619 research outputs found
Advanced medium-order modelling for the prediction of the three-dimensional wake shed by a vertical axis wind turbine
No full textA recently formulated model for the treatment of the evolution of the wake of aerodynamic bodies has been implementedinto the wind turbine simulation software QBlade with the aim ofmodelling near and far wake behavior with a so-called mediumorder model. The paper first presents the vortex particle treatment of the wake. Shed and trailing vortex elements generatedby a lifting line model of the turbine blade are allowed to freelyconvect under the action of the freestream, body and wake influence. Induced velocities are calculated with use of a regularizedBiot-Savart kernel. The method is validated against experimentscarried out in the large-scale wind tunnel of the Politecnico diMilano on a H-type turbine architecture. Wake velocities and periodic unsteadiness are predicted relatively well by the methodfor two tip speed ratios. It is observed that higher order effects such as vortex stretching and viscous interaction must beimplemented into the model in order to accurately predict wakeevolution. A recently developed vortex particle multilevel multiintegration method has been implemented which approximatesthe far-field influence of the particles and reduces significantlythe computational expense. The paper also reports on the implementation of higher order effects into this optimization framework to account for evolution of vortex particle strength and theinclusion of viscous effects into the model, which are shown to beparticularly relevant for vertical axis wind turbines
Benchmark of a novel aero-elastic simulation code for small scale VAWT analysis
No full textAfter almost 20 years of absence from research agendas, interest in the vertical axis wind turbine (VAWT) technology is presently increasing again, after the research stalled in the mid 90’s in favour of horizontal axis turbines (HAWTs). However, due to the lack of research in past years, there are a significantly lower number of design and certification tools available, many of which are underdeveloped if compared to the corresponding tools for HAWTs. To partially fulfil this gap, a structural FEA model, based on the Open Source multi-physics library PROJECT::CHRONO, was recently integrated with the Lifting Line Free Vortex Wake method inside the Open Source wind turbine simulation code QBlade and validated against numerical and experimental data of the SANDIA 34m rotor. In this work some details about the newly implemented nonlinear structural model and its coupling to the aerodynamic solver are first given. Then, in a continuous effort to assess its accuracy, the code capabilities were here tested on a small scale, fast-spinning (up to 450 rpm) VAWT. The study turbine is a helix shaped, 1kW Darrieus turbine, for which other numerical analyses were available from a previous study, including the results coming from both a 1D beam element model and a more sophisticated shell element model. The resulting data represented an excellent basis for comparison and validation of the new aero-elastic coupling in QBlade.
Based on the structural and aerodynamic data of the study turbine, an aero-elastic model was then constructed. A purely aerodynamic comparison to experimental data and a BEM simulation represented the benchmark for QBlade aerodynamic performance. Then, a purely structural analysis was carried out and compared to the numerical results from the former. After the code validation, an aero-elastically coupled simulation of a rotor self-start has been performed to demonstrate the capabilities of the newly developed model to predict the highly nonlinear transient aerodynamic and structural rotor respons
Effects of Airfoil's Polar Data in the Stall Region on the Estimation of Darrieus Wind Turbine Performance
No full textInterest in vertical-axis wind turbines (VAWTs) is experiencing a renaissance after most major research projects came to a standstill in the mid 1990s, in favor of conventional horizontal-axis turbines (HAWTs). Nowadays, the inherent advantages of the VAWT concept, especially in the Darrieus configuration, may outweigh their disadvantages in specific applications, like the urban context or floating platforms. To enable these concepts further, efficient, accurate, and robust aerodynamic prediction tools and design guidelines are needed for VAWTs, for which low-order simulation methods have not reached yet a maturity comparable to that of the blade element momentum theory for HAWTs' applications. The two computationally efficient methods that are presently capable of capturing the unsteady aerodynamics of Darrieus turbines are the double multiple streamtubes (DMS) theory, based on momentum balances, and the lifting line theory (LLT) coupled to a free vortex wake model. Both methods make use of tabulated lift and drag coefficients to compute the blade forces. Since the incidence angles range experienced by a VAWT blade is much wider than that of a HAWT blade, the accuracy of polars in describing the stall region and the transition toward the "thin plate like" behavior has a large effect on simulation results. This paper will demonstrate the importance of stall and poststall data handling in the performance estimation of Darrieus VAWTs. Using validated CFD simulations as a baseline, comparisons are provided for a blade in VAWT-like motion based on a DMS and a LLT code employing three sets of poststall data obtained from a wind tunnel campaign, XFoil predictions extrapolated with the Viterna-Corrigan model and a combination of them. The polar extrapolation influence on quasi-steady operating conditions is shown and azimuthal variations of thrust and torque are compared for exemplary tip-speed ratios (TSRs). In addition, the major relevance of a proper dynamic stall model into both the simulation methods is highlighted and discussed
Comparison of experimental and numerically predicted three-dimensional wake behaviour of a Vertical Axis Wind Turbine
No full textThe evolution of the wake of a wind turbine contributes significantly to its operation and performance, as well as to those of machines installed in the vicinity. The inherent unsteady and three-dimensional aerodynamics of Vertical Axis Wind Turbines (VAWT) have hitherto limited the research on wake evolution. In this paper the wakes of both a troposkien and a H-type VAWT rotor are investigated by comparing experiments and calculations. Experiments were carried out in the large-scale wind tunnel of the Politecnico di Milano, where unsteady velocity measurements in the wake were performed by means of hot wire anemometry. The geometry of the rotors was reconstructed in the open-source wind-turbine software QBlade, developed at the TU Berlin. The aerodynamic model makes use of a lifting line free-vortex wake (LLFVW) formulation, including an adapted Beddoes-Leishman unsteady aerodynamic model; airfoil polars are introduced to assign sectional lift and drag coefficients. A wake sensitivity analysis was carried out to maximize the reliability of wake predictions. The calculations are shown to reproduce several wake features observed in the experiments, including blade-tip vortex, dominant and submissive vortical structures, and periodic unsteadiness caused by sectional dynamic stall. The experimental assessment of the simulations illustrates that the LLFVW model is capable of predicting the unsteady wake development with very limited computational cost, thus making the model ideal for the design and optimization of VAWTs
Über den Einfluss von wasserdampfverdünnten vorgemischten Wasserstoffflammen für zukunftsweisende Gasturbinen Anwendungen
No full textAufgrund der limitierten Verfügbarkeit fossiler Rohstoffe, ist der Umschwung auf neue Energieträger, neben der Entwicklung effizienter Energiespeichertechniken und einer intelligenten Lastverteilung, eine der größten Herausforderungen des 21. Jahrhunderts. Erzeugt aus regenerativen Energiequellen ist Wasserstoff hierfür ein aussichtsreicher Kandidat. Auch wenn Wasserstoff ein herausragender Energieträger ist, ist seine Handhabung nicht trivial. Daher ist es bis heute mit konventionellen Methoden noch nicht möglich eine herkömmliche Gasturbine mit Wasserstoff zu betreiben. Insbesondere die hohen Flammengeschwindigkeiten erhöhen das Risiko eines Flammenrückschlages erheblich. Daher ist es notwendig neue Verbrennungskonzepte zu untersuchen um ein tiefgreifendes Verständnis für die Verbrennungsvorgänge von neuartigen Treibstoffen zu entwickeln.
In der vorliegenden Arbeit wird eine Methode untersucht um eine effiziente Verbrennung von Wasserstoff in Gasturbinenprozessen zu ermöglichen. Dafür wird dem Verbrennungsprozess Wasserdampf beigefügt, welcher die Verbrennungstemperatur erheblich absenkt und gleichzeitig den Wirkungsgrad erhöht. Zusätzlich wird ein Gasturbinenzyklus vorgestellt, der eine saubere und effiziente Verbrennung von Wasserstoff erlaubt und gleichzeitig als effizienter Energiespeicher dient.
Im Rahmen dieser Arbeit wird der Verbrennungsprozess von Wasserstoff unter hohen Dampfmengen analysiert. Mit Hilfe von detaillierter Chemie und laminaren Vormischflammen wird Luft und reiner Sauerstoff als Reaktionspartner untersucht. Dazu werden in einem ersten Schritt diverse Reaktionsmechanismen untersucht und mit Literaturdaten zu laminaren Brenngeschwindigkeiten und Zündverzugszeiten verglichen. Es wird außerdem gezeigt, dass Wasserdampf den Verbrennungsprozess hauptsächlich als Stoßpartner beeinflusst, indem er starken Einfluss auf die Erzeugung von Radikalen nimmt. Dadurch wird die Brenngeschwindigkeit und die Selbstzündung stark gehemmt. Außerdem wird nachgewiesen, wie Wasserdampf positiv auf die Bildung schädlicher Stickoxide einwirkt. Bei der Untersuchung von turbulenten Vormischflammen beeinflusst Wasserdampf maßgeblich das Flammenbild. So breitet sich die Wärmefreisetzungszone weiter aus, die Flammenfront verdickt sich und die Flamme nimmt an Länge zu. Die Grobstruktursimulationen (LES) können im Vergleich mit den Experimenten das Strömungsfeld und die Flammencharakteristik sehr gut reproduzieren.
Der Beitrag dieser Arbeit ist ein tief gehender Einblick in den Verbrennungsprozess von Wasserstoff unter hohen Wasserdampfeinflüssen, der die Eignung von Wasserdampf für eine emissionsreduzierte und sogar emissionsfreie Verbrennung aufzeigt. Darüber hinaus liefert diese Arbeit validierte Entwicklungsmethoden für die Konzeptionierung von zukünftigen mit Wasserstoff betriebenen Gasturbinenbrennkammern.Following the limited availability of fossil fuels and growing environmental concerns, energy transition is beside energy storage and grid balancing one of the major challenges of the 21th century. Among the variety of alternative and renewable fuels, hydrogen is a promising candidate if produced from water and excess electricity or biomass. However, hydrogen combustion properties differ significantly from established fossil fuels. For example, it is not possible to operate safely a traditional gas turbine on high hydrogen content fuels, due to the high risk of flashback. In particular it is challenging to retrofit conventional gas turbine applications to allow for an efficiently and clean usage of hydrogen or hydrogen-rich fuels. In fact, it calls for a new generation of combustion technologies based on deep understanding of new fuels and their combustion behavior.
The present thesis addresses the possibility to utilize hydrogen as a efficient and clean gas turbine fuel. At nearly stoichiometric conditions steam is added directly into the combustion process, which significantly reduces the flame temperature. Moreover, a possible future gas turbine cycle is discussed that allows for the efficient combustion of hydrogen, energy storage and grid balancing. The scope of the thesis is to investigate the premixed combustion of pure hydrogen diluted with varying amounts of steam for gas turbine applications. The combustion process is modeled accurately using detailed chemical description of the complex oxidation reactions. In order to assess this several reaction mechanisms were identified and compared. Their respective performances are assessed based on laminar premixed flame calculations and auto-ignition events under dry and steam diluted conditions, for which experimentally determined measurements are available. A detailed study enabled to identify the effect of steam focusing on the third-body reactions with a significant increase of some key radial concentrations. In addition, high steam concentration results in a modification of the nitrogen oxides formation pathways, resulting in significantly lower emission concentrations. For the assessment of turbulent flames it is shown that the heat release spreads, the flame front thickens and the flame extends slightly further downstream with the addition of steam. In comparison with the OH* chemiluminescence images the simulated flame shape and positions are well in line with the experiments. In conclusion it is shown that the LES together with the assembled detailed reaction mechanism is able to predict the flow field and oxidation process of steam diluted hydrogen flames.
Within the scope of this thesis, new insight into the combustion process of highly steam diluted hydrogen flames is given, showing that steam dilution is a promising alternative for low- or zero-emission combustion. An additional outcome of this study is a accurate and validated set of tools and methods for a further design process of a hydrogen powered gas turbine cycle
Lösungen zur passiven und aktiven Strömungskontrolle an Rotorblättern für Windenergieanlagen
No full textDie vorliegende Arbeit beschreibt die Forschungsaktivitäten des Autors auf dem Feld der Windkraftanlagen-Aerodynamik, mit besonderem Augenmerk auf der passiven und aktiven Strömungsbeeinflussung. Die Beschreibung der theoretischen Grundlagen wird von einer detaillierten Literaturrecherche begleitet, in welcher viele verschiedene Methoden zur Strömungsbeeinflussung analysiert werden. Dabei wird besonders darauf geachtet, neben den aerodynamischen Aspekten auch technische Gesichtspunkte zu berücksichtigen, um eine aussagekräftige Vergleichbarkeit der Ansätze zu erhalten. Die erfolgversprechendsten Lösungen werden im Anschluss mittels strömungssimulationen(CFD), Windkanaltests und aeroelastischen Simulationen genauer untersucht. Mit Hilfe dieser Untersuchungen ist es möglich den Nutzen strömungsbeeinflussender Maßnahmen für den Einsatz an Windkraftanlagen abzuschätzen. Abschließend wird im letzten Kapitel ein fiktives, mit Elementen zur Strömungsbeeinflussung ausgestanztes "smartes" Rotorblatt vorgestellt und diskutiert. Das Ziel des Autors ist es, dass diese Arbeit als Grundlage für weitergehende Forschungsarbeiten an bestehenden Elementen zur Strömungsbeeinflussung dient und Anstöße für die Entwicklung neuer Konzepte liefert, welche die Nachteile der jetzigen Lösungen überwinden.The current thesis describes the research work of the author in the field of wind turbine aerodynamics and especially passive and active flow control solutions for wind turbine applications. The basic theory description is accompanied by a thorough literature review where many different flow control solutions are briefly analyzed. Special efforts have been made to include a wide range of technical fields, apart from aerodynamic performance, in the analysis process and thus try to provide a comprehensive study. The best performing solutions are then thoroughly analyzed by means of flow simulations (CFD), wind tunnel tests as well as aeroelastic simulations. Through this analysis it is possible to estimate the merits of such solutions for wind turbine applications. A final chapter with a fictive wind turbine "smart" rotorblade, equipped with flow control solutions is finally presented and discussed in this thesis. The aim of the author is that this work will be able to function as a base for further research of other flow control solutions as well as for the development of new flow control concepts, which will overcome current drawbacks
Numerical analysis of the flame dynamics and combustion instabilities of premixed flames
No full textQuasi One-Dimensional Modelling of Turbulence and Interaction of Combustion Chambers in a Shockless Explosion Combustor
No full textThis thesis deals with the modelling of two-dimensional coupling of quasi one-dimensional domains and turbulence within a quasi one-dimensional combustion chamber. Also an interpolation-free finite volume moving mesh method is described.
First, the basic framework of a gas turbine is introduced including an uncommon approach for constant volume combustion: the shockless explosion combustion (SEC). In a preceding work a simulation code for this combustion process solving quasi one-dimensional reactive Euler equations with a finite volume (FV) Riemann solver has been developed and was extended for the thesis at hand.
A network model is presented, allowing for the investigation of interaction of multiple pulsating combustion chambers of an SEC gas turbine with the plenums and each other. It couples the quasi one-dimensional domains using boundary conditions and flux corrections such that interactions of slanted combustion chambers with the plenums are possible. A series of simulations utilising this model is carried out to show possible fields of research for this tool.
As the simulation of combustion processes are especially sensitive to spacial resolution but complex chemistry also imposes restrictions on the number of grid cells a feature for adaptive remeshing is described. It uses the moving mesh idea within the FV solver. As interpolation introduces too much numerical diffusion a flux correction is given which evolves governing equations and mesh simultaneously without changing the Euler equations themselves. The performance of this feature is demonstrated with simulations of a detonation and a cyclic SEC.
Finally, the prerequisites for the research of the starting process of an SEC gas turbine are created by including molecular transport and turbulence in the SEC-code. Towards this aim, the one-dimensional turbulence (ODT) model is adjusted for this application. The ODT-line on which the stochastic eddy events, representing the turbulence, occur is aligned with the streamwise direction of the long-stretched combustion chamber. Also ODT is used as a stand-alone and subgrid-scale model. The main features of turbulence and ODT are compared to the new variant ODT-FHD. This study reveals that the ODT-FHD is able to generally reproduce the correct dependency of turbulence on mean flow velocity along with a plausible distribution of eddy sizes and kinetic energies. While lacking the possibility to generate new extrema of flow properties along the ODT-line it incorporates turbulent diffusion very well. The influence of the three model parameter is shown in addition to the simulation of a turbulent flame and a turbulent single-tube SEC
Deep Learning-basierte kardiovaskuläre Modellierung
No full textFluid dynamics play a vital role in understanding and treating a variety of cardiovascular diseases. Using high resolution medical-imaging data as well as modern flow simulation software, patient-specific blood flow in the heart and/or the vascular system can be calculated. From this data, clinically relevant parameters can be derived to supplement diagnosis and treatment. However, a widespread application of blood flow modelling in clinical practice remains unachieved despite numerous works detailing its potential to improve the diagnosis and treatment of various cardiovascular diseases. One obstacle for a clinical translation lies in the high computational demand of the blood flow modelling process, which is predominantly based on computational fluid dynamics (CFD). CFD simulations however require a considerable amount of computational power and runtime to provide accurate results. Applying these methods in clinical practice appears therefore unfeasible, since the required computational resources are rarely found in common healthcare facilities. As a possible solution to these practicability concerns, this work presents two artificial neural networks (ANN) which derive clinically relevant hemodynamic parameters from patient-specific geometry data. Compared to CFD, ANNs require little computational power and deliver results within seconds, thus potentially being more suited for a widespread clinical application.
The ANN-based modelling approach presented in this work focuses on two physically related cardiovascular diseases, namely coarctation of aorta (CoA) and aortic valve disease (AVD). In both diseases,an energy loss occurs due to a pathological constriction of available flow cross-section area within the aorta, potentially causing heart-failure if left untreated. As has been shown in previous studies, modelling patient-specific pressure within the pathological region can be of substantial benefit for diagnostics and treatment outcome. The ANN-based modelling approach presented here provides this information by calculating patient-specific pressure in the aorta, which is already a key parameter in clinical decision making, and wall-shear-stress (WSS) on the aortic wall. To assess the feasibility of this ANN-based modelling, the ANN-based computations were compared with CFD simulations for 37 (CoA) and 23 (AVD) test cases. Good agreement between both methods was found, however, several cases showed a substantial discrepancy. All in all, ANN-based modelling appears to be a feasible alternative to costly numerical methods, although further improvements are necessary for a clinical application.Im Bereich der kardiovaskulären Medizin trägt das Gebiet der Strömungsmechanik maßgeblich zum Verständnis und zur Therapie verschiedenster Herz-Kreislauf Erkrankungen bei. Mithilfe hochauflösender medizinischer Bildgebung und modernen Strömungssimulationsverfahren ist es sogar möglich, bei einem bestimmten Patienten die Strömung im Herzen und/oder im Gefäßsystem zu berechnen. Dadurch lassen sich wiederum klinisch relevante Informationen ableiten, die der Therapieplanung und Durchführung zugutekommen. Allerdings bleibt eine großflächige Anwendung solcher patientenspezifischer Modellierungsverfahren bisher aus, obwohl mehrere Forschungsgruppen das Potential solcher Verfahren, die klinische Diagnostik und Therapie zu verbessern, an verschiedensten Beispielen zeigen konnten. Ein Hindernis bei der Translation solcher Methoden in die Klinik ist der vergleichsweise hohe rechentechnische Aufwand. Bisher basieren fast alle Ansätze zur Modellierung patientenspezifischer Blutströmung auf numerischen Simulationsverfahren, welche hohe Anforderungen bezüglich Rechenleistung und Laufzeit stellen. Ein Einsatz solcher Methoden mit den an einem Krankenhaus üblicherweise real verfügbaren Mitteln ist daher schwierig bis unmöglich. Als mögliche Lösung für dieses Problem stellt diese Arbeit zwei künstliche neuronale Netze (engl. „artifical neural network“, kurz ANN) vor, welche klinisch relevante strömungsmechanische Parameter aus patientenspezifischen Geometriedaten berechnen. ANNs haben dabei den Vorteil, kaum nennenswerte Rechenleistung zu brauchen und Ergebnisse in wenigen Sekunden zu liefern.
Das hier entwickelte Berechnungsverfahren konzentriert sich dabei auf zwei strömungsmechanisch verwandte Herzkreislauf-Erkrankungen, nämlich Aortenisthmusstenosen (ISTA) und Aortenstenosen (AS). Bei beiden Erkrankungen entsteht ein Energieverlust aufgrund einer pathologischen Engstelle in der Aorta, welcher auf lange Sicht zu schwerer Herzinsuffizienz führen kann. Informationen über die Druckverhältnisse im betroffenen Bereich können, wie in vorherigen Arbeiten gezeigt wurde, einen substanziellen Mehrwert für die Behandlung solcher Patienten erbringen. Die hier vorgestellten ANNs berechnen dazu den patientenspezifischen Druck in der Aorta, welcher bereits jetzt nach klinischen Richtlinien zur Behandlungsentscheidung herangezogen wird, sowie die Wandschubspannungen auf der Aortenwand. Um die Nutzbarkeit der ANNs für den klinischen Gebrauch einschätzen zu können, wurden ANN-basierte Ergebnisse mit numerischen Simulationen für 37 ISTA- und 23 AS-Testfälle verglichen. Es fand sich eine gute Übereinstimmung bei der Vorhersage des Drucks und der Wandschubspannung, allerdings waren in einigen Einzelfällen hohe Abweichungen zwischen beiden Simulationsmethoden zu beobachten. Insgesamt konnte gezeigt werden, dass eine ANN-basierte Berechnung eine vielversprechende Alternative zu aufwendigen numerischen Simulationen sind, jedoch bedarf es noch einiger Verbesserungen für eine klinische Anwendung.BMBF, 13GW0208B, ArtiCardi
- …
