Numerical simulation and optimization of single-phase turbulent flow in chevron-type plate heat exchanger with sinusoidal corrugations
Numerical simulation and optimization of single-phase turbulent flow in chevron-type plate heat exchanger with sinusoidal corrugations
复制标题
DOI：
10.1080/10789669.2011.558167
复制
发表时间：
2011-04
期刊：
影响因子：
--
通讯作者：
Weizhe Han;K. Saleh;V. Aute;G. Ding;Y. Hwang;R. Radermacher
中科院分区：
文献类型：
--
作者：
Weizhe Han;K. Saleh;V. Aute;G. Ding;Y. Hwang;R. Radermacher
本文采用数值模拟的方法研究了具有正弦波纹的人字形板式换热器内单相湍流流动的热流体动力学特性。计算区域包含一个湍流通道，湍流模型采用剪切应力输运κ-ω模型。数值模拟结果的努塞尔数和摩擦因子进行了比较有限的实验数据和现有的关联，以验证数值模型的准确性。使用单相水作为工作流体进行验证。采用近似辅助优化技术进行了双目标优化，节省了大量计算量，确保了优化的可行性。以螺旋桨深度、螺旋桨角度、螺旋桨桨螺距和流体入口速度为设计变量，采用最大熵设计方法选取200个样本建立了元模型，获得了单位长度的传热系数和压降。采用多目标遗传算法作为优化器。优化结果以Pareto解集的形式给出，表明了该方法在整个设计空间中的优势，以及在最大传热系数和最小单位长度压降这两个优化目标之间的权衡。此外，帕累托最优设计进行了验证，对数值模拟直接获得的值。近似辅助优化结果表明，所有优化设计均在设计空间内具有最大的放大因子值，且最优换热角随最大换热系数的增大而增大。
The thermal-hydrodynamic characteristics of a single-phase turbulent flow in chevron-type plate heat exchangers with sinusoidal-shaped corrugations have been numerically investigated in this article. The computational domain contains a corrugation channel, and the simulations adopted the shear-stress transport κ-ω model as the turbulence model. The numerical simulation results in terms of Nusselt number and friction factor were compared with limited experimental data and existing correlations in order to verify the accuracy of the numerical model. The validation was performed using single-phase water as the working fluid. A bi-objective optimization was carried out using an approximation-assisted optimization technique, which provided considerable computational savings to ensure the feasibility of optimization. The corrugation depth, corrugation angle, corrugation pitch, and fluid inlet velocity were identified as design variables, and 200 samples were selected using the maximum entropy design method to build the metamodel for obtaining the heat transfer coefficient as well as the pressure drop per unit length. A multi-objective genetic algorithm was utilized as the optimizer. The optimization results were presented in the form of Pareto solution set, which clearly showed its dominance over the entire design space and the tradeoff between the two optimization objectives: maximizing heat transfer coefficient and minimizing drop per unit length. Also, the Pareto optimal designs were validated against the values directly obtained from numerical simulations. The approximation-assisted optimization shows that all optimal designs have largest enlargement factor values inside the design space, and the optimal corrugation angle increases with the increase of maximum heat transfer coefficient.