Neural-network based fault diagnosis of hydraulic forging presses in China

Abstract

The paper describes the utilization of neural networks for fault diagnosis of hydraulic forging presses which may have an impact on the effective utilization of the over 2000 presses in use in China. The technical descriptions of the presses and the 47 major possible faults are presented. For diagnosing these faults the neural network with 30 000 iteration training was utilized and it provided a 99% accuracy in identifying causes of the failures of hydraulic forging presses.

Additional information

Notes on contributors

G. SALVENDY

To whom correspondence should be addressed.