Research on Aluminum Electrolytic Multi-Fault Diagnosis Method Based on Immune Genetic Algorithm

Abstract:

As for a wide variety of faults that happen frequently during the aluminum electrolysis process, a new method of multi-fault diagnosis method using neural network based on immune genetic algorithm (IGA) is proposed. IGA has the abilities of searching for global optima and better convergence. By applying these abilities and the diagnosis characteristics of the aluminum electrolysis process, the study builds the layered fault diagnosis model structure . The results of simulations show that this model is of the better ability of convergent on whole solution space and the capacity of fast learning than that of the traditional fault diagnosis model, therefore, the method worths applying widely.

• Periodical: Advanced Materials Research (Volumes 706-708)
• Pages: 1159-1162
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.706-708.1159
• Online since: June 2013
• Authors: Jie Jia Li, Wen Yue Guan
• Keywords: Aluminium Electrolysis, Fault Diagnosis, Immune Genetic Algorithm

References

[1] Jiejia Li,Yang Li,Chengdong Wu. The Application of Fuzzy Neural Network in the Aluminum Eelectrolysis Fault Diagnosis.Journal of Shenyang JianzhuUniversity(Natural Science),2005.04.

[2] Naixiang Feng. Aluminum Electrolysis.Beijing:Chemical Industry Press[M], 2006：50-58.

[3] Jianxin Liu.Discussion about the hazard of anode effect in aluminumelectrolysis [J]. Metallurgical Collections, 2005.4(2):41-44.

[4] Yinghui Gao,Jia Wang,Bin Liu.Fault Diagnosis of Frequency Conversion Speed Adjusting Pump Based on Radial Basic Function Neural Network. Measurement and Control Automation, 2011(05)-0047-03.

[5] Haykin,S. Neural Network Principle. Beijing:China Machine Press, 2004.1.

[6] Weirong Wang,Peng Jing,Kang Huang.Immune Genetic Algorithm and Its Application on Multi-object Optimization Design. Mechanical Engineering and Automation, Dec, 2006,06-0010-03.