Purpose

Numerical simulations are performed to determine the heat transfer characteristics of slot jet impingement of air on a concave surface. The purpose of this paper is to investigate the effect of protrusions on the heat transfer by placing semi-circular protrusions on the concave surface at several positions. After identifying appropriate locations where the heat transfer is a maximum, multiple protrusions are placed at desired locations on the plate. The gap ratio, curvature ratio (d/D) and the dimensions of the plate are varied so as to obtain heat transfer data. The curvature ratio is varied first, keeping the concave diameter (D) fixed followed by a fixed slot width (d). A surrogate model based on an artificial neural network is developed to determine optimum locations of the protrusions that maximize the heat transfer from the concave surface.

Design/methodology/approach

The scope and objectives of the present study are two-dimensional numerical simulations of the problem by considering all the geometrical parameters (H/d, d_p, Re, θ) affecting heat transfer characteristics with the help of networking tool and numerical simulation. Development of a surrogate forward model with artificial neural networks (ANNs) with a view to explore the full parametric space. To quantitatively ascertain if protrusions hurt or help heat transfer for an impinging jet on a concave surface. Determination of the location of protrusions where higher heat transfer could be achieved by using exhaustive search with the surrogate model to replace the time consuming forward model.

Findings

A single protrusion has nearly no effect on the heat transfer. For a fixed diameter of concave surface, a smaller jet possesses high turbulence kinetic energy with greater heat transfer. ANN is a powerful tool to not only predict impingement heat transfer characteristics by considering multiple parameters but also to determine the optimum configuration from many thousands of candidate solutions. A maximum increase of 8 per cent in the heat transfer is obtained by the best configuration constituting of multiple protrusions, with respect to the baseline smooth configuration. Even this can be considered as marginal and so it can be concluded that first cut results for heat transfer for an impinging jet on a concave surface with protrusions can be obtained by geometrically modeling a much simpler plain concave surface without any significant loss of accuracy.

Originality/value

The heat transfer during impingement cooling depends on various geometrical parameters but, not all the pertinent parameters have been varied comprehensively in previous studies. It is known that a rough surface may improve or degrade the amount of heat transfer depending on their geometrical dimensions of the target and the rough geometry and the flow conditions. Furthermore, to the best of authors’ knowledge, scarce studies are available with inclusion of protrusions over a concave surface. The present study is devoted to development of a surrogate forward model with ANNs with a view to explore the full parametric space.

中文翻译：

---

射流形态的数值模拟与优化

目的

进行数值模拟以确定缝隙空气撞击凹面时的传热特性。本文的目的是通过在凹面上的多个位置放置半圆形凸起来研究凸起对传热的影响。在确定最大热传递的适当位置之后，将多个突起放置在板上的所需位置。改变间隙比，曲率比（d / D）和板的尺寸以获得热传递数据。首先改变曲率比，保持凹面直径（D）固定，然后保持固定的槽宽（d）。开发了基于人工神经网络的替代模型，以确定突出的最佳位置，以使从凹面传热最大化。

设计/方法/方法

本研究的范围和目的是借助网络工具和数值模拟，通过考虑影响传热特性的所有几何参数（H / d，d _p，Re，θ），对该问题进行二维数值模拟。开发具有人工神经网络（ANN）的代理正向模型，以探索整个参数空间。定量确定突起是否伤害或帮助热传递，以防止凹面的撞击射流。通过使用穷举搜索和替代模型来替代耗时的前向模型，可以确定可以实现更高热传递的凸起位置。

发现

单个突起几乎不影响传热。对于固定直径的凹面，较小的射流具有较高的湍流动能和较大的热传递。人工神经网络是一种功能强大的工具，它不仅可以通过考虑多个参数来预测碰撞传热特性，还可以从成千上万个候选解决方案中确定最佳配置。相对于基线平滑配置，由多个突起组成的最佳配置可实现最大传热增加8％。

创意/价值

冲击冷却过程中的热传递取决于各种几何参数，但是，在先前的研究中，并非所有相关参数都已全面改变。众所周知，粗糙表面会根据其目标的几何尺寸以及粗糙的几何形状和流动条件而改善或降低传热量。此外，据作者所知，很少进行研究，包括在凹形表面上形成突起。本研究致力于开发具有人工神经网络的代理正向模型，以探索整个参数空间。