The cooling performance of biological water-silver nanofluid (NF) in three double-pipe heat exchangers with converging sinusoidal inner wall (SCDHX), converging inner wall (CDHX), and plain inner wall (PDHX) was examined numerically using the first and second laws of thermodynamics. The two-phase mixture model is employed to conduct the simulations. Based on the results, the sinusoidal wall increases the flow mixing, and thereby the convective heat transfer coefficient in SCDHX enhances by 50% and 18% over the PDHX type for Res of 500 and 2000, respectively. Moreover, the highest NF frictional entropy generation rate (${\dot{S}}_{f,m,c}$) was obtained for SCDHX; 67% and 80% higher than those for CDHX and PDHX, respectively. The efficiency criteria ratio (η) of SCDHX was obtained as roughly 1.1 for Res of 500 and 1000, which is 27.27% higher than that for CDHX type. Also, the efficiency criteria ratio of SCDHX over the CDHX was in the range of 1.21-1.41. Besides, a robust soft computing, namely the Gaussian process regression (GPR) approach, was proposed to accurately estimate the total entropy of cold NF, the total entropy of hot NF, and the performance ratio based on the nanoparticle concentration and Reynolds number.

使用第一和第二个数值研究了具有收敛正弦内壁（SCDHX），收敛内壁（CDHX）和平面内壁（PDHX）的三个双管换热器中生物水-银纳米流体（NF）的冷却性能热力学定律。采用两相混合模型进行模拟。根据结果​​，正弦壁增加了流动混合，因此对于Re s 为 500 和 2000，SCDHX 中的对流传热系数比 PDHX 类型分别提高了 50% 和 18% 。此外，最高的 NF 摩擦熵产生率 (${\dot{\mathrm{小号}}}_{F,米,C}$) 是为 SCDHX 获得的；分别比 CDHX 和 PDHX 高 67% 和 80%。在Re s 为 500 和 1000 时， SCDHX 的效率标准比 (η) 约为 1.1，比 CDHX 型高 27.27%。此外，SCDHX 与 CDHX 的效率标准比在 1.21-1.41 范围内。此外，提出了一种稳健的软计算，即高斯过程回归（GPR）方法，可以根据纳米粒子浓度和雷诺数准确估计冷NF的总熵、热NF的总熵和性能比。