Prediction of minimum surface roughness in end milling mold parts using neural network and genetic algorithm


Oktem H., ERZURUMLU T., ERZİNCANLI F.

MATERIALS & DESIGN, cilt.27, sa.9, ss.735-744, 2006 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 27 Sayı: 9
  • Basım Tarihi: 2006
  • Doi Numarası: 10.1016/j.matdes.2005.01.010
  • Dergi Adı: MATERIALS & DESIGN
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Sayfa Sayıları: ss.735-744
  • Anahtar Kelimeler: end milling, cutting parameters, mold parts, surface roughness, neural network, genetic algorithm, injection molding, MODEL, MACHINABILITY
  • Kocaeli Üniversitesi Adresli: Evet

Özet

This paper presents an approach for determination of the best cutting parameters leading to minimum surface roughness in end milling mold surfaces of an ortez part used in biomedical applications by coupling neural network and genetic algorithm. In doing this, design of experiments, neural network and genetic optimization technique are utilized in integrated purpose. A series of cutting experiments for mold surfaces in one component of ortez part are conducted to obtain surface roughness values. A feed forward neural network model is developed exploiting experimental measurements from the surfaces in the mold cavity. The neural network model is trained and tested in MATLAB. Genetic algorithm coupled with neural network is employed to find optimum cutting parameters leading to minimum surface roughness without any constraint. For this purpose, a simulation model for the component of ortez part was created to determine the critical regions to be used in roughness measurements and to produce a plastic product. Additional measurements were performed to validate optimum values and their corresponding to roughness value predicted by genetic algorithm with the values obtained from experiments in the mold cavity and on plastic product. From this, it is clearly seen that a good agreement is observed between the predicted values and experimental measurements. (c) 2005 Elsevier Ltd. All rights reserved.