Conventional CFD Research Articles

Machine-learning aided calibration and analysis of porous media CFD models used for rotating packed beds

The proposed research is an attempt to advance the state-of-the-art of the numerical modelling of RPB by combining Computational Fluid Dynamics and Machine Learning approaches. The latter creates an accurate framework that should help quantify the potential of Rotating Packed Beds (RPB) technology to intensify conventional CO2 capture processes. To this end, a direct sensitivity analysis is detailed to supplement a machine-learning (ML) algorithm built for calibrating resistance coefficients needed for porous media modelling. The algorithm is used to improve CFD predictions of dry pressure drop in rotating packed beds (RPBs) for a wide range of operating conditions. The sensitivity derivatives with respect to the packing resistance coefficients are demonstrated for the first time in RPBs, which is not available in the current CFD open source and commercial codes. In this regard, sensitivity differential equations are derived from three-dimensional Navier-Stokes equations for porous media in a rotating reference frame. These sensitivity equations are discretized using a finite volume scheme and solved to obtain the sensitivity pressure drop differences at the packing edges. The results are validated against the predictions of the analytical sensitivity analysis and the finite difference approximation. After, the Newton – Gauss method that employs the sensitivity pressure drop derivatives, is used to minimize the error (cost function) between the pressure drop obtained from CFD simulations and the available experimental data. This is achieved by tuning the packing resistance coefficients to the RBPs' operating conditions (gas flowrate and rotating speed) and correlate them using an artificial neural network (ANN). The results of the proposed approach show a significant improvement in porous media-based CFD predictions of RPBs' pressure drop across a wide range of operating conditions and this over conventional porous media-based CFD models. This is necessary for CFD models to be reliably used as a tool that can efficiently improve existing RPBs' designs and/or participate in RPBs' design innovation.

An XAI Framework for Predicting Wind Turbine Power under Rainy Conditions Developed Using CFD Simulations

Renewable energy sources are essential to address climate change, fossil fuel depletion, and stringent environmental regulations in the subsequent decades. Horizontal-axis wind turbines (HAWTs) are particularly suited to meet this demand. However, their efficiency is affected by environmental factors because they operate in open areas. Adverse weather conditions like rain reduce their aerodynamic performance. This study investigates wind turbine power prediction under rainy conditions by integrating Blade Element Momentum (BEM) theory with explainable artificial intelligence (XAI). The S809 airfoil’s aerodynamic characteristics, used in NREL wind turbines, were analyzed using ANSYS FLUENT and symbolic regression under varying rain intensities. Simulations at a Reynolds number (Re) of 1 × 106 were performed using the Discrete Phase Model (DPM) and k–ω SST turbulence model, with liquid water content (LWC) values of 0 (dry), 10, 25, and 39 g/m3. The lift and drag coefficients were calculated at various angles of attack for all the conditions. The results indicated that rain led to reduced lift and increased drag. The innovative aspect of this research is the development of machine learning models predicting changes in the airfoil coefficients under rain with an R2 value of 0.97. The proposed XAI framework models rain effects at a lower computational time, enabling efficient wind farm performance assessment in rainy conditions compared to conventional CFD simulations. It was found that a heavy rain LWC of 39 g/m3 could reduce power output by 5.7% to 7%. These findings highlight the impact of rain on aerodynamic performance and the importance of advanced predictive models for optimizing renewable energy generation.

Numerical simulation of wind turbine wake characteristics by flux reconstruction method

In the paper, a CFD method that employs the flux reconstruction (FR) method and the actuator line method (ALM) is proposed, to develop a novel numerical framework to improve the efficiency and accuracy of turbulent wake predictions of horizontal axis wind turbines (HAWT). The PyFR solver, which is an open-source program that uses the FR method is applied. The ALM is for the first time implemented in the FR solver to simulate HAWTs. To increase the fidelity of the wake prediction, the nacelle and tower of the HAWT are also modeled, and the proposed numerical simulation method is named as AL-PyFR(N&T) model. The NTNU “Blind Test 1” wind turbine is used to validate the proposed model. First, the mesh and time step dependence of the AL-PyFR(N&T) model are verified. Second, the accuracy of the proposed model is verified by comparing the velocity deficits and turbulent kinetic energy results at different downstream locations with the measured values and the results from other CFD studies. By comparison of the results with and without the effect of the nacelle and tower, it is found that tower wake plays a significant role in wake asymmetry. Crucially, the AL-PyFR(N&T) model demonstrates a capacity for accurate wake predictions with fewer mesh requirements compared to conventional CFD models. This efficiency marks a significant advancement for high-fidelity numerical simulations of offshore wind farms.

Rotor Performance Predictions for Urban Air Mobility: Single vs. Coaxial Rigid Rotors

This work details the development and validation of a methodology for high-resolution rotor models used in hybrid Blade Element Momentum Theory Unsteady Reynolds Averaged Navier–Stokes (BEMT-URANS) CFD. The methodology is shown to accurately predict single and coaxial rotor performance in a fraction of the time required by conventional CFD methods. The methodology has three key features: (1) a high-resolution BEMT rotor model enabling large reductions in grid size, (2) a discretized set of momentum sources to interface between the BEMT rotor model and the structured URANS flow solver, and (3) leveraging of the first two features to enable highly parallelized GPU-accelerated multirotor CFD simulations. The hybrid approach retains high-fidelity rotor inflow, wake propagation, and rotor–rotor interactional effects at a several orders of magnitude lower computational cost compared to conventional blade-resolved CFD while retaining high accuracy on steady rotor performance metrics. Rotor performance predictions of thrust and torque for both single and coaxial rotor configurations are compared to test the data that the authors obtained at the NASA Langley 14- by 22-ft. Subsonic Tunnel Facility. Simulations were run with both fully turbulent and free-transition airfoil performance tables to quantify the associated uncertainty. Single rotor thrust and torque were predicted on average within 4%. Coaxial thrust and power were predicted within an average of 5%. A vortex ring state (VRS) shielding phenomenon for coaxial rotor systems is also presented and discussed. The results support that this hybrid BEMT-URANS CFD methodology can be highly parallelized on GPU machines to obtain accurate rotor performance predictions across the full spectrum of possible UAM flight conditions in a fraction of the time required by conventional higher-fidelity methods. This strategy can be used to rapidly create look-up tables with hundreds to thousands of flight conditions using a three-dimensional multirotor CFD for UAM.

Numerical study to investigate the effect of sample size sensitivity on porous materials with respect to different geometrical parameters

Porous materials play a crucial role in various industrial applications, where precise determination of geometric parameters like dynamic tortuosity, as well as viscous and thermal characteristic lengths, is essential for optimizing acoustical performance. Up to this point, there has been no published size sensitivity test that recommends the ideal sample size for analyzing these parameters in an industrial environment. The present article analyzes different reconstructed porous material samples using conventional CFD and coupled FEA-CFD simulations and suggests optimal sample size and specific simulation setup optimized for industrial purposes, which grant high accuracy and reasonable computational cost. The present paper analyzed several samples, which were reconstructed by micro-CT and analyzed in Star-CCM+ simulation environment. It was found that the optimal sample size (1.5×1.5×1.5 mm or 2.5×2.5×2.5 mm) is different for specific average pore sizes. It was proven through various numerical simulations based on reconstructed porous material samples that running only one coupled CFD-FEM simulation is satisfactory to directly determine all investigated parameters. It was demonstrated that the computationally demanding numerical model could be simplified in specific cases using a rigid body and highlight the limitations.

Exascale Computational Fluid Dynamics in Heterogeneous Systems

Abstract Exascale computing has extended the reach of resolved flow simulations in complex, heterogeneous systems far beyond conventional computational fluid dynamics capabilities. As a result, unprecedented pore and microscale resolution have been achieved in domains that have been traditionally modeled by, and limited to, continuum, effective medium approaches. By making use of computational resources on the new exascale supercomputer, Frontier, at the Oak Ridge Leadership Computing Facility, we performed flow simulations that have pushed the limits of domain-to-resolution ratios by several orders of magnitude for heterogeneous media. Our approach is an incompressible, Navier–Stokes CFD solver based on adaptive, embedded boundary (EB) methods supported by the Chombo software framework for applied partial differential equations (PDEs). The computational workhorse in the CFD application code is an elliptic solver framework in Chombo for pressure-Poisson and viscous, Helmholtz terms that leverages a PETSc-hypre software interface tuned for accelerator-based platforms. We demonstrate scalability of the approach by replicating a unit cylinder packed with microspheres to achieve over 400 × 109 degrees-of-freedom simulated. These simulations model domain lengths of over 20 meters with channel volumes of over 400 cm3 and containing millions of packed spheres with 20 micron grid resolution, challenging current understanding of what it means to be a representative elementary volume (REV) of the continuum scale in heterogeneous media. We also simulate a range of Reynolds numbers to demonstrate wide applicability and robustness of the approach.

Integration and optimization of the after-treatments systems to reduce the acoustic footprint of the ships

Exhaust Gas Cleaning Systems (EGCS) such as scrubbers are mandatory and extensively used to abate SOx in exhaust gases when high sulphur content fuel oil is employed in the marine engine in order to comply with international ship emission regulations, both in new and existing ships. Currently, about 13% of bulker, container, and tanker ships have a scrubber installed, despite the fact that their installation on board is challenging due to their large dimensions to be fitted into the funnel and the complexity of the system, since the exhaust line must control both the chemical and acoustic emissions. In the presented work a combined FEM, CFD simulations and GA optimization methodology aimed at the integration of the abatement system, while optimizing the acoustic properties, is developed. The methodology is first assessed on an industrially-relevant scenario that involves the use of a Genset mock-up equipped with a reference Diesel Oxidation Catalyst (DOC) and a scrubber for the abatement of both NOx and SOx, showing that acoustic performances of the DOC are reliably modelled by the FEM-CFD methodology, which has a significantly reduced computational cost as compared to conventional CFD modelling of acoustic properties. The GA optimization is carried out to improve the DOC acoustic properties showing that it is possible to confer the silencing effect to the after-treatment systems, thus eliminating the traditional silencers from the exhaust line. This leads to a compact exhaust line that integrates the EGCS while maintaining efficient both the chemical pollution abatement capability, and silencing effects to guarantee full compliance (i.e., acoustic and chemical) with the international regulations.

Application of Weighted Least Squares Meshless Method to Incompressible Flows

Meshless methods have been successfully applied to solve a large number of inviscid and viscous compressible flows in both two and three dimensional flow fields. However, the conventional methods of solution for compressible flow equations cannot be applied to incompressible flow equations. This is due to the basic difference in the nature of the pressure variable in the compressible and incompressible flows. Pressure acts as a thermodynamic variable in case of compressible flows and can be easily updated by using the equation of state. In contrast, in case of incompressible flows, the pressure is merely responsible for enforcing the incompressibility constraint and hence cannot be updated by the equation of state. This has led to the development of numerical methods based on Helmholtz Decomposition to solve for the fluid flows in the incompressible flow regimes. Here, we have applied weighted least squares meshless method to solve for the flows in the incompressible flow regime using the projection method based on Helmholtz Decomposition. The standard test cases of two dimensional lid driven cavity flow at Reynolds numbers of 100, 400 and 1000 have been solved using the weighted least squares meshless method. A good match between the results of the test case using the meshless methods and conventional CFD methods for the solution of the incompressible flows has been obtained.

Physics-informed deep learning for simultaneous surrogate modeling and PDE-constrained optimization of an airfoil geometry

We use a physics-informed neural network (PINN) to simultaneously model and optimize the flow around an airfoil to maximize its lift to drag ratio. The parameters of the airfoil shape are provided as inputs to the PINN and the multidimensional search space of shape parameters is populated with collocation points to ensure that the Navier–Stokes equations are approximately satisfied throughout. We use the fact that the PINN is automatically differentiable to calculate gradients of the lift-to-drag ratio with respect to the airfoil shape parameters. This allows us to optimize with the L-BFGS gradient-based algorithm, which is more efficient than non-gradient-based algorithms, without deriving an adjoint code. We train the PINN with adaptive sampling of collocation points, such that the accuracy of the solution improves as the solution approaches the optimal point. We demonstrate this on two examples: one that optimizes a single parameter, and another that optimizes eleven parameters. The method is successful and, by comparison with conventional CFD, we find that the velocity and pressure fields have small pointwise errors and that the method converges to optimal parameters. We find that different PINNs converge to slightly different parameters, reflecting the fact that there are many closely-spaced local minima when using stochastic gradient descent. This method can be applied relatively easily to other optimization problems and avoids the difficult process of writing adjoint codes. It is, however, more computationally expensive than adjoint-based optimization. As knowledge about training PINNs improves and hardware dedicated to neural networks becomes faster, this method of simultaneous training and optimization with PINNs could become easier and faster than using adjoint codes.

Review of vortex lattice method for supersonic aircraft design

AbstractThere has been a renewed interest in developing environmentally friendly, economically viable, and technologically feasible supersonic transport aircraft and reduced order modeling methods can play an important contribution in accelerating the design process of these future aircraft. This paper reviews the use of the vortex lattice method (VLM) in modeling the general aerodynamics of subsonic and supersonic aircraft. The historical overview of the vortex lattice method is reviewed which indicates the use of this method for over a century for development and advancements in the aerodynamic analysis of subsonic and supersonic aircraft. The preference of VLM over other potential flow-solvers is because of its low order highly efficient computational analysis which is quick and efficient. Developments in VLM covering steady, unsteady state, linear and non-linear aerodynamic characteristics for different wing planform for the purpose of several different types of design optimisation is reviewed. For over a decade classical vortex lattice method has been used for multi-objective optimisation studies for commercial aircraft and unmanned aerial vehicle’s aerodynamic performance optimisation. VLM was one of the major potential flow solvers for studying the aerodynamic and aeroelastic characteristics of many wings and aircraft for NASA’s supersonic transport mission (SST). VLM is a preferred means for solving large numbers of computational design parameters in less time, more efficiently, and cheaper when compared to conventional CFD analysis which lends itself more to detailed study and solving the more challenging configuration and aerodynamic features of civil supersonic transport.

Numerical Investigation on the Thermal Protection Characteristics of a New Active Jet Design Parameter for Hypersonic Flight Vehicle

This study investigated the thermal protection performance of an active jet thermal protection system (PRsAJ-TPS) based on a new jet design parameter PRs for hypersonic flight vehicles (HFVs). The new parameter PRs is defined as the relationship between the jet flow total pressure and the free flow total pressure behind the outer bow shock. A 20° tilted nozzle design is employed together with the PRs to form the PRsAJ-TPS. Theoretical and numerical analysis is performed to prove the advantages of using the PRs. A conventional in-house CFD solver with the k ‐ ω SST turbulence model is utilized to perform the calculation. The influence of different PRs (working in the short penetration mode) and flight angles of attack on the performance of the PRsAJ-TPS is also studied. The simulation results confirmed that with a constant PRs, the Mach disk location stays the same and the normalized heat flux reduction is similar at different flight conditions. Nearly linear relationships exist between the new design parameter PRs and thermal protection performance indicators. The PRsAJ-TPS also exhibits good protection when flight angle of attack (AoA) varies between 0° and 40°, with the best results achieved at AoA = 0 ° . This study provides valuable information for the engineering application of the jet-based TPS to future HFVs.

Design and Thermal-Hydraulic Evaluation of Integrated Steam Generator with Serpentine-Tube Configuration for Prevention of Sodium-Water Reaction

Although sodium-cooled fast reactors (SFRs) can have inherently enhanced safety features, there is a concern regarding unexpected events arising from a sodium-water reaction (SWR) in a Rankine cycle steam generator. Innovative concepts of integrated steam generators (ISGs), including an intermediate heat exchanger, have been suggested for eliminating an SWR by employing an intermediate fluid between liquid sodium and steam/water system. In this study, we propose an ISG designed with a novel tube configuration: a serpentine tube-type integrated steam generator (S-ISG). Furthermore, we developed a dedicated one-dimensional thermal-hydraulic analysis code for the specific design evaluation of the S-ISGs. The evaluation results are reported with reasonable explanations and compared with the results obtained using a conventional CFD tool to analyze local thermal-hydraulic characteristics that cannot be shown in the one-dimensional analyses. The proposed S-ISG has an inherent advantage for modularization as well as prevention of SWR; therefore, it is characterized by its ability to improve operational efficiency even in the case of repair, despite the low space efficiency. The results demonstrate that the component developed in this study can serve as an option to maintain cost efficiency while enhancing the public acceptability of an SFR.

Fast and accurate prediction of air temperature and velocity field in non-uniform indoor environment under complex boundaries

The thermal environment in a built indoor space is usually non-uniform in temperature and many other index distributions. Simulation, analysis and understanding the non-uniform indoor environment is critical to a healthy, energy-efficient and comfort indoor space. Conventional CFD model consumes huge computational resources and time for a final result, which may not be suitable for coupling with energy system, and some other applications that needs fast simulation. This study proposes a new model that is built upon Proper Orthogonal Decomposition (POD) method and Radial Basis Function (RBF), with a special Taguchi method served for data preparation. A series of tests are conducted and the results show that the POD-RBF model is fast in computation, accurate in prediction, robust in implementation, and flexible in applications. The construction of the model takes less than 3 s and each case computation takes less than 1 s. The averaged model error in predicting temperature and velocity field is about 0.3% and 8% across different tests.

On the fast fluid dynamics and fractional step methods to predict the coupled indoor temperature and velocity fields

The two-way coupling of temperature and flow fields in the indoor environment makes the conventional CFD simulation computationally intensive, especially when the transient evolution of temperature and velocity fields is of interest in areas such as unsteady ventilation. This study, by applying different splitting CFD methods (fast fluid dynamics (FFD) and fractional step (FS) method) to solve the momentum equation to decouple velocity and pressure fields, further explored different strategies to tackle the coupled temperature equation, and validated the different sequences of operation based on different combinations of temperature/velocity solution schemes. Each operational sequence of different combined numerical schemes was validated on a mixed convection airflow benchmark case with a heated floor. The results showed that different operational sequences using the numerical schemes which considered the strong coupling effect of temperature and flow fields could predict the final steady-state result well, while for the transient simulation that need to track the physical flow/thermal fields variation process precisely, the operational sequence which introduce the effect of buoyancy into volumetric flux to solve pressure Poisson equation showed better accuracy and owned the capability of predicting the same transient evolution process as conventional CFD method. Based on the comprehensive comparison, several promising numerical schemes under proper operational sequence were recommended and the highest computing efficiency was provided by sequence D of RIPC + Iterative which saved 68% computing time compared to the conventional CFD method.

Conditional Generative Adversarial Networks for modelling fuel sprays

In this study, the probabilistic, data driven nature of the generative adversarial neural networks (GANs) was utilized to conduct virtual spray simulations for conditions relevant to aero engine combustors. The model consists of two sub-modules: (i) an autoencoder converting the variable length droplet trajectories into fixed length, lower dimensional representations and (ii) a Wasserstein GAN that learns to mimic the latent representations of the evaporating droplets along their lifetime. The GAN module was also conditioned with the injection location and the diameters of the droplets to increase the generalizability of the whole framework. The training data was provided from highly resolved 3D, transient Eulerian–Lagrangian, large eddy simulations conducted with OpenFOAM. Neural network models were created and trained within the open source machine learning framework of PyTorch. Predictive capabilities of the proposed method was discussed with respect to spray statistics and evaporation dynamics. Results show that conditioned GAN models offer a great potential as low order model approximations with high computational efficiency. Nonetheless, the capabilities of the autoencoder module to preserve local dependencies should be improved to realize this potential. For the current case study, the custom model architecture was capable of conducting the simulation in the order of seconds after a day of training, which had taken one week on HPC with the conventional CFD approach for the same number of droplets (200,000 trajectories).

Rapid prediction of airborne gaseous pollutant transport in aircraft cabins based on proper orthogonal decomposition and the Markov chain method

Rapid prediction of airborne gaseous pollutant transport is important for designing a safe indoor environment. Current models generally solve the airflow fields by CFD first and then predict the transport of a pollutant in fixed airflow patterns. Every time the air-supply parameters are adjusted, the airflow field must be re-solved by CFD, which is time-consuming. This study proposed a model to improve the prediction efficiency. The model first applies proper orthogonal decomposition to the sampled airflow fields, to construct a database related to all the airflow fields in the sample ranges, and then uses the Markov chain method to obtain the airflow field with the desired air-supply parameters for construction of a transport probability matrix. Finally, the airborne gaseous pollutant transport can be predicted quickly in the fixed airflow pattern. The proposed model was applied to an aircraft cabin model, first with a single gaseous pollutant source and then with two sources, for validation of the proposed model. The results show that the proposed model can predict both the airflow field and the transport of a gaseous pollutant with outcomes similar to those obtained by the conventional CFD method, but with a much shorter computing time. When the database has been prepared in advance, the use of the model reduces the computing time by more than 90%. Further improvement of the proposed model in terms of accuracy and extension of the model to prediction of pollutant transport within unsteady airflow fields will be the main objectives of future work.

A novel hybrid framework coupling the discrete velocity based-method, the particle-based method and the Grad's distribution function-based method for supersonic rarefied flows

Supersonic rarefied flow has an intrinsically multiple-scale nature due to the large variations of density and characteristic length of the flow structures. Although particle-based methods such as direct Monte Carlo simulation (DSMC) present efficient performance in supersonic rarefied flows, the statistical noise as the nature of stochastic method causes a large amount of averaging steps to be required. To minimise the contamination of statistical noise and preserve the efficiency of capturing the multiscale rarefied effects, a finite volume based hybrid framework is proposed to couple the discrete velocity method (DVM), the particle-based method and the Grad's distribution function-based moment gas kinetic method (G13-MGKS). Benefiting from the coupled treatment both in the spatial and velocity space, the DVM-based method is applied to low-velocity areas close to the wall, which greatly eases the statistical noise and numerous steps of statistical averages can be avoided. Besides, the partition of different sub-methods conveniently determined by the initial flow field from conventional CFD solver has been verified without compromising any accuracy. Further applying the Grad's distribution function-based method in the area with moderate rarefied effect, the present hybrid DVM-Particle-Grad (HDPG) method exhibits the high efficiency and achieves a tenfold speedup ratio compared to the DVM.

Implementation of a velocity decomposition method coupled with volume-of-fluid method for simulating free-surface flows

An efficient velocity decomposition method (VDM) is proposed for simulating free surface flows based on OpenFOAM platform, which is primarily intended for marine hydrodynamics applications. In this method, it is assumed that the free surface can be treated as a source region, which introduces rotating velocity related to the wave elevation. In the implemented solver called cpVDFoam, the decomposed potential flow velocity is computed by solving the Laplace equation without considering the effect of free surface, and the remaining non-potential velocity is determined with a complimentary Navier–Stokes (N-S) equation. Further, the free surface is modeled using volume-of-fluid (VOF) method in the viscous domain for obtaining the non-potential velocity. Flow field analysis shows that non-potential velocity magnitude contributed by freesurface is positive correlation with wave elevation, and which value shall decrease to zero in the zero wave elevation region. The efficacy of the proposed VDM is validated by comparing the simulation results with the results obtained using standard OpenFOAM solver interFoam and earlier reported experiment. All results indicate that the present method exhibited similar accuracy as the conventional CFD method but less sensitive to the domain size.

Prediction of Erotion Rate in Two Elbows for Coal-Air Flow Based on Computational Fluid Dynamics Simulation

Corrosion is one of the problems cause the material fail, leaks in the equipment and pipe systems in the industry which caused distrub in productivity of production in industry. There are some ways to prediction erosion in pipes, with the conventional or manual and CFD. CFD is to prediction erosion in turbulent air flows with variations in coal particle size and velocity of flowing air. The results of prediction erotion based on CFD is numerical and graphic, by displayed numerical results into visual results. These results include about the velocity, pressure, and temperature distribution in the air flow. So, CFD can be used to prediction pipe because it has wide coverage and effective in display numerical and visual results. Discreate Phase Model is the model of CFD with a variable stream velocity of 10 m/s, 15 m/s, 20 m/s and a solid loading of 5%, 10%, 15%, 20%, 25%. From the results, the maximum erosion rates increased with increasing sizes of particles and stream velocities. However, the location of the maximum erosion rate in the elbows region was independent on those mentioned parameters.

Multi-objective optimization of TPMS-based heat exchangers for low-temperature waste heat recovery

The transformation to a truly sustainable energy system will require taking better advantage of the waste heat. Integrating heat exchangers with the triply periodic minimal surface (TPMS) is a promising and efficient way to build waste heat recovery systems that harness heat emissions from the low pitch thermal systems. This is mainly due to the low hydrodynamic resistance and pressure drop in the TPMS while securing good heat transfer at low-temperature gradient. This study establishes a computational design and analysis of heat and mass transfer inside a heat exchanger based on the TPMS structure and determine thermal effectiveness, heat transfer coefficient, and pressure drop inside the channel. The non-linearity dependence of results to several design variables makes obtaining the optimal design configuration solely using conventional CFD or experimental study nearly impossible. Hence, a multi-objective optimization workflow based on a Genetic Algorithm for laminar flow is employed to reveal the underlying relationships between design variables for the optimal configurations. The results illustrate the local sensitivity of important parameters such as the heat transfer coefficient, Nusselt number, and thermal performance of the heat exchanger against various design variables. It is shown that the pressure drop is directly affected by gas inlet velocity, viscosity, and density, from high to low, respectively. The Pareto frontiers for the optimal thermal performance are extracted, and the correlation between design objectives is determined. This methodology provides a promising framework for heat exchangers’ design analysis, including multi-objective goals and design constraints.