Key Laboratory of Digital Manufacturing Technology and Application, The Ministry of Education, Lanzhou University of Technology, Lanzhou, 730050, China; School of Mechanical and Electronical Engineering, Lanzhou University of Technology

---

Abstract—At present, although most fault diagnosis methods of rotating machinery is qualitatively used, it is gravely lacking in quantitative accuracy. So a novel algorithm GAHPSO combining with the advantages of genetic algorithm (GA), simulated annealing (SA) and particle swarm optimization (PSO) was provided to train neural network (NN). The proportion and blend methods were applied to the novel algorithm. Information entropy was used to take fault signals. Four kinds of spectral entropies and six kinds of typical rotor faults were used as input and output data. NN classifier based on GA-HPSO was set up. The simulation results indicate that GA-HPSO has a better ability to escape from a local minimum and is more effective than the conventional single algorithm. It can rapidly and accurately realize fault data classification. It provides a new method for fault diagnosis.

Index Terms—Information Entropy; GA-HPSO-NN; Fault Diagnosis

[PDF]

Cite: Bin Peng, Chang-an Hu, Rong-zhen Zhao, Yu-qiao Zheng, "The Investigation of Fault Diagnosis Based on GA-HPSO-NN," Journal of Computers vol. 7, no. 10, pp. 2511-2517, 2012.