A Comparison for Grain Size Calculation of Cu-Zn Alloys with Genetic Programming and Neural Networks
Yükleniyor...
Tarih
2015
Yazarlar
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
Polish Acad Sciences Inst Physics
Erişim Hakkı
info:eu-repo/semantics/openAccess
Özet
Neural Networks (NN) and Genetic Programming (GP) were used as a new method for formulation of grain size of electrodeposited Cu1-xZnx alloys as a function of Zinc and Copper content both electrolyte and the alloy films produced by electrodeposition technique. To predict grain size a great number of different expression models genetic programming and neural network were conducted. Each model differs from the other with their linking function, number of genes, head size, and chromosomes. To generate databases for the new grain size formulations, testing and training sets in total of 134 samples were selected at different Zn and Cu ratios of components. 6 different input parameters were selected and the output parameter was grain size of the electrodeposited Cu-Zn alloys. The testing and training sets consisted of randomly selected 106 and 28 for the proposed models. All results in the models indicated an applicable performance for predicting grain size of the alloys and found reliable. The predicted model showed that all of the input parameters effected on the resulting grain size. The NN and GEP based formulation results are compared with experimental results and found to be quite reliable with a very high correlation (R2 = 0.995 for GEP and 0.999 for NN).
Açıklama
International Conference on Computational and Experimental Science and Engineering (ICCESEN) -- OCT 25-29, 2014 -- Antalya, TURKEY
Anahtar Kelimeler
Formulation, Prediction
Kaynak
Acta Physica Polonica A
WoS Q Değeri
Q4
Scopus Q Değeri
Q3
Cilt
128
Sayı
2B
