Acta Mechanica Slovaca 2012, 16(2):54-60 | DOI: 10.21496/ams.2012.019

Application of Genetic Algorithm for Optimization of Neural Networks for Selected Tribological Test

Tomasz Trzepieciński^a*, Hirpa Lemu Gelgele^b

^a Rzeszow University of Technology, Department of Materials Forming and Processing, ul. W. Pola 2, 35-959 Rzeszów, Poland

^b University of Stavanger, Department of Mechanical and Structural Engineering and Materials Science, 4036 Stavanger, Norway

In this work was presented the method of determination of the friction coefficient by using multilayer artificial neural networks on the basis of experimental database obtained from the strip drawing test. Using genetic algorithm the optimization of number of input variables of artificial neural networks has been done. As an input parameters for training artificial neural networks following parameters has been used: surface parameters of the sheet and dies, sheet material parameters and clamping force. Some results have pointed out that genetic algorithm has been successfully appled to optimization of training set.

Keywords: Friction, friction coefficient, genetic algorithm, artificial neural networks

Received: January 15, 2012; Accepted: July 15, 2012; Published: October 31, 2012  Show citation