Local Flow Characteristics Research Articles

Thermal Stratification Prediction in Reactor System Based on CFD Simulations Accelerated by A Data-driven Coarse-grid Turbulence model

Thermal stratification in large enclosures is an integral phenomenon to nuclear reactor system safety. Currently, the effective model for thermal stratification utilizes a multi-scale method that integrates 1-D system-level and 3-D CFD code, which offers thermal stratification details while supplying system-level data across various domains. Nonetheless, harmonizing two codes that operate on different spatial and temporal scales presents a significant challenge, with high-resolution CFD simulations requiring substantial computational resources. This study introduced a data-driven coarse-grid turbulence model based on local flow characteristics at a significantly coarser scale targeting improved efficiency and accuracy in reactor safety analysis concerning thermal stratification. A machine learning framework has been introduced to expedite the RANS-solving process by coupling of OpenFOAM and TensorFlow, which entails training a deep neural network with fine-grid CFD-generated data to predict turbulent eddy viscosity. The feasibility of the developed data-driven turbulence model was proven through the SUPERCAVNA experimental facility problem validation.

Study of Flow Characteristics and Anti-Scour Protection Around Tandem Piers Under Ice Cover

The impact of an ice-covered environment on the local flow characteristics of a bridge pier was studied through a series of flume tests, and the dominant factors affecting the scour pattern were found to grasp the change laws of the local hydrodynamic characteristics of the bridge pier under the ice cover. At the same time, because the scour problem of the pier foundation is a technical problem throughout the life-cycle of the bridge, to determine the optimal anti-scour protection effect on the foundation of the bridge pier, active protection scour plate was used to carry out scour protection tests, and its structural shape was optimized to obtain better anti-scour performance. The test results show that the jumping movements of sediment particles in the scour hole around the pier are mainly caused by events Q2 and Q4, which are accompanied by events Q1 and Q3 and cause the particle rolling phenomenon, where Q1 and Q3 events are outward and inward interacting flow regimes, and Q2 and Q4 events are jet and sweeping flow regimes, respectively. The power spectral attenuation rate in front of the upstream pier is high without masking effects, while strong circulation at the remaining locations results in strong vorticity and high spectral density, in particular, when the sampling time series is 60 s (i.e., f = 1/60), the variance loss rates under ice-covered conditions at the front of the upstream pier, between the two piers, and at the tail end of the downstream pier are 0.5%, 4.6%, and 9.8%, respectively, suggesting a smaller contribution of ice cover to the variance loss.

Insights into the relationship between particulate flow characteristics and local erosion behaviour under waterjet: The role of particle-fluid-surface interaction

In this study, the erosion characteristics of eutectic high entropy alloy under the liquid–solid two-phase flow is investigated by a coupled numerical-experimental method, in order to reveal the intrinsic relationship between microscopic erosion mechanism (experimental test) and the related particle-fluid-surface interaction (CFD modelling). Furthermore, instead of the common relation between average erosion rate and mainstream flow velocity, the relationship between local erosion distribution and local particulate flow characteristics is clarified. The erosion rate and erosion pattern match well between the experiment and simulation. The results demonstrate that NiCoCrFeNb0.45 exhibits superior anti-erosion performance when compared to other widely used equipment metals. The erosion profile agrees well with the shape of surface velocity distribution, indicating a close relationship between surface erosion behaviour and flow characteristics. In particular, the erosion pattern at a normal angle appears as a symmetric ring due to the flow stagnation phenomenon, with slight erosion damage in the centre area and severe erosion in the surrounding, whereas the erosion profile at an oblique angle displays an elliptic shape, with severe erosion damage in centre. Notably, the primary erosion mechanism varies dramatically in different surface regions, induced by the various impact velocities and angles associated with the corresponding changes in the flow field. This study provides a deeper understanding of erosion behaviour in terms of the interrelationship among the material mechanical properties, particulate flow field and particle-surface impingement behaviour.

Analysis of the exergy transfers in subcritical and transcritical CO[formula omitted] ejectors and their connection with the performance of the refrigeration cycle

Ejector refrigeration systems (ERSs) operating with carbon dioxide are notoriously challenging from both design and operation perspectives. In particular, the ejector itself gives rise to special attention since it directly affects the performance of the global system. In this study, the connections between local exergy transfers within the ejector and global cycle performance are investigated by using a novel multi-scale numerical approach. Computational Fluid Dynamics (CFD) is used to generate accurate predictions of the ejector operation with access to the local flow and heat transfer features. The exergy tube analysis is applied onto the simulation results, linking local transfers within the ejector to the results at the component- and cycle-scale. A state-of-the-art ejector thermodynamic model is then calibrated onto the CFD results, and employed to determine the physically possible operation of the ERS. In addition, the metastability impact on the exergy transfers and overall system performance is investigated by comparing the CFD results of the classical Homogeneous Equilibrium Model with those obtained with the Homogeneous Relaxation Model. Notably, it is found that at fixed ejector inlet conditions, there exits an outlet pressure that maximizes the exergy transfers from primary to secondary streams. Interestingly, this work shows for the first time that the maximum exergy gain of the secondary stream appears to correlate with the maximum coefficient of performance of the system, independently of the on- or off-design operation of the ejector. Moreover, metastability generally deteriorates the performance at both scales. The results of the present study serve to pave the way towards better ejector design objectives.

Dynamic characteristics of the hydrogen injector-ejector unit in a PEM fuel cell system

The ejector, as a fluid-dynamics controlled passive component, is a promising hydrogen recirculation device in proton exchange membrane (PEM) fuel cell systems. Nevertheless, its performance is significantly affected by dynamic operating conditions and the behavior of the injector. This study investigates the dynamic performance of an ejector integrated with a hydrogen injector in a 100 kW PEM fuel cell system. A dynamic computational fluid dynamics (CFD) model of the integrated injector-ejector unit is developed using a dynamic mesh method and validated through theoretical analysis and experimental data. A thorough analysis of the global performance and local fluid flow characteristics of the injector-ejector unit is conducted. The results show that the hydrogen injector, with a throat diameter of 2.00 mm, exhibits an adjustable linear flow range from 0 to 1.62 g/s under an inlet pressure of 2.0 MPa. Additionally, the primary mass flow rate can be linearly regulated by adjusting the valve gap, while the secondary mass flow rate exhibits asynchronous behavior with the primary flow. The velocity and temperature distributions within the injector-ejector unit undergo significant changes under dynamic conditions. These findings contribute to optimizing the design and control strategies of hydrogen supply components in PEM fuel cell systems.

A Comparison of Local and Global Strategies for Exploiting Field Inversion on Separated Flows at Low Reynolds Number

The prediction of separated flows at low Reynolds numbers is crucial for several applications in aerospace and energy fields. Reynolds-averaged Navier–Stokes (RANS) equations are widely used but their accuracy is limited in the presence of transition or separation. In this work, two different strategies for improving RANS simulations by means of field inversion are discussed. Both strategies require solving an optimization problem to identify a correction field by minimizing the error on some measurable data. The obtained correction field is exploited with two alternative strategies. The first strategy aims to the identification of a relation that allows to express the local correction field as a function of some local flow features. However, this regression can be difficult or even impossible because the relation between the assumed input variables and the local correction could not be a function. For this reason, an alternative is proposed: a U-Net model is trained on the original and corrected RANS results. In this way, it is possible to perform a prediction with the original RANS model and then correct it by means of the U-Net. The methodologies are evaluated and compared on the flow around the NACA0021 and the SD7003 airfoils.

Inertial oscillation modelling of low-level jets: an application to the complex terrain and double-nosed wind profiles

This study investigates the role of inertial oscillations in the evolution of a nocturnal Low-Level Jet (LLJ) in complex terrain and explores the impacts of local perturbations on wind dynamics. Specifically, a conceptual model based on inertial oscillations (Van de Wiel et al. J Atmos Sci 67(8):2679–2689 (2010)) is used to replicate the evolution of an LLJ in a gentle-sloping valley ensuring to capture its long-period dynamics under weak synoptic forcing. The analysis is performed on an already-analysed case study from the Mountain Terrain Atmospheric Modeling and Observations (MATERHORN) data set, taking advantage of the known local flow characteristics and the existence of a temporary anomaly in the LLJ shape called double-nosed LLJ. In an attempt to capture this last flow feature, a model modification is introduced, revealing appropriate to capture the double-nosed shape of the LLJ. Further observational studies will be needed to corroborate the operational use of this model and explore its application potential in different wind and energy sectors.

Hydrodynamic Analysis of Pipelines Transporting Capsule for Onshore Applications

Abstract— Rapid depletion of energy resources hasimmensely affected the cargo transportation industry.Efforts have been made to develop newer economic andenvironmental friendly modes for this purpose. Bulk solidscan be transported for long distances effectively in pipelines.Raw materials can be stored in spherical containers(commonly known as capsules) transported through thepipeline. For economical and efficient design of anytransportation mode, both the local flow characteristics andthe global performance parameters need to be investigated.Published literature is severely limited in establishing theeffects of local flow features on system characteristics ofHydraulic Capsule Pipelines (HCPs). The present study focuses on using a well-validatedComputational Fluid Dynamics (CFD) based solver tosimulate the unsteady turbulent flow of a sphericalcapsule as a simplified case in HCPs for onshoreapplications. A novel numerical model has beenemployed in the present study with the aid of thedynamic mesh technique for calculating the pressureand the velocity variations within such pipelines. Thenumerical model comprises a straight test section. Thenumerical model for capsule flow yields realisticresults for the global flow parameters as compared tothe experimental data from the test rig developed inthe present study. The influence of the carrying fluid(water) on the capsule has been verified by estimatingthe shear forces and the friction coefficient. Inaddition, novel models have been developed forpredicting the wet friction coefficient considering thestart-up effect in such systems.

Local characteristics of bubbly flow in rods assembly 3 × 3

This study focuses on the experimental investigation of local characteristics of upward bubble flow in a 3 × 3 square rod bundle assembly. The experiments were conducted at Reynolds numbers ranging from 4000 to 11000 and gas void fractions up to 10%. The distribution of the gas phase around the central rod of the assembly was measured using a frontal conductivity probe, while the wall shear stress on the central rod was obtained through the electrodiffusion method. It is shown that the gas phase distribution does not exhibit the symmetry characteristic of the assembly cross-section. Furthermore, as the distance from the spacer grid increases, a reconfiguration of void fraction profiles is observed. This explains the non-monotonic dependence of wall shear stress and wall shear stress fluctuations on the distance to the spacer grid.

Enhancing the SST turbulence model for predicting heat flux in hypersonic flows through symbolic regression

In the context of hypersonic flows, shock-wave/turbulent boundary-layer interactions (SWTBLIs) can lead to substantial aerodynamic heating. The commonly used Reynolds-averaged Navier–Stokes (RANS) method in engineering applications faces challenges in accurately predicting heat transfer in these conditions, as the assumptions made by Morkovin’s assumption in the RANS method are not applicable in hypersonic SWTBLI flows. This article addresses these challenges by introducing a variable turbulent Prandtl number (Prt) model for non-adiabatic walls, established through field inversion and symbolic regression (SR) techniques. The methodology begins with field inversion for an oblique SWTBLI case at Mach 5. The corrected field, denoted as β(x), obtained from this inversion, is integrated with selected local flow characteristics to derive a corrected expression using SR. Following the necessary adjustments, this expression is incorporated into the RANS solver. Various cases with different wall cooling ratios, Mach numbers, and Reynolds numbers are chosen to assess the generalization ability of the variable Prt model. The outcomes demonstrate a substantial improvement in the model’s ability to predict heat transfer in hypersonic SWTBLI flows without compromising the baseline model’s performance in predicting the undisturbed boundary layer. The correction effect remains notably enhanced even in complex three-dimensional cases.

Numerical and experimental study of homogenization mechanism of high shear rotor‐stator mixer

AbstractUtilizing computational fluid dynamics (CFD) for analytical purposes, this study developed a fundamental model employing the multiple reference frame (MRF) method, facilitated by the CFX simulation platform. The investigation conducted numerical simulations of the flow field within the rotor‐stator assembly of a high shear mixer, guided by the Navier–Stokes equations and the standard k‐ε turbulence model. To quantify the homogenization efficacy of the high shear mixer under scenarios with and without energy consumption considerations, the study introduced two distinct parameters: the mixing index (γ) and the energy ratio mixing index (λ). The impact of structural parameters such as the number of rotor‐stator teeth, radial clearance, and tooth apex‐base axial clearance on the local flow characteristics—velocity, pressure, turbulent kinetic energy, shear rate distribution, net power consumption, and the specified indices—was meticulously analyzed. This analysis aimed to identify the optimal structural configurations for the mixer, considering energy efficiency and mixing effectiveness, and to determine the relative influence of these structural variations on the mixing index (γ) and the energy ratio mixing index (λ).

Coupling Global Parameters and Local Flow Optimization of a Pulsed Ejector for Proton Exchange Membrane Fuel Cells

Proton exchange membrane fuel cells (PEMFCs), as an important utilization of hydrogen energy, contribute to the sustainable development of global energy. Pulsed ejectors have a high potential for improving the hydrogen utilization of PEMFCs in the full operating range by circulating unconsumed hydrogen. In this study, a pulsed ejector applied to a 120 kW fuel cell was designed, and the flow characteristics were analysed using computational fluid dynamics (CFD). Based on the data from the CFD model, the global optimization of the ejector was carried out using the Gaussian process regression (GPR) surrogate model and the grey wolf optimization (GWO) algorithm. The local structure was then further optimized using an adjoint method coupling streamlining modification that takes into account the local flow characteristics. The CFD results showed that, under a fixed structure, increasing the pressure difference between the secondary flow and the ejector outlet would promote boundary layer separation, shorten the shockwave chain length, change the effective flow area of the secondary flow, and lower the entrainment ratio (ER). The analytical results from the GPR model indicated significant interactions among the structural parameters. The globally optimized ejector using GPR and GWO improved the hydrogen entrainment ratio from 1.42 to 3.12 at the design point. Furthermore, the results of streamlining local optimization show that the entrainment ratio increased by 1.67% at the design point and increased by up to 3.99% over the full operating range compared to the optimized ejector by global optimization.

Field inversion machine learning augmented turbulence modeling for time-accurate unsteady flow

Field inversion machine learning (FIML) has the advantages of model consistency and low data dependency and has been used to augment imperfect turbulence models. However, the solver-intrusive field inversion has a high entry bar, and existing FIML studies focused on improving only steady-state or time-averaged periodic flow predictions. To break this limit, this paper develops an open-source FIML framework for time-accurate unsteady flow, where both spatial and temporal variations of flow are of interest. We augment a Reynolds-Averaged Navier–Stokes (RANS) turbulence model's production term with a scalar field. We then integrate a neural network (NN) model into the flow solver to compute the above augmentation scalar field based on local flow features at each time step. Finally, we optimize the weights and biases of the built-in NN model to minimize the regulated spatial-temporal prediction error between the augmented flow solver and reference data. We consider the spatial-temporal evolution of unsteady flow over a 45° ramp and use only the surface pressure as the training data. The unsteady-FIML-trained model accurately predicts the spatial-temporal variations of unsteady flow fields. In addition, the trained model exhibits reasonably good prediction accuracy for various ramp angles, Reynolds numbers, and flow variables (e.g., velocity fields) that are not used in training, highlighting its generalizability. The FIML capability has been integrated into our open-source framework DAFoam. It has the potential to train more accurate RANS turbulence models for other unsteady flow phenomena, such as wind gust response, bubbly flow, and particle dispersion in the atmosphere.

Experimental study on the effect of sound waves on the heat transfer characteristics of heated pipes

This study has experimentally investigated the heat transfer characteristics of a heated steel pipe (diameter = 20 mm) placed horizontally in a vertical experimental cylindrical pipe under longitudinal sound waves. The ranges of sound parameters used in this study are as follows: sound frequency f = 210–1000 Hz, and sound pressure level SPL = 125–135 dB. The heat transfer coefficient and Nusselt number of the cross- section of the heated pipe are used to describe the influence of different sound parameters on the heat transfer behavior of the heated pipe. The results show that under the action of low-frequency strong sound wave (f < 230 Hz), the distribution of local Nusselt number of heated pipe decreases (0-90°) at first and then increases (90-180°), while under the action of high-frequency strong sound wave (f > 500 Hz), the distribution of local Nusselt number first increases and then decreases. The relative Nusselt number decreases exponentially with the increase of the Strouhal number. In addition, the oscillating flow induced by sound waves is greatly influenced by the natural convection. The local enhancing or weakening effects of strong sound waves with different parameters on the heat transfer in a hot steel pipe are related to the characteristics of local flow.

Assessment of Machine Learning Wall Modeling Approaches for Large Eddy Simulation of Gas Turbine Film Cooling Flows: An a Priori Study

Abstract In this work, a priori analysis of machine learning (ML) strategies is carried out with the goal of data-driven wall modeling for large eddy simulation (LES) of gas turbine film cooling flows. High-fidelity flow datasets are extracted from wall-resolved LES (WRLES) of flow over a flat plate interacting with the coolant flow supplied by a single row of 7-7-7 shaped cooling holes inclined at 30 degrees with the flat plate at different blowing ratios (BR). The WRLES are performed using the high-order Nek5000 spectral element computational fluid dynamics (CFD) solver. Light gradient boosting machine (LightGBM) is employed as the ML algorithm for the data-driven wall model. Parametric tests are conducted to systematically assess the influence of a wide range of input flow features (velocity components, velocity gradients, pressure gradients, and fluid properties) on the accuracy of ML wall model with respect to prediction of wall shear stress. In addition, the use of spatial stencil and time delay is also explored within the ML wall modeling framework. It is shown that features associated with gradients of the streamwise and spanwise velocity components have a major impact on the prediction fidelity of wall model, while the effect of gradients of wall-normal velocity component is found to be negligible. Moreover, adding flow feature information from an x-y-z spatial stencil significantly improves the ML model accuracy and generalizability compared to just using local flow features from the matching location. Overall, highest prediction accuracy is achieved when both spatial stencil and time delay features are incorporated within the data-driven wall modeling paradigm.

A Fibrin-Thrombin Based In Vitro Perfusion System to Study Flow-Related Prosthetic Heart Valves Thrombosis

Prosthetic heart valve (PHV) replacement has increased the survival rate and quality of life for heart valve-diseased patients. However, PHV thrombosis remains a critical problem associated with these procedures. To better understand the PHV flow-related thrombosis problem, appropriate experimental models need to be developed. In this study, we present an in vitro fibrin clot model that mimics clot accumulation in PHVs under relevant hydrodynamic conditions while allowing real-time imaging. We created 3D-printed mechanical aortic valve models that were inserted into a transparent glass aorta model and connected to a system that simulates human aortic flow pulse and pressures. Thrombin was gradually injected into a circulating fibrinogen solution to induce fibrin clot formation, and clot accumulation was quantified via image analysis. The results of valves positioned in a normal versus a tilted configuration showed that clot accumulation correlated with the local flow features and was mainly present in areas of low shear and high residence time, where recirculating flows are dominant, as supported by computational fluid dynamic simulations. Overall, our work suggests that the developed method may provide data on flow-related clot accumulation in PHVs and may contribute to exploring new approaches and valve designs to reduce valve thrombosis.

Simulation of supersonic axisymmetric base flow with a data-driven turbulence model

Axisymmetric base flow involves massive flow separation which is challenging for a typical Reynolds-averaged Navier-Stokes (RANS) turbulence model. Furthermore, a supersonic condition imposes additional difficulties to a turbulence model in predicting strong compressibility effects associated with the flow separation phenomena. A data-driven approach is pursued here to improve a turbulence model. The Spalart-Allmaras model is corrected from an optimization process, and then the model correction is mapped to local flow features through a machine learning process. For the optimization-combined machine learning method, the field inversion and machine learning framework is adapted here with three major modifications. First, a zonal-field inversion method is suggested to avoid the undesired modification of attached flow in the optimization process for the model correction. Second, the RANS model is modified by adjusting a net balance between the production and destruction terms in the RANS equation. Third, the local Mach number is included in the machine learning process to incorporate compressibility effects of the supersonic flow. The current study indicates that the improved RANS model reduces the eddy viscosity in the separated region, compared to the baseline RANS model. The reduction of the eddy viscosity helps to predict the base expansion of the supersonic flow and the size of the recirculation region. Current computations are compared with relevant literature data including wake velocity profiles and the base pressure in the wide range of the supersonic Mach number.

Experimental investigation on local flow structures of upward cap-bubbly flows in a vertical large-size square channel

Gas–liquid two-phase flows are typical in engineering systems involving heat and mass transfers. The cross-sectional geometries and sizes of these flow channels vary, exerting a notable impact on the thermofluid behavior. To develop thermofluid models, it is imperative to obtain local measurements of two-phase flows and gain insights into their characteristics from these measurements. Numerous experimental studies have investigated the local flow characteristics in circular pipes and rectangular channels; however, few studies on square channels have been conducted, particularly for flow regimes beyond bubbly flows, such as cap-bubbly flows. Given the importance of two-phase flows in large square channels for advanced nuclear reactors, such as the economic simplified boiling water reactor (ESBWR), we experimented with upward cap-bubbly flows in a large square channel. Detailed cross-sectional distributions of the void fractions, axial gas velocities, and interfacial area concentrations for two bubble-size groups were obtained at three axial locations using local measurements of a four-sensor optical probe. Based on the database, the cap-bubbly flow characteristics were understood, including flow development in a large square channel. In addition, the existing drift-flux correlations for large circular pipes predicted the measured void fraction with an accuracy of approximately ± 10 %, whereas the existing interfacial area concentration correlations reasonably correlated with the measured interfacial area concentration with an accuracy of approximately ± 30 %. Furthermore, the database was used to calculate the covariances of the void fractions. The correlations for large circular pipes significantly underestimated the void fraction covariance of the larger bubble group, probably because of the peculiar void fraction distribution of the group in a large square channel. The other covariances were well predicted.

DUST COLLECTION EFFICIENCY UNDER IN-PHASE TURBULENCE CONDITIONS IN THE PRESENCE OF PHASE TRANSITIONS

The article presents the results of a study of the dust collection efficiency in a regular packing layer of a rotoclone-type hybrid apparatus on the particle inertia index in the presence of phase transition processes. The effect of phase transitions on the efficiency of gas purification processes under inphase turbulence conditions is considered. Based on the conditions of stability of the in-phase vortex formation regime of the gas phase in a regular packing layer, the dominant influence of turbulent diffusion mechanisms on the intensity of phase transition processes and, as a consequence, on the dust collection efficiency has been established. In this case, the predominant influence is in the area of highly dispersed and finely dispersed dust fractions. The results of the experimental studying the patterns of dust collection efficiency depending on the particle inertia index in the presence of phase transition processes are presented. The studies were carried out for three phase transition modes: condensation of vaporsfrom a vapour-gas mixture, evaporation of a spraying liquid as well as a neutral mode. Using A.N. Kolmogorov’s hypothesis,the basic equations and results of calculating the numerical values of local characteristics of a turbulent gas flow in a regular packing layer are presented under the assumption of isotropic turbulence realized in a layer of a regular packing, whose local characteristics do not depend on coordinates and direction.

Numerical study and machine learning on local flow and heat transfer characteristics of supercritical carbon dioxide mixtures in a sinusoidal wavy channel PCHE

The ambient temperature has a great influence on the practical application of the supercritical carbon dioxide(S-CO2) Brayton cycle. The introduction of mixtures is an effective way to change the critical characteristics of the working fluid, which allows the system to better match the ambient temperature. The mixtures not only affect the overall performance of the power system, but also affect the flow and heat transfer processes in various heat exchange equipment. In this investigation, the physical model of a sinusoidal wavy channel printed circuit heat exchangers (PCHE) is built, and numerical simulations are conducted firstly on the flow and heat transfer characteristics of three S-CO2 mixtures (CO2-He, CO2-Ke and CO2-Xe). Effects of the mole fraction and type of the additive gasses on the thermal-hydraulic performance are discussed, as well as the variations of the local heat transfer coefficients and pressure drop along the flow direction. It's found that transfer coefficient and pressure drop of CO2-He increase with the increase of the additive gas mole fraction n, with maximum heat increases of 396 % and 1147 % against pure S-CO2, respectively. With the increase of n, heat transfer coefficient and pressure drop of CO2-Kr and CO2-Xe decrease. Maximum decreases of 71 % and 38 % can be seen for the heat transfer coefficient and pressure drop against pure CO2 for CO2-Kr, and 80 % and 64 % against pure CO2 for CO2-Xe, respectively. CO2-Xe yields the best thermal-hydraulic performance of the PCHE among the three mixtures. Secondly, machine learning is adopted to predict the local flow and heat transfer characteristics of S-CO2 mixtures in response to the significant variation of S-CO2 mixtures along the flow direction within the PCHE. Four machine learning models including Support vector machine (SVR), Artificial neural network (ANN), Random forest (RF) and Extreme Boosting Tree (XGBoost) are used and the corresponding prediction performance are analysed. It is found that machine learning is efficient and accurate in the prediction. Among the four machine learning models, XGBoost has a strong fitting ability to the local Nu and f, with an R2 of 0.9992 and 0.9718 on the test set, respectively. The prediction results of the XGBoost model can reflect the variations of local Nusselt number and friction factor along the flow direction well. However, ANN has stronger generalization ability in the prediction under new working conditions. The use of machine learning can greatly help the design and optimization of PCHEs with S-CO2 mixtures as working fluids.