Home/MST. Modeling and Simulation Technology/MST. Modeling and Simulation Technology – Posters

The Science of Making Torque from Wind (TORQUE 2018)

20 June  |  21 June  |  22 June  |  Events View
AA | CM | DSE | MET | MST | WWT

Loading Events
This event has passed.

Room: BL.27 Building
Topic: MST. Poster Session
Form of presentation: Poster
Duration: 75 minutes

Authors:
Christian Høeg, Palle Flydtkjær, Zili Zhang

Abstract:
Mooring line modelling plays an important role in predicting the dynamic response of the floating offshore wind turbines (FOWTs), especially under extreme load conditions. This paper investigates the influence of different mooring line models on the stochastic dynamic responses of a spar-type FOWT. A 16-degree-of-freedom (16-DOF) aero-hydro-servo-elastic model for the spar-type FOWT is first established using Euler-Lagrangian approach, taking into consideration the full coupling of the blade-drivetrain-tower-spar vibrations, a collective pitch controller and a generator controller. Three different mooring line models have been established and incorporated into the 16-DOF model, namely the linear spring model, the quasi-static model and the lumped-mass model, the last of which include the hydrodynamic loads, inertial force and damping force of the mooring cable. Stochastic dynamic analysis of the coupled FOWT-mooring line system is carried out using 3 different turbulent wind conditions and 4 different sea states and a total of 2160 10-min simulations. The mean value, the standard deviation and the extrapolated extreme value of the structural responses (blades, tower and mooring cable) as well as the fatigue equivalent loads are compared for the three different mooring line models.

Authors:
Laurent van den Bos, Benjamin Sanderse, Lindert Blonk, Wim Bierbooms, Gerard van Bussel

Abstract:
Various load cases are prescribed for offshore wind turbines during the design phase of the turbine. The effect of various uncertain parameters has to be studied to provide meaningful and complete conclusions. The technique that is often applied in this regard is binning, where the probability density function under consideration is binned and in each bin random simulations are run to estimate the loads. We propose to apply polynomial interpolation instead, which is a more efficient mathematical technique that allows to more accurately predict the loads on the turbine for specific load cases. The techniques are demonstrated using BLADED on a specific load scenario.

Authors:
Gonzalo Sanchez-Arriaga, Alejandro Pastor-Rodríguez, Ricardo Borobia-Moreno, Roland Schmehl

Abstract:
The LAgrangian Kite SimulAtor (LAKSA) is a freely available software for the dynamic analysis of tethered vehicles, such as kites and drones, applied to airborne wind energy generation. This numerical tool involves four simulators. The one, two and four-line kite simulators, which consider flexible but inelastic tethers, are based on a minimal coordinate Lagragian formulations and can be used for the analysis of fly and ground generation systems, kite-based traction systems, and kitesurfing applications, respectively. The configuration of the mechanical system in the fourth simulator can be defined by the user, who can select the number of kites and the properties of the elastic and flexible tethers linking them. For all the tools, kites (or drones) are taken as rigid bodies and thedynamic equations of the tether-bridle-kite systems, together with the user-defined and time-dependent control variables, are solved self-consistently. Academic and research analysis can take advantage of the modularity of all the simulators and their inputs and outputs interfaces, which follow a common and simple architecture.

Authors:
Joannes Olondriz, Josu Jugo, Iker Elorza, Santiago Alonso-Quesada, Aron Pujana-Arrese

Abstract:
Model based control design method is a very successful design paradigms, which requires good system models, usually non-linear. The control design process usually starts with the achievement of linear models from those full non-linear models around a steady state solution. In this work, an alternative tool is presented to validate Floating Offshore Wind Turbine (FOWT) linear models. This easily implementable tool is based in the application of a chirp signal to the non-linear FOWT model to obtein more accurate linear models.

Authors:
Sung-Ho Hur, Bill Leithead

Abstract:
The Matlab/Simulink model of the Supergen (Sustainable Power Generation and Supply) Wind 5 MW exemplar wind turbine, which has been employed by a number of researchers in various institutions and Universities over the last decade, is improved, especially in speed, to facilitate wind farm modelling. Note that wind farm modelling usually involves duplicating wind turbine models. The objective is achieved through various stages, including prewarping, discretisation using Heun’s method in addition to Euler method, and conversion to C. Results are presented to demonstrate that improvement in speed is significant and that the wind turbine model can now be used for wind farm modelling more efficiently. It is important to highlight that improvement in speed is achieved without compromising the complexity of the turbine model; that is, each turbine included in a wind farm is not simplified or compromised.

Authors:
Bahri Uzunoglu

Abstract:
An optimal Bayesian update strategy that implements the subjective opinions of several experts are introduced for preventive maintenance of wind turbines while single expert opinion has been introduced by the author in the previous studies. This work is introducing the opinion of the wind farm manager or technician via subjective opinions based on a Bayesian adaptive update strategies for optimal preventive maintenance. Subjective opinion will be implemented to Bayesian cycles while experts can impact the distribution parameters with no knowledge of statistics but just by presenting their opinion as belief, disbelief or uncertainty. Statistical parameters such as minimal time of maintenance and cost of new strategy will be impacted by the interaction of wind farm manager and technician that interact with quantitative data with their opinions. The approach employs and complements the quantitative data from turbine Supervisory control and data acquisition (SCADA).

Authors:
Matt Harrison

Abstract:
Long, flexible wind turbine blades experience significant deflections during operation. Historically, the blade element momentum model in Bladed assumed that the blade is split into a set of radial aerodynamic sections which are aligned perpendicular to the pitch axis, and that any rotation of the sections due to deflection, pre-bend or pre-sweep in the blade can be ignored. For modern-day flexible rotor blades, under extreme loading conditions, this assumption no longer holds. Including this rotation when calculating inflow velocities and angle of attack, and when resolving the resultant forces at the element, causes a significant change in axial force (Fz) along the pitch axis, and therefore the blade out of plane bending moment (My). This in turn drives a change in overturning moments on the hub (which is driven by differences in bending moments between blades). An example 160m rotor shows an increase of 68% in hub overturning moment during an extreme gust event when aerofoil rotation due to deflection is included. This is verified through comparison to the alternative free vortex wake rotor model in Bladed. DNV GL Bladed includes the rotation of aerofoils by default at version 4.8

Authors:
Marco Gomiero, Uwe Schmidt Paulsen, Ernesto Benini, Torben Larsen

Abstract:
In this work, aero-elastic solver predictions with HAWC2 are compared with measured data from a VESTAS V52 wind turbine situated at DTU campus Risø. Nearly one year of various measured wind conditions were considered for selection of loads and performance simulations. A new methodology for adjusting strain gauge (SG) calibrations originally from blade pull testing over time is presented. As a result, we show and discuss the different predictions on power performance and compare results with measured blade loads, under the condition of adjusting blade SG pull test calibration for temperature and time degradation effects.

Authors:
Jonh Prospathopoulos

Abstract:
In the present paper the effect of an oscillating trailing edge flap on the wind turbine blade loading and performance is investigated. Two URANS solvers are used to simulate sinusoidal trailing edge flaps with a length of 10% chord and 10o deflection, centered at 75% span and extending at 15% of the blade. The predictions reveal significant 3D effects on the hysteresis, as well as the location and magnitude of the maximum amplitude of loading. The increase of flapping frequency from 1P to 3P results in less amplitude and larger hysteresis. A stronger influence of flap is observed at the negative flap angles

Authors:
Daniel Matzke

Abstract:
In this contribution the validation of the load calculation of a full scale mbs model of a 2.75 MW generic research wind turbine by comparison with comprehensive measurements from a 4MW system test bench is presented. The aim of the discussed entire system model is to produce load time series of components’ internal loads to serve as input for more detailed component models such as FE-tooth contact or FE-bearing models. Therefore the paper focuses on the validation of said internal components loads and the displacements and deformations these loads depend on. It is evaluated which model parameters influence these calculation objectives as well as the modelling methods and level of detail best used to represent them meaningfully.

Authors:
Jennifer Marie Rinker, Morten Hartvig Hansen, Torben Juul Larsen

Abstract:
This paper presents a model calibration investigation using a wide range of available experimental data. The wind turbine under investigation was the V52 research turbine located at Denmark Technical University (DTU) Risø campus. The experimental data included static and dynamic tests for both the entire wind turbine and the isolated blades. Each set of experimental data was used to calibrate a part of the final model. There are three main aspects of this paper. First, the different experiments are outlined, including an overview of the experimental procedures and the key results. Second, the model calibration procedure for each set of experimental data is explained. Third, a recommendation for an overall calibration procedure is presented for future researchers and the key outcomes of our calibration investigation are discussed.

Authors:
Schröder Laura

Abstract:
Previous studies have suggested the use of reduced order models calibrated by means of high-fidelity load simulations as means for quick wind turbine load assessments; the best performing surrogate modelling approach in terms of balance between accuracy and computational cost has been the polynomial chaos expansion (PCE). Regarding the growing interest in advanced machine learning applications, the potential of using Artificial Neural-Network (ANN) based surrogate models for improved simplified load assessment is investigated in this study. First results show that a feedforward neural network (FNN) with two hidden layers and 10 neurons per layer, trained with the Levenberg Marquardt algorithm is able to estimate blade root flapwise damage-equivalent loads (DEL) more accurately and faster than a PCE trained on the same data set. Further research will focus on further model improvements by applying different ANN model architectures and training techniques, as well as expanding the work with more load components.

Authors:
Sebastian Reisch, Georg Jacobs, Dennis Bosse, Amin Loriemi

Abstract:
When comparing the drivetrain configurations of today´s wind turbines on the market it is obvious that no rotor bearing concept has become clearly established. The failure statistics vary widly and recommended technical upgrades are often not validated on system level. To validate simulation models used to initially design or improve a rotor support concept and to monitor, understand and evaluate the mechanical behavior of a rotor support in the field a force transmission analysis can be useful.The parallel investigation of a rotor bearing support system on a system test bench and in a simulation model enables a detailed analysis of the mechanical interaction, e.g. between the main bearing and the gearbox in case of a three-point suspension. This enables the identification of possible evaluation and monitoring variables for field operation.

Authors:
Joseph Robert Saverin, David Marten, George Pechlivanoglou, Christian Navid Nayeri, Christian Oliver Paschereit

Abstract:
The current paper describes an aerodynamic model for treatment of wind turbine wakes. For accurate treatment of the wake evolution for multiple diameters downstream of the turbine, along with interaction with local flow features, a model with low numerical diffusion has been chosen, a vortex particle method (VPM), which inherently allows treatment also of shearing effects and viscous diffusion. Treatment of blade loading is facilitated with the use of a lifting-line model. Reduction of the computation cost from order N^2 to order N has been achieved by implementing the model within a multilevel framework. Details of correct specification of distributed and shed vortical elements in the blade wake are provided, along with details as to ensuring numerical consistency and accuracy in the far wake. In addition the model has been highly parallelized, so that relatively quick simulations at high fidelity can be achieved on the order of seconds. The model has been formulated as generally as possible in order to be able to treat arbitrary potential-flow type kernels. This allows for future application of the method to flows with boundary interaction. The ability of the model to produce results of comparable accuracy to CFD is demonstrated with experimental measurements.

Authors:
M. Gaunaa, N. N. Sørensen, N. Frost-Jensen Johansen, A. S. Olsen, C. Bak R. B. Andersen

Abstract:
The production loss due to a damaged Leading Edge (LE) can be quite significant. Therefore it is of interest for manufacturers to find better materials, construction methods or coatings for the LE. In order to do this, accelerated test of the LE erosion can be done in a Rain Erosion Tester (RET), which is essentially a three bladed rotor with a horizontal rotor plane running through falling water droplets at a relatively high rotational speed.  Since the RET is a fairly new contraption, there is not much information on the droplet path and the conditions of the droplets hitting the blades e.g. are they disintegrated/split up into smaller droplets before they impact on the blade surface. The present paper describes a series of high speed video recordings of droplets in a RET and the comparison and analysis of the measured data with simplified engineering models as well as full CFD simulations.

Authors:
Hyungyu Kim, Insu Paek, Kwansu Kim, song yuan

Abstract:
Dynamic wind farm simulation tool developed from the previous study is used tostudy the annual power increase of the active induction wind farm control at an offshore site inKorea. The turbulent wind is generated from the veers model and the measured averaged winddata for 16 wind directions and used to estimate the power output of a virtual wind farm with 3rows and 3 columns. A simple open loop wind farm control method was applied to find out theannual power increase considering the farm layout change according to the wind directionchange. The power gains were found to be changed with different scale and shape parametersbut the power increase was found to be very small.

Authors:
Jennifer Marie Rinker

Abstract:
This paper presents an open-source tool that can be used to simulate turbulence boxes that are constrained by measured data, which is useful for wind turbine model validation. The tool, which is called PyConTurb for “Python Constrained Turbulence”, uses a novel algorithm based on the Kaimal Spectrum with Exponential Coherence method, and the algorithm can effciently generate turbulence boxes under a wide variety of measurement constraints. The theoretical background for the novel technique is presented along with a few notes on its implementation in Python. The utility of PyConTurb is demonstrated using real data measured using three-dimensional sonic anemometers at the Denmark Technical University Risø campus. The presented results demonstrate that PyConTurb can successfully generate turbulence boxes from real measured data, including recreating the desired spatial coherence relationships between the simulated and measured time series. PyConTurb is shown to be a promising tool for investigating new spatial coherence models and for future one-to-one wind turbine validation studies.

20 June  |  21 June  |  22 June  |  Events View
AA | CM | DSE | MET | MST | WWT

Go to Top