A new approach to determine proper mean and fluctuating inlet boundary conditions is proposed. It is based on data driven techniques, i.e., machine learning approach, and its goal is to use any known information about the downstream flow to reconstruct the unknown or incomplete inlet boundary conditions for a numerical simulation. The European Research Community On Flow, Turbulence And Combustion (ERCOFTAC) test case of the swirling flow inside a conical diffuser is investigated. Despite its relatively simple geometry, it constitutes a very challenging test case for numerical simulations due to incomplete experimental data and to the delicate balance between core flow recirculation and boundary layer separation. Simulations are performed using both Reynolds averaged Navier–Stokes (RANS) and large-eddy simulations (LES) turbulence methods. The mean velocity and turbulence kinetic energy profiles obtained with the machine learning approach in RANS are found to be in very good agreement with the experimental measurements and the numerical predictions are greatly improved as compared to the previous results using basic inlet boundary conditions. They are indeed comparable to the best previous RANS using empirical ad hoc inlet conditions to accurately simulate the downstream flow. In LES, in addition to the mean velocity profiles, the machine learning approach also allows us to properly reconstruct the fluctuating part of the turbulent field. In particular, the methodology allows us to circumvent the lack of turbulent correlations associated with classical inlet synthetic turbulence.
The head losses inside the draft tube of a bulb turbine can represent an important portion of the total energy losses due to the low heads at which these machines operate. However, the complexity of the flow makes it a numerical challenge to simulate. Previous studies have shown that Large Eddy Simulations (LES) based on mean inlet conditions improve flow prediction inside the draft tube’s cone but diverge further downstream [4]. Moreover, uncertainties and the lack of detailed experimental data at the inlet have been identified as the main reason for these discrepancies and precise measurements close to the walls are necessary to correctly validate the numerical simulations [8, 4]. Thanks to detailed experimental measurements, the objective of this paper is to enhance the accuracy of previous results by performing LES computations to numerically simulate the flow inside the draft tube of a bulb turbine, and to investigate the influence of inlet conditions using an innovative approach. A two-criteria based mesh adaptation [14] along with an element-wise masking strategy is used to assure a good spatial discretization level while reducing the computational cost. Partial experimental data imposed at the inlet are often not sufficient to achieve a proper downstream flow prediction with a LES. The real challenge thus consists in the economical generation of proper mean and fluctuating inlet flow fields. We first show that a simple homogeneous and isotropic synthetic turbulence field added to the mean experimental profiles may improve the prediction of the downstream flow, but this is only achieved through empirical adjustments. Therefore, we also investigate the use of machine learning procedure to automatically generate proper inlet mean and fluctuating fields.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.