The effects of using hybrid nanofluids and of helical coil pitch (λ) in a 3D shell and tube heat exchanger (STHE) are investigated. The algorithm used in this study is Phase Coupled SIMPLE and the method used is Eulerian. Nanofluid flow with Reynolds (Re) numbers of 10,000, 15,000, and 20,000, nanoparticles with volume fractions (ϕ) of 2 and 4%, and λ = 20, 25, 40, and 50 mm are investigated. The highest numbers related to the thermal index (Nu) and effectiveness occurred in the λ = 20 mm and the maximum ϕ and Re. In the case of λ = 20 mm, the maximum Nusselt number is 15.8%, 26%, and 45.3% more than that of 25, 40, and 50 mm, respectively. However, in the same case, in comparison between the ϕ = 4% and ϕ = 0, the Nu increases by 45.7%, 61.7%, and 76%. The present study shows that combining using hybrid nanofluids and changing the geometry of STHE, as an innovative approach can positively increase efficiency. Finally, the results are used for training an artificial neural network (ANN). In this regard, for finding the optimum neuron numbers in the hidden layer, the optimum feed-forward network is obtained to predict the efficiency of the material.

研究了在 3D 管壳式换热器 (STHE) 中使用混合纳米流体和螺旋线圈​​节距 (λ) 的影响。本研究中使用的算法是 Phase Coupled SIMPLE，使用的方法是 Eulerian。研究了雷诺数 ( Re ) 为10,000、15,000和 20,000 的纳米流体流动，纳米流体的体积分数 ( φ)为 2% 和 4%，λ = 20、25、40 和 50 mm。与热指数 (Nu) 和有效性相关的最高数字出现在 λ = 20 mm 和最大φ和Re中。在 λ = 20 mm 的情况下，最大努塞尔数分别比 25、40 和 50 mm 多 15.8%、26% 和 45.3%。然而，在同样的情况下，比较φ  = 4% 和φ = 0时，Nu 增加 45.7%、61.7% 和 76%。本研究表明，结合使用混合纳米流体和改变 STHE 的几何形状，作为一种创新方法可以积极提高效率。最后，结果用于训练人工神经网络（ANN）。在这方面，为了找到隐藏层中的最佳神经元数量，获得最佳前馈网络来预测材料的效率。