Inhalt des Dokuments
Center for Computational Science and Engineering, Berkeley, USA
AMReX: a software framework for multiphysics modelling
Adaptive mesh refinement (AMR) is one of several techniques for
adapting the resolution of a simulation in particular regions of the
domain. Block-structured AMR specifically refines the mesh by defining
locally structured regions with finer spatial, and possibly temporal,
This combination of locally structured meshes within an irregular global hierarchy is in some sense the best of both worlds in that it enables regular local data access while enabling greater flexibility in the overall computation.
Block-structured AMR was first designed for solving hyperbolic conservation laws with explicit time-stepping. However, as the complexity of the system being simulated increases, effective use of AMR becomes more challenging. Algorithmic complexity and evolving architectures pose addiitonal challenges. To meet these challenges, we have developed AMReX, a software framework for block-structured AMR algorithms on current and future architectures. AMReX supports a number of different time-stepping strategies and spatial discretizations for algorithms based on continuum and/or particle representations.
Current AMReX applications include accelerator design, additive manufacturing, astrophysics, combustion, cosmology, microfluidics, materials science and multiphase flow. In this talk I will focus on how AMReX supports a wide variety of multiphysics applications with different algorithmic priorities while maintaining performance across multiple high-performance architectures.
Über Ann Almgren: Ann Almgren is a senior scientist and the Group Lead of the Center for Computational Sciences and Engineering in the Applied Math Department. Her primary research interest is in computational algorithms for solving PDE’s for fluid dynamics in a variety of application areas. Her current projects include the development and implementation of new multiphysics algorithms in high-resolution adaptive mesh codes that are designed for the latest multicore architectures.
Paniagua Perez, Purdue:
Development of novel turbine designs at Purdue University
The presentation will begin describing some design challenges for future turbines, related to conventional gas turbines, pressure gain combustion and alternative green cycles. Regarding pressure gain combustion, the harsh flow conditions delivered by the combustor require alternative designs. We will review several strategies developed by the Purdue Experimental Turbine Aerothermal Lab: subsonic inlet, supersonic inlet and bladeless turbine. The subsonic inlet enables high efficiency and high power extraction, as well as acceptable pressure flections to the rotor. The turbine endwalls were parametrized and optimized with a multistep approach first using a steady approach with a wide range of parameters, and then with a narrower range using unsteady evaluations using a commercial 3D Reynolds-Averaged Navier-Stokes solver. Finally, an overall engine analysis of the gas turbine with conventional and pressure gain combustion demonstrates the superiority of pressure gain combustion integrated with the optimized turbine. A reduced-order model of the gas turbine models the the combustion process of the rotating detonation combustor, losses through the diffuser, and turbine losses. We will also present the bladeless turbine to harness power from supersonic axial inflow without inlet swirl, allowing for power extraction from harsh environments with minimal maintenance costs.
The computational work is integrated with a detailed experimental campaign performed at the Purdue Experimental Turbine Aerothermal Lab (PETAL). This experimental tri-sonic facility can operate continuously and also perform transients, suited for precise heat flux, performance, and optical measurement techniques. PETAL owns two modular wind tunnels, with two separate settling chambers and two sonic valves. The two different wind tunnels have three different test sections: LEAF (Linear Experimental Aerothermal Facility), BRASTA (Big Rig for Aerothermal Stationary Turbine Analysis), STARR (Small Turbine Aerothermal Rotating Rig) to service both fundamental and applied research. LEAF is completely transparent for optical imaging, and detailed calibration of both intrusive and optical diagnostics, aimed at technology readiness levels (TRLs) of 1–2. BRASTA was designed with multiple optical access to perform proof of concepts as well as validation of turbine component performance for relevant non-dimensional parameters at TRLs of 3–4. STARR comprises a two-stage turbine module, specifically designed to ensure accurate efficiency measurements, with a direct drive (no gearbox) high speed AC electric motor, enabling engine representative transient operation. The test section Reynolds number (Re) extends from 60,000 to 3,000,000, based on a characteristic length of 0.06 m. The adequate setting of the inlet massflow and sonic valve position enables a very wide range of inlet Mach number, from Mach 0.1 to Mach 6.5, with massflows up to 30kg
Prof. Anton Schiela:
Algorithmen zur Optimierung mit partiellen Differentialgleichungen
In vielen Anwendungen spielen partielle Differentialgleichungen eine wichtige Rolle, und oft reicht es nicht, diese simulieren zu können. Man möcht die zugrunde liegenden Prozesse auf optimale Art und Weise steuern oder regeln, man möchte unbekannte Paramter aus Messdaten identifizieren, oder Simulation und Messdaten in Übereinstimmung bringen. Solche und ähnliche Aufgaben führen auf Optimierungsprobleme mit partiellen Differentialgleichungen.
In diesem Vortrag sollen eine Reihe von algorithmischen Ansätzen für diese Problemklasse vorgestellt, und anhand von Anwendungsbeispielen diskutiert werden. Zur effizienten Lösung dieser Probleme ist es entscheidend, die zugrundeliegende mathematische Struktur der Gleichungen zu nutzen. Dies führt bei unterschiedlichen Fragestellungen immer wieder zu neuen mathematischen Herausforderungen und neuen Algorithmen.
Vladimir Dulin: Investigation of coherent structures in swirling jets and flames by using 2D and 3D particle image velocimetry
Flow swirl is commonly used for stabilization of flames in combustors. For high swirl rates a breakdown of the vortex core occurs resulting in a dramatical intensification of local mixing. Besides, a central recirculation zone, formed due to the breakdown, captures hot combustion products and provides heat exchange between them and the fresh mixture. Another feature of swirling jets is the phenomenon of vortex core precession. Moreover, the precession in swirling jets becomes very intensive after the vortex breakdown and formation of central recirculation zone. It is considered to be associated with rotation of a coherent structure, consisting of helical vortices and resulted from a global flow instability. The present talk reports on the experimental study of coherent structures, including processing vortex core, in turbulent swirling jets and premixed flames by using 2D PIV and PLIF and time-resolved 3D PIV. The focus is placed on quantitative analysis of the vortex breakdown effect on energy of different azimuthal/helical modes and turbulent mixing rate in the non-reacting swirling jets and on flame shape in the jets with combustion.
Russel Quadros: Passive flow control of shock-wave/boundary-layer interaction using surface roghness: A numerical study
Shock-wave/boundary-layer interactions (SBLIs) are commonly observed in high speed engineering applications such as air intakes, turbo-machinery cascades, helicopter blades, supersonic nozzles, and launch vehicles. The nature of the incoming boundary layer, which interacts with the shock, has a significant impact on the flow topology and aerodynamic performance of the aerospace vehicle. It is advantageous to have a transitional boundary layer interacting with the shock as compared to a laminar or a turbulent boundary layer. A transitional interaction could bridge the gap between the large separation size obtained in a laminar interaction and the high drag associated with the turbulent interaction.
Using direct numerical simulation, we trip the incoming laminar boundary layer through hemispherical surface roughness elements, and impinge an oblique shock wave while the boundary layer is still in transition. In order to analyse the effect of shock strength on the interaction, we consider two cases of shock-generator angles (3 and 6 degrees). We compare our results with available experimental data for cases with and without the impinging shock. We observe a suppression in the mean separation region for the transitional interaction case irrespective of the shock strength. Also, lower levels of instantaneous separation are observed in the transitional interaction as compared to the turbulent interaction.
Daniel Mira Martinez: Towards the modelling of multiphysics reactive systems using high-fidelity numerical simulations
This work introduces the state-of-the-art on the numerical and modelling framework available in the multiphysics code Alya to simulate practical combustion systems. The numerical framework is based on the Finite Element method using low-dissipation schemes with skew-symmetric properties. The approach is based on the low-Mach number approximation of the Navier-Stokes equations using both an implicit and explicit formulation. The turbulent combustion model is based on steady and unsteady flamelet models designed for partially premixed conditions and fuels characterizing autoignition and flame propagation. Results of practical examples of combustion systems with emphasis on code validation and accuracy assessment of physical models will be presented along with multiphysics applications for problems with conjugate heat transfer and fluid/solid thermal interactions.
Alessandro Sorce: Dynamics and Controls of innovative energy systems
The presentation aims at providing the audience
with the fundamentals in dynamic modelling of energy systems, with
application to the recent challenges in advanced cycles.
The availability of robust and reliable simulation tools, validated on experimental data, is a necessary step for the development and assessment of control strategies to be deployed in advanced systems,
avoiding the risk of instabilities and keeping under threshold the main critical parameters over time.
Through practical examples, the audience is provided with insights on dynamic phenomena of energy systems, which are often not considered at the design phase, and which may later become a technology barrier. Finally, the latest advances in dynamic modelling approaches and tools are outlined.
Alberto Traverso: Advanced techniques for power plant monitoring and diagnostics
The presentation will provide an overview of the monitoring and diagnostic process, starting from the field measurement validation and selection technique. Then will be introduced the two main modelling approaches, knowledge-based and data-driven, which can be adopted for the creation and expert system observing the phenomena to be monitored or assessing the component health conditions. For both modelling strategies, some example referred to SoA systems (e.g. GT and Combined Cycle) and advanced systems (e.g. Fuel Cell systems) will be presented. An overview of the analysis of residuals strategies and recent advances in the diagnostic process will complete the presentation.
Guillermo Paniagua: Unsteadiness in supersonic turbines
In view of decreasing weight and costs, current turbomachinery designs tend to increase the stage loading and to shorten the distance between blade rows. Hence, the understanding of unsteady flows is a necessary step towards the improvement of turbine-based propulsion. In this talk we will present the fundamental physical mechanisms abating the performance of transonic turbines. Then we will present design strategies to mitigate the unsteady forcing based on the redesign of the turbine passages as well as through active control strategies.
Guillermo Paniagua: Flow control research experience in compressible flows
This talk will summarize the three different areas of research that targeted the control of trailing edge shocks. A first proposal to control the fish tail shock waves consists on, pulsating coolant blowing through the trailing edge of the airfoils. A linear cascade representative of modern turbine bladings was specifically designed and constructed to be tested at subsonic and supersonic regimes (0.8, 0.95, 1.1 and 1.2) together with two engine representative Reynolds numbers at various blowing rates. In the second proposal we will review the experience of mounting DBD’s also at the airfoil trailing edge. Finally we will describe the control of the boundary layer on a compression ramp by means of DC electrical arc discharges, with the ultimate goal of modulating the compressible flow with the minimum amount of energy.
Pedro Mellado: Small-Scale Turbulent Mixing at the Top of the Planetary Boundary Layer
The understanding and the representation of small-scale turbulent mixing in the entrainment zone at the top of the planetary boundary layer remains a challenge. This challenge becomes sometimes quite problematic, like in the case when stratocumulus clouds cap the boundary layer. This seminar will present two examples showing how direct numerical simulation can be used to address part of that challenge. First, in the simpler case of cloud-free conditions, we will see that the entrainment zone can be better described in terms of two length-scales, instead of just one. This result helps explain the observed dependence of entrainment-zone properties on weak- and strong tropospheric stratification regimes. The second example considers the role of evaporative cooling at the stratocumulus top. We will see how mixing enhancement by wind shear can render evaporative cooling as important as other turbulence sources at cloud top. This implication is twofold: evaporative cooling can be relevant in the cloud-top dynamics but only when interacting with other processes (like wind shear), and wind shear needs to be added to the analysis of the stratocumulus-topped boundary layer.
Prof Marino: Aeroacoustics of supersonic jets: state of the art and future perspectives
Aeroacoustics of jet (sub- and supersonic) is among the fundamental topics considered in the framework of the analysis of acoustic phenomena coupled to fluid dynamic sceneries. The interest spans from the basic investigation aimed at the analysis of the sound generation mechanisms, concerning the complex fluid flow structures which outcome as an acoustic field, to the possible strategies to reduce the high noise encountered in many applications. The presentation focuses on the main characteristics of supersonic under expanded cold jets and puts in evidence the level of comprehension reached nowadays. The main numerical and experimental results are reviewed. back
Giorgia Sinibaldi: Experimental analysis on the aeroacousitcs of an under-expanded impinging jet
An experimental analysis on the acoustics and the fluid dynamics of an under-expanded circular jet impinging on a perpendicular plate is carried out. The purpose is to analyzed the discrete components of the noise which characterized this kind of flows. Starting from an “infinite" distance of the impinging surface, i.e. the free jet, the plate is moved towards the jet exit. First, the transition from the screech tones, typical of the free jet, to the impinging tones is shown. Then, the generating mechanism of the impinging tones is investigated. In particular, different acoustic fields are observed to be related to a modification of the jet flow structure, leading to a change in the generation mechanism of the impinging tones. The analysis is carried out by means of acoustic measurements, Particle Image Velocimetry (PIV) and wall pressure measurements. A wide range of nozzle pressure ratios and nozzle-to-plate distances are explored to obtain an exhaustive overview of the impinging jet behavior.