Refine
Document Type
- Doctoral thesis (2)
- Scientific article (1)
Has Fulltext
- yes (3)
Is part of the Bibliography
- no (3)
Language
- English (3)
Keywords
- Particle image velocimetry (3) (remove)
Institute
Physics-informed neural networks (PINN) are machine-learning methods that have been proved to be very successful and effective for solving governing equations of fluid flow. In this work we develop a robust and efficient model within this framework and apply it to a series of two-dimensional three-component (2D3C) stereo particle-image velocimetry datasets, to reconstruct the mean velocity field and correct measurements errors in the data. Within this framework, the PINNsbased model solves the Reynolds-averaged-Navier-Stokes (RANS) equations for zeropressure-gradient turbulent boundary layer (ZPGTBL) without a prior assumption and only taking the data at the PIV domain boundaries. The TBL data has different flow conditions upstream of the measurement location due to the effect of an applied flow control via uniform blowing. The developed PINN model is very robust, adaptable and independent of the upstream flow conditions due to different rates of wall-normal blowing while predicting the mean velocity quantities simultaneously. Hence, this approach enables improving the mean-flow quantities by reducing errors in the PIV data. For comparison, a similar analysis has been applied to numerical data obtained from a spatially-developing ZPGTBL and an adverse-pressure-gradient (APG) TBL over a NACA4412 airfoil geometry. The PINNs-predicted results have less than 1% error in the streamwise velocity and are in excellent agreement with the reference data. This shows that PINNs has potential applicability to shear-driven turbulent flows with different flow histories, which includes experiments and numerical simulations for predicting high-fidelity data.
The Taylor-Couette (TC) flow, the flow confined between two concentric independently rotating cylinders, is used as a perfect model to investigate shear flow over concave surfaces and one of the paradigmatic systems of the physics of fluids. In this thesis, an experimental investigation of the turbulent TC flow in a very wide gap geometry with a radius ratio 𝜂 = 0.1 is performed. The physical and dynamic behavior of the flow is studied in a geometry that has rarely been investigated before the current study, which makes this study unique. The study aims to understand the effect of curvature on the TC flow, particularly in cases where the circumferential length of the inner cylinder is smaller than the gap width. The flow is studied in the different rotation regimes: counter-rotating, co-rotating, and purely inner cylinder rotating regimes up to shear Reynolds numbers Re_s≤ 150000. The flow field has been qualitatively studied using visualization techniques. By probing the different flow parameters, familiar coherent TC flow patterns appear, in addition to newly observed patterns we assume only exist for very wide gap TC flows. For a more detailed quantitative study, a time-resolved velocity field measurement has been conducted using the High-speed Particle Image Velocimetry technique through the system end plate. The radial and azimuthal velocity components in the 2D horizontal plane are measured at different axial positions, in order to scan the axial variance of the flow. The recorded flow field is used to compute the angular momentum transport in terms of the quasi-Nusselt number (Nu_ω). The results show a maximum of Nu_ω for low counter-rotating rates of −0.011 ≤ μ_max ≤ −0.0077, which is associated with large-scale structures that span the entire gap. Moreover, the Nu_ω decreases for counter-rotation rates higher than μ_max until it reaches a minimum value and then tends to increase again for higher counter-rotation cases. The space-time behavior of the turbulent flow field for the high counter-rotating cases shows the existence of newly observed patterns next to the outer cylinder wall that propagates inward, enhancing the angular momentum transport and resulting in a second maximum in transport for higher counter-rotating rates. For the pure rotating inner cylinder, the momentum transport scaling with the shear rate Nu_ω ∼ Re_s^α has been studied, and it shows a transition in scaling to 𝛼 = −0.76 for all flows with Re_(s )≥ 25000. This new scaling reveals the transition of the flow from the classical turbulent regime to the ultimate one, where this transition is accompanied by a clear change in the flow behavior. Moreover, the flow in the co-rotating regime and particularly in the centrifugal stable regime (𝜇 > +0.01) is investigated. The Velocity measurements show the presence of disturbed flow near the inner cylinder, where the measured velocity profiles showed a clear deviation from those predicted by laminar flow for flows up to 𝜇 = +0.04.
Stratified vortices can be found from small to large scales in geophysical and astrophysical flows. On the one hand, tornadoes and hurricanes can lead to devastation and even a large
number of casualties. On the other hand, vortices can distribute heat and momentum in the atmosphere which is important for a habitable environment on Earth. In the astrophysical context, accretion disks (from which solar systems are formed) can be seen as stratified vortices. In such systems, understanding the mechanisms that can result in an outward transport of angular momentum is a central problem. For a planet or star to be formed in a disk, angular momentum has to be carried away from its center to allow matter aggregation by gravity; otherwise, its rotation speed would be far too large, avoiding this matter aggregation (and the consequent star formation) to happen. In such gas systems, turbulence is the most likely mechanism to achieve such a large angular momentum transport. However, it was shown that the flow profile of accretion disks is stable with respect to purely shear instabilities, and the question arises about how the turbulence can be generated. Among other candidates, the strato-rotational instability (SRI) has attracted attention in recent years. The SRI is a purely hydrodynamic instability that can be modeled by a classical Taylor-Couette (TC) system with stable density stratification due to axial salinity or temperature gradients.
In this thesis, a combined experimental and high-performance computing study of new specific behaviors of the strato-Rotational Instability (SRI) is performed. The density stratification causes a change in the marginal instability transition when compared to classical non-stratified TC systems, making the flow unstable in regions where – without stratification – it would be stable. This characteristic makes the SRI a relevant phenomenon in planetary and astrophysical applications, particularly in accretion disk theory.
Despite many advances in the understanding of strato-rotational flows, the confrontation of experimental data with non-linear numerical simulations remains relevant, since it involves
linear aspects and non-linear interactions of SRI modes which still need to be better understood. These comparisons also reveal new non-linear phenomena and patterns not yet observed in the SRI, that can contribute to our understanding of geophysical flows.