Simulation and accurate prediction of thermal efficiency of functionalized COOH-MWCNT/water nanofluids by artificial neural network using experimental data

In this investigation, experimental data of nanofluids have been modeled by the Artificial Neural Network (ANN) method for Functionalized COOH-MWCNT nanoparticles based on water in the heat exchanger. Nanofluids' thermal efficiency is a function of partial pressure difference and Reynolds number. To model this data, an ANN with two hidden layers with 7 and 8 neurons in the first and second layers, respectively, is used. The tansing transfer function is used in both hidden layers. The type of modeling program continues until appropriate accuracy is achieved and also adjusts the structure of the ANN by specifying the desired precision. A comparison between the obtained results of experimental data and the results of ANN shows that the ANN model can predict efficiency well. The regression coefficient of 98.9 % is obtained from the modeling shows the high accuracy of the modeling method. The maximum value of MSE is equal to 0.00295. In this investigation, it is also observed that nanofluids efficiency increases with increasing of vol- ume fraction of nanoparticles (phi). This means that the heat transfer coefficient increases further than the partial pressure difference. The lowest range of the margin of deviation (MOD) is ob- tained in the range of (+1 and-1).