine Learning and Cybernetics. Proceedings of 2004 International Conference on . Vol. 5[C] . 2004. pp: 2672-2675.

DOI: 10.1109/icmlc.2004.1378288

[7] Liu Chenglong , Yang Tianyu. Study on Method of GPS Height Fitting Based on BP Artificial Neural Network[J]. Journal of Southwest Jiaotong University, Vol. 42 No. 2. 2007. pp.148-151.

[8] Wang Jiexian. A Method for Fitting of Conicoid in Industrial Measurement, Geomatics and Information Science of Wuhan University, Vol. 32, No. 1, (2009).

[9] Gu Chuan, Pan Guorong, Shi Guigang, Chen Xingquan, Parameter Identification of Surface Fitting Based on Genetic Algorithm. Geomatics and Information Science of Wuhan University , Vol. 34, No. 8, Aug. (2009).