Mixed Convective Heat Transfer Characteristic in Vented Cavity Under Dynamic Flow Modulation by CFD and Neural Network Model Approaches
- 9 April 2025
- journal article
- Published by Wiley in Heat Transfer
- Vol. 54 (5) , 3072-3087
Abstract
Two prominent tools of current engineering research: computational fluid dynamics (CFD) and neural network (NN) model are employed to investigate the mixed convection phenomenon within a representative compact thermal system. The thermal system of interest is a vented cavity with inlet and outlet ports incorporating a rotating cylinder as a dynamic flow modulator. The dynamic state of the modulator is characterized by speed ratio (ψ)—a ratio of the cylinder's peripheral speed to the mean intake air velocity. The governing mass, momentum, and energy equations are discretized and solved in a dimensionless format using the Galerkin finite element method to represent the thermal field and flow field in terms of heat transfer and pressure drop characteristics. These characteristics are investigated using five key parameters, including the dynamic flow conditions which are characterized by Reynolds number (Re), Richardson number (Ri), speed ratio (ψ), and the position of the modulator (xc, yc). The accuracy of the CFD model is ensured through validation against established literature. The results obtained from the CFD framework are utilized to construct and compare the performance of two NN models, namely the artificial neural network (ANN) and the adaptive neuro‐fuzzy inference system (ANFIS), with varying numbers of neurons in the hidden layer and several training algorithms. A notable agreement has been found between the numerical outcomes and the predictions generated by distinct NN models. “Bayesian regulation” (BR) algorithm for ANN and “Trapezoidal membership function” (trimf) for ANFIS yield the most accurate results. Moreover, the research findings indicate that the ANN method may provide a faster and more precise estimation of thermal and pressure drop characteristics compared to the ANFIS method, making it a highly suitable approach for real‐time applications in compact thermal systems. An analysis of the Pearson coefficient in the present context highlights that cylinder's position influences both the heat transfer and pressure drop most, with speed ratio, Reynolds number, and Richardson number following in descending order of impact.Keywords
This publication has 37 references indexed in Scilit:
- Mixed convection heat transfer by nanofluids in a cavity with two oscillating flexible fins: A fluid–structure interaction approachApplied Mathematical Modelling, 2020
- Thermal optimization of shell and tube heat exchanger using porous bafflesApplied Thermal Engineering, 2020
- Design of computational intelligent procedure for thermal analysis of porous fin modelChinese Journal of Physics, 2019
- Performance investigation of the cross-flow closed-type heat-source tower using experiments and an adaptive neuro-fuzzy inference system modelEnergy and Buildings, 2019
- Comparison of ANN, Regression Analysis, and ANFIS Models in Estimation of Global Solar Radiation for Different Climatological Locations2018
- Optimization of heat transfer enhancement of nanofluid in a channel with winglet vortex generatorApplied Thermal Engineering, 2015
- Multiple regression, ANN and ANFIS models for prediction of backbreak in the open pit blastingEngineering with Computers, 2012
- ANN and ANFIS models for performance evaluation of a vertical ground source heat pump systemExpert Systems with Applications, 2010
- An experimental and numerical investigation of mixed convection from a heat generating element in a ventilated cavityExperimental Thermal and Fluid Science, 2007
- Prediction of flow fields and temperature distributions due to natural convection in a triangular enclosure using Adaptive-Network-Based Fuzzy Inference System (ANFIS) and Artificial Neural Network (ANN)International Communications in Heat and Mass Transfer, 2007