Determination of Heat Transfer Coefficients at Different pH Values of a Nanofluids and Modeling with Decision Tree Algorithm

It is important to be able to use the energy in a more beneficial way by increasing the heat transfer in the in-pipe flows. Because, with the technological developments, there is an increasing energy demand in the industry sector. For this reason, researchers have been working on new generation heat transfer fluids in recent years. In our study, nanoparticle production of CuO (copper oxide) was performed. Scanner electron microscope (SEM) image analysis and X-ray diffraction method analysis (XRD) analysis were performed to show that the material produced has the properties of nano material. A nanofluid was obtained using pure water, ethanol and ethylene glycol materials with CuO nanoparticles. Heat transfer coefficients were determined at different pH values of the obtained nanofluid. In the experimental studies, the Re value was 887 and 2290, whereas the heat transfer coefficient value was 349,821 (W/m² ° C) and 374,253 (W/m²°C), respectively. The pH value was 7.84 and 9.95, while the heat transfer coefficient was 349,821 (W/m²°C) and 374,253 (W/m²°C), respectively. Predictive models were obtained by using decision tress (DT) algorithms for heat transfer coefficients calculated by experimental studies. In order to determine the validity of the obtained models, mean square error (MSE), root mean square error (RMSE) and mean absolute percentage error (MAPE) analysis were performed. As a result, it was observed that the heat transfer coefficient value of the nanofluid decreased as the pH values increased. It was calculated that the heat transfer coefficient of the nanofluid obtained at different Reynolds values was 13.3% higher than the heat transfer coefficient of pure water.It has been shown that the KA algorithm, which is a computational intelligence method, was successful in estimating the thermophysical properties of nanofluids according to the value of 0.891 MAPE.