Search

A machine learning model is trained on HAWC2 time-domain aeroelastic load simulations in order to provide a load surrogate model from static rotor loads input to lifetime wind turbine design loads, to be utilized in fast design loads evaluation for design optimization. The simulations are performed on the IEA-3.4-130-RWT with a range of design variations for tip-speed-ratio, pitch-ramp settings, and blade length scaling. The surrogate model is shown to provide accurate predictions of lifetime blade and turbine ultimate and fatigue loads for design variations within the training design space.

As wind turbine blades become longer and more flexible, the structural models employed in current aeroelastic simulation tools, usually based on beam formulations, are challenged. Effects previously considered negligible in the response of these beam-like structures might now become significant to obtain accurate predictions of their behaviour. To assess this, a test matrix covering key geometric and material characteristics of wind turbine blades is defined and three test cases are analysed as an initial stage of this study. For each case, the static deflection of the beam model from the aeroelastic tool HAWC2 is compared against higher-fidelity shell/solid finite element models. Three structures are studied: an isotropic box beam, a composite box beam and a 12.6 m wind turbine blade from DTU. A good match is observed for the box beam test cases, which study the effects of out-of-plane and distortion warping and bend-twist coupling. Important aspects of high-fidelity models (such as end effects) are highlighted. The study of the DTU 12.6 m blade shows a significant difference (15%) in the torsion natural frequencies of the beam model, though a good agreement in twist angle under torsional loading. More pronounced differences in twist angle are observed when a load distribution based on the steady state response is applied, likely resulting from challenges in modelling complex structural features and three-dimensional effects.

Artificial Neural Networks (ANNs) are being applied as a faster alternative to Computational Fluid Dynamics (CFD) for wind turbine engineering wake models. Unfortunately, ANNs can fail to generalize if the data is insufficient. Physics-Informed Neural Networks (PINNs) can improve convergence while lowering the required data amounts. This paper investigates the PINN methodology systematically by considering varying amounts of data and physics collocation points. This work considers the rotationally symmetric Reynolds Averaged Navier-Stokes (RANS) formulation. Initially, a baseline fully data-driven ANN is studied to determine a suitable network size. Then, multiple PINN-based wake surrogates are trained with continuity and momentum conservation knowledge, varying amounts of data, and physics collocation points. It was found that including physics information under the best circumstances could improve accuracy by 18% at the cost of increasing the training time by a factor of 116. The findings imply that physics information can improve neural network based wake surrogates.

Greater landward velocities were recorded after six megathrust earthquakes in subduction zone regions adjacent to the ruptured portion. Previous explanations invoked either increased slip deficit accumulation or plate bending during postseismic relaxation, with different implications for seismic hazard. We investigate whether bending can be expected to reproduce this observed enhanced landward motion (ELM). We use 3D quasi-dynamic finite element models with periodic earthquakes. We find that afterslip downdip of the brittle megathrust exclusively produces enhanced trenchward surface motion in the overriding plate. Viscous relaxation produces ELM when a depth limit is imposed on afterslip. This landward motion results primarily from in-plane elastic bending of the overriding plate due to trenchward viscous flow in the mantle wedge near the rupture. Modeled ELM is, however, incompatible with the observations, which are an order of magnitude greater and last longer after the earthquake. This conclusion does not significantly change when varying mantle viscosity, plate elasticity, maximum afterslip depth, earthquake size, megathrust locking outside of the rupture, or nature and location of relevant model boundaries. The observed ELM consequently appears to reflect faster slip deficit accumulation, implying a greater seismic hazard in lateral segments of the subduction zone.

Item does not contain fulltext, 17 januari 2023

Inside wind farms, wake effects are the primary source of turbine interactions, and as such, they constitute one of the most important aspects of wind farm operations. The two most widespread methods for calculating the wind farm wake flow are computational fluid dynamics (CFD) methods and engineering models. Both methods have drawbacks; CFD methods can be very accurate but are computationally expensive. Vice versa, engineering models sacrifice accuracy by simplifying the physics, thereby improving computational efficiency. In cases where many evaluations of the flow are needed, this trade-off is a hindrance. One such case is wind farm layout optimization problems. It has been shown that the estimation of wake flows can be improved by surrogate modeling. Recently, Artificial Neural Networks (ANN) have been demonstrated to predict accurate wakes over a mesh. In this work, a new mesh-free ANN-based wake model is proposed. This new model can predict the flow everywhere in the domain, and as it employs smooth activation functions, it is suitable for gradient-based optimization. Two ANNs were trained with data generated by Reynolds-Averaged Navier-Stokes with Actuator Disc simulations for several yaw angles. The first ANN predicts streamwise wake velocity induction/deficit, the second ANN predicts added wake turbulence intensity. Both ANNs predict with a low error.

The essay questions whether and to what extent the law of obligations provides balanced and consistent solutions to the exceptional difficulties in fulfilling the obligations caused by the pandemic emergency. To this end, the direct application of the principle of solidarity is excluded and cannot justify judicial decisions that are incompatible with the letter of the law or contrary to the rationality of the system. On the other hand, the use of equity is allowed as a criterion for integrating the legislative provision and implementing a rule that has already been positivized in accordance with its rationale. From this perspective, the essay examines the temporary or partial impossibility of performance. Liability is also excluded when the contingencies related to the Pandemic, while not making the performance absolutely impossible, have prevented the debtor from acquiring the means that he was required to use. Without prejudice to the general principle according to which the lack of financial resources does not constitute an impossibility of performance, on the basis of an interpretation of the emergency legislation, the debtor exceptionally is not liable if he proves that the delay in the fulfillment of pecuniary obligations is not attributable to his fault, but to the observance of the measures for containing the infections. Where instead the debtor has the money necessary to pay but, precisely because of the economic crisis caused by the pandemic, the payment would expose the survival of his company to the risk of serious and irreparable consequences, the duty of good faith, which also binds the creditor, prevents him from demanding timely fulfillment, indirectly excluding the debtor's liability.