Developing GMDH-Type Neural Network Models for Interrupted Plate Fins

فاقد چكيده فارسي

In the present research,the group method of data handling (GMDH) neural networks (NNs) were developed to model heat transfer and pressure drop in the interrupted plate fins. The models were developed for interrupted plate fins with different geometric parameters including; fin length (r), fin interruption (s), fin pitch (p), and fin thickness (t). Hence, the investigated operating parameters such as Reynolds number (Re), spanwise spacing ratio (p/t), and streamwise spacing ratio (s/r) were used as input data. The target variables are Nusselt number (Nu) and friction factor (f) as heat transfer and pressure drop characteristics of the fluid. The numerical-validate data of previous work used for training the GMDH-type NNs. The error values between the estimated and target data were found 1.33 % and 6.93 % for Nu and f, respectively. Furthermore, the precision of the GMDH results was compared with typical multilayer perceptron (MLP) neural networks. The results show the superiority of the GMDH in predicting Nu. However, a less prediction accuracy was observed for f.