In this study, Multi-Layer Perceptron (MLP) model of the Artificial Neural Network (ANN) is used to predict the thermo - hydraulic performance of a circular perforated absorber plate with cross flow arrangement. The perforations on the absorber plate are selected as 24, 36 and 54 and the diameters of the holes are 5 mm, 8 mm and 10 mm. The flow rate of inlet air is varied from 3000 to 21000. The results predicted by ANN are compared with the values obtained from experimental analysis. In order to generate the ANN results, 64 data sets are analyzed with 45 samples for training, 9 samples for testing and 9 samples for validation. Lavenberg Marquardt (LM) algorithm consisting of 8 inputs and 1 output with 11 neurons in the hidden layer is used. The number of neurons in the hidden layer have been optimized by least training error. The adopted model has root mean square error (RMSE) of 0.76 and R 2 of 0.9972. An average 1.58% error is obtained for the ANN model compared to the experimental thermohydraulic efficiency. ANN can be used successfully to predict the performance of the solar air heater with a circular, perforated absorber plate.
All Science Journal Classification (ASJC) codes
- Fluid Flow and Transfer Processes