A Predictive Surrogate Model for Heat Transfer of an Impinging Jet on a Concave Surface

TL;DR

The paper addresses predicting heat transfer ($Nu$) for a concave-surface cooling scenario under pulsating impingement. It evaluates multiple MOR and DL approaches, including FFT-ANN for constant-frequency $Nu$ and POD-LSTM for random-frequency $Nu$, and compares time-series predictions using LSTM and Transformer, plus local-field reconstruction with POD-LSTM. Results show Transformer provides longer-horizon, more accurate forecasts for the average $Nu$ under random-frequency forcing, while POD-LSTM effectively reconstructs local heat-transfer fields with acceptable error (max ~8%). The work demonstrates the viability of advanced MOR and ML techniques to accelerate CFD-based heat-transfer analysis and offers a path toward efficient surrogate models for complex cooling geometries.

Abstract

This paper aims to comprehensively investigate the efficacy of various Model Order Reduction (MOR) and deep learning techniques in predicting heat transfer in a pulsed jet impinging on a concave surface. Expanding on the previous experimental and numerical research involving pulsed circular jets, this investigation extends to evaluate Predictive Surrogate Models (PSM) for heat transfer across various jet characteristics. To this end, this work introduces two predictive approaches, one employing a Fast Fourier Transformation augmented Artificial Neural Network (FFT-ANN) for predicting the average Nusselt number under constant-frequency scenarios. Moreover, the investigation introduces the Proper Orthogonal Decomposition and Long Short-Term Memory (POD-LSTM) approach for random-frequency impingement jets. The POD-LSTM method proves to be a robust solution for predicting the local heat transfer rate under random-frequency impingement scenarios, capturing both the trend and value of temporal modes. The comparison of these approaches highlights the versatility and efficacy of advanced machine learning techniques in modelling complex heat transfer phenomena.

Figures (17)

• Figure 1: Schematic of the computational model
• Figure 2: Structured grid for the model with a closer look at the boundary layer mesh near the concave surface
• Figure 3: Variation of local Nusselt number for different meshes
• Figure 4: Utilized boundary conditions for the CFD simulation
• Figure 5: The performance of turbulence models to predict the local Nu number distribution