North China University of Technology, Beijing, China

---

Abstract—The diagnosis system for the alumina reduction1 was developed on the basis of BP neural network with optimization by genetic algorithm. The neural network used the characteristic vectors composed of the frequency energy calculated from cell resistance as 10 inputs and three cell statuses as 3 outputs. The neural network was certified by industrially sampling data. The results showed the accuracy ratio was larger than 80%, which can meet the requirements in the aluminum production. The diagnosis software was designed and applied in an aluminum smelter.

Index Terms—Alumina reduction, diagnosis, neural network.

[PDF]

Cite: Shuiping Zeng, Lin Cui, Jinhong Li, "Diagnosis System for Alumina Reduction Based on BP Neural Network," Journal of Computers vol. 7, no. 4, pp. 923-928, 2012.