Uluslararası Doğu Anadolu Fen Mühendislik ve Tasarım Dergisi

Yazarlar: Aslı DURMUŞOĞLU, Sibel GÜNEŞ, Ersin KARAKAYA

Konular:Mühendislik, Makine

Anahtar Kelimeler:Heat transfer enhancement,Fluid characteristics,ANN

Özet: In this paper, we present an Artificial Neural Networks (ANNs) model which is prone to accurately predict the friction factor and Nusselt number of a tube with loose-fit perforated twisted tapes. Experimantal tests were realized using the tapes with three different rates of pitch length of twisted tape to inner diameter of tube (y/D=2, 2. 5, 3), two different twisted tape width rates (W/D= 0. 9285, 0. 9642) and three different rates of hole to inner diameter (d/D= 0. 0714, 0.107, 0.143) in a range of Reynolds number 4860 to 24,130 under uniform heat flux conditions. The ANN model was improved and validated using a databank containing experimental datasets. The back propagation algorithm, which is the best training algorithm, is recognized to be the most extensive learning method for ANN. This algorithm is used for training and testing of the network. The ANNs results were found to good compliance with the experimental data. Value of the coefficient of multiple determination were obtained. The R2 values are 0,9992 for Nusselt number and 0,9995 for friction factor.

---