دانلود رایگان مقاله سیستم استنتاج فازی عصبی تطبیقی انتقال حرارت نانو سیال در لوله گرم

Abstract

In this paper, estimating of hydrodynamics and heat transfer nanofluid flow through heated tube has been conducted by using Adaptive Neuro-Fuzzy Inference System (ANFIS). The CFD data related to three types of nanofluids (Al2O3, SiO2 and TiO2) flow in horizontal tube with 19 mm diameter and 2000 mm length. Heat flux around tube is fixed at 5000 W/m2, the range of Reynolds number is (3000–30,000) and volume concentrations are (1% and 2%). ANFIS model has three input data presented by Reynolds number, volume concentration of nanofluids and materials and two output presented predicting friction factor and Nusselt number in the tube. The simulation results of proposed algorithm have been compared with CFD simulator in which the mean relative errors (MRE) are 0.1232% and 0.1123 for friction factor and Nusselt number respectively. Finally, ANFIS models can predict hydrodynamics and heat transfer of the higher accuracy than the developed correlations.

6. Conclusions

There are two parts in this article; the first is the numerical study of turbulent nanofluid flow in the circular heated tube. The influence of Reynolds number (Re), nanofluid volume concentration (ϕ) and the nanofluids type on the friction factor and Nusselt number were studied. The second part is included the intelligent study using ANFIS to find friction factor and Nusselt number through circular heated tube. It can be concluded that: 1. Friction factor increases with increasing of volume concentrations but decreases with increasing of Reynolds number. 2. Nusselt number increases with increasing of both volume concentrations and Reynolds number. 3. Al2O3 nanofluid has higher friction factor than SiO2 and TiO2, furthermore, SiO2 nanofluid has higher Nusselt number than Al2O3 and TiO2. 4. ANFIS is completed iterations with (0.00926 s) time but CFD is completed iterations with (183 s), so reducing time by using ANFIS for the case undertaken. 5. The prediction of the friction factor and Nusselt number with the ANFIS models is in good agreement with the CFD analysis with maximum error of less than 0.1282.