Simulation of Turbulence and RoughnEss in Additive Manufactured parts

Periodic Reporting for period 2 - STREAM (Simulation of Turbulence and RoughnEss in Additive Manufactured parts)

Reporting period: 2021-04-01 to 2022-09-30

Additive manufacturing (AM) process offers tremendous gains over conventional subtractive manufacturing in heat exchanger design, key issue of thermal engine efficiency. The STREAM project aims at designing novel modeling strategies for the performance prediction of additive-manufactured heat exchangers. The consortium consists of two laboratories CNRS-CORIA and CNRS-LEGI, which have a long experience in high-fidelity multi-physics turbulent flow modeling and TEMISTh, a SME which develops customized solutions for heat exchangers. From the fluid dynamics point of view, AM often introduces important wall roughness, which depends strongly on the manufacturing process itself, and which impacts heat transfer and pressure loss across the device. It is therefore mandatory to design Computational Fluid Dynamics (CFD) models with a sufficient level of accuracy to predict the performances of heat exchangers. RANS (Reynolds-Averaged Navier-Stokes) and LES (Large-Eddy Simulation) are two complementary turbulence modeling approaches that are good candidates for such challenge. In these approaches, wall modeling often relies on statistical analysis, leading to law-of-the-wall models that are widely used in the prediction of internal flows. However, these models need to be extended and validated for wall roughness generated by additive manufacturing. To this aim, STREAM built a large database of high-fidelity roughness-resolving Large-Eddy Simulations that have been analyzed to derive a hierarchy of well-parametrized statistical wall models. An original stochastic wall model has been derived based on the analysis of the total shear stress distribution at the wall. The resulting statistical and stochastic models, usable in both roughness-modeled RANS and LES approaches, have been extensively validated a priori by comparison with the high-fidelity database and a posteriori on classical heat exchanger applications: Fuel-Cooled Oil Cooler, Air-Cooled Oil Cooler, Surface Air-Cooled Oil Cooler.

The first part of the project was dedicated to the first objective dealing with the identification of the key physical parameters and the best methodology to model the turbulence especially in rough additive manufactured parts (like tubes, plate, fins...). The key physical statistical moments that define a rough surface have been identified and categorized from an extensive literature review. From this review, parameters such as the effective slope (ES) are very important for the modelling of the impact of roughness on the boundary layer. For the second objective, which deals with the improvement of rough-wall modelling based on resolved-rough unsteady simulations, a rough-surface generator (RSG) and analyzer, and a body-fitted mesh generator have been created and included in an automated simulation workflow to build a large resolved-roughness (RR) Large-Eddy Simulation (LES) database. This database has been extensively post-processed and analyzed in order to derive a hierarchy of statistical and stochastic rough-wall models. These models have been validated a priori by comparing with the RR-LES database and a posteriori by performing RANS and LES simulations of various heat exchanger configurations. The RANS simulations were first performed with standard rough-wall models to have a reference and then with the modified model. Finally, a design methodology for additive-manufactured heat exchangers was derived which enables to include the new modelling approaches. With this methodology, three different representative heat exchangers were designed from parameters given by the topic manager.

These results were mainly disseminated through publications (Journal of Turbomachinery), participation to international conferences (ASME 2022, ETMM13, FFHMT22) and training courses (VKI lecture series). The exploitation of these results will be done through software patent (YALES2) and following projects focused on the modelling and design of additively manufactured heat exchangers for the new generation of aero-engines.

The expected impact as planned in the description of work is listed below.
1. Definition of the predominant physical parameters impacting the predictive character of CFD simulations => achieved
2. Assessment of the existing wall models to predict pressure loss and heat flux for heat exchangers simulations for both RANS and LES approaches => achieved
3. Improvement of wall models to consider both roughness and complex geometries effects => achieved
4. Development of a roughness structure and mesh generation tool for numerical LES-RANS benchmarks => achieved and integrated in YALES2 code
5. Optimization of the architecture of the exchanger regarding the objective of dimensioning => achieved with a dedicated methodology
6. Local optimization of the flow and heat transfer considering the orientation of the walls, and therefore their roughness => partially achieved as optimization is not local yet
7. Adaptation of the physical model to the type of heat exchanger => not needed in the STREAM configurations

Beyond to these expected outcomes, the STREAM project results had a greater impact, which has led to new projects, new collaborations and new topics in line with STREAM:
– 2022-2025, NEMO project (TEMISTH, IRT M2P) funded by PIA (Plan d’Investissement d’Avenir) in France: investigation of post-processing techniques on performances of heat exchangers.
– 2021-2024, WINGS project (CENAERO, VKI, SAFRAN AEROBOOSTERS) funded by Belgium: design and modeling of additively manufactured heat exchangers. This project will build upon the STREAM results to perform the design of heat exchangers in the framework of the development of the new RISE engine.
– 2023-2028, YALES2-AE project (CORIA/SAFRAN/GDTECH) funded by SAFRAN: this project is focused on the industrialization of the YALES2 software for aeronautical applications.
– SAFRAN TECH & SAFRAN AEROBOOSTER contacted CORIA for the modeling of sub- and super-critical heat exchangers.
– TOTAL ENERGIES has started a collaboration in 2022 with both CNRS-CORIA and TEMISTH on heat exchangers.

Additional perspectives can also be drawn from the STREAM project and pave the way for future projects:
– The STREAM project has been subjected to the same difficulties encountered by experts in accurately approaching & quantifying additive manufacturing roughness despite the fact that key parameters have been identified. More effort should be put on the characterization and quantification of additive manufacturing roughness.
– The STREAM project has also shown that existing correlations for the friction factor and for the Nusselt number can be used on some narrow validity range especially for pressure losses. Correlations for Nusselt number are far less reliable. New projects should focus on building new databases and new correlations for these heat exchangers.
– The STREAM project has clearly highlighted the very high impact of the printing direction on the friction factor and Nusselt number. This dependency is directly expressed by very different Effective Slope while having the same other roughness parameters. This flow direction sensitivity can lead to different impacts of the roughness on different parts of a heat exchanger and it has to be taken into account. This issue is linked to local optimization, which is not yet possible in current tools.

Periodic Reporting for period 2 - STREAM (Simulation of Turbulence and RoughnEss in Additive Manufactured parts)

Related documents

Share this page

Download