An Artificial Neural Network-Based Model for Predicting Heat Transfer Coefficient in Heating-Cooling Systems with Chemically Synthesized Nanofluid
نویسندگان
1 Department of Chemical Engineering, Borujerd Branch, Islamic Azad University, Borujerd, I.R. IRAN
2 Department of Chemical Engineering, Borujerd Branch, Islamic Azad University, Borujerd, I.R. IRAN
3 Department of Chemical Engineering, Arak Branch, Islamic Azad University, Arak, I.R. IRAN
4 Department of Chemical Engineering, Arak Branch, Islamic Azad University, Arak, I.R. IRAN
doi
10.30492/ijcce.2025.2039625.6792چکیده
The purpose of this researsh is to investigate the efficiency of simultaneous use of /water hybrid nanofluid (NF) in concentrations of 0.1 to 0.5% by volume and using blade turbulators with a 45-degree angle with different geometric shapes in the Reynolds number range of 3746 to 10041 in Convective Heat Transfer Coefficient (CHTC) and Nusselt number in the Double Pipe Heat Exchanger (DPHE). SEM photography methods were used to identify and determine the size of synthesized NanoParticles (NPs). Heat transfer (HT) experiments were designed using the Response Surface Methodology (RSM). Additionally, an Artificial Neural Network (ANN) with the Levenberg-Marquardt (LM) algorithm was utilized for data prediction. The most optimal increase in the hybrid NF with a concentration of 0.5% by volume with a Reynolds number of 10041 and a pitch distance of 5 mm with two-sided blades is 57 and 55% for the CHTC and the Nusselt number, respectively. The highest and lowest Thermal Performance Factor (TPF) for NF with concentration 0.5 and Reynolds numbers 3746 and 10041 are equal to 0.9085 and 4.0454. The obtained values for the Margin of Deviation (MOD), correlation coefficient (R2), and Mean Squared Error (MSE) using RSM and an ANN with a 4-22-1 topology are as follows: ±10 and ±0.01%, 0.9938, 1, 0.1, and 0.000000441, respectively.
