IJE TRANSACTIONS B: Applications Vol. 30, No. 11 (November 2017) 1458-1467    Article in Press

PDF URL: http://www.ije.ir/Vol30/No11/B/18.pdf  
downloaded Downloaded: 0   viewed Viewed: 161

M. Momeni, M. B. Laruccia, X. Liu, A. Atashnezhad and S. Ghaheri
( Received: June 08, 2017 – Accepted: September 08, 2017 )

Abstract    This study is designed to consider the two important yet often neglected factors, which are factory recommendation and bit features, in optimum bit selection. Iimage processing techniques to consider the bit features have been used. A mathematical equation, which is derived from a neural network model, is used for drill bit selection to obtain the bit’s maximum penetration rate that corresponds to the optimum parameters for drilling. In the end, the bit with the maximum penetration rate is chosen. The results of this study showed that bit pattern can be inserted in the calculation through a proper bit image processing techniques. This is to ensure that each unique bit can be discriminated from other bits. The values of mean square error 0.0037 and coefficient of determination (R2) 0.9473 were found for the rate of penetration model. The image processing techniques were used to extract the bit features. The artificial neural network black box was converted to white box in order to extract a mathematical equation and visibility of the model.


Keywords    Bit selection, Artificial neural network, Image processing techniques, Genetic algorithm, Optimum drilling operation


Download PDF 

International Journal of Engineering
E-mail: office@ije.ir
Web Site: http://www.ije.ir