Software Modelling Limitations |
 
|
Software Modelling Limitations |
|
There are a number of limitations associated the modelling capability as it currently stands. The ability to model floating wind turbines is a relatively recent addition to the software, and the feature will grow and develop over time. In the meantime, users should be familiar with the underlying assumptions which the current model is based on. Limitations are sub-divided into aerodynamic and hydrodynamic for ease of reference.
•AeroDyn is capable of modelling one-, two-, or three-bladed rotors, but from a Flexcom perspective, a three-bladed configuration is implicitly assumed. Given that three-bladed turbines are the most common type, there are no immediate plans to support one- or two-bladed rotors.
•AeroDyn is designed to model horizontal-axis wind turbines only. It is not possible to model vertical-axis wind turbines, so there are no immediate plans to make this functionality available in Flexcom.
•If you choose the rigid blade model, blade geometries are approximated as rigid profiles. This means that variations in aerodynamic loading due to dynamic blade deformations are not captured. Hence blade rotational inertia effects are not included in the simulation, only the static weight of the blades is taken into account. The shaft is not modelled explicitly, and is instead represented by a rigid element which connects the hub location to the top of the tower. This element is non-rotational and its sole purpose is to transfer the aerodynamics loads from the hub to the tower. As the low-speed shaft torque is unknown in the rigid blade model, it is derived from the generator torque and the gearbox ratio, and this leads to a smoother signal than would be observed in reality. None of these issues are present for the flexible blade model.
•As noted in the section on Floating Platform Loads, hydrodynamic loading on the platform is based on information which is derived from a radiation-diffraction analysis. These simulations compute pressure integrals based on Bernoulli’s equation around the surface of a floating body, allowing forces and moments at the body's centre of gravity to be derived. Flexcom is not capable of performing a radiation-diffraction simulation, so you will need access to a suitable hydrodynamic solver. There are several established commercial codes available such as WAMIT and ANSYS Aqwa, while NEMOH is a popular open-source tool.
•A linear solution technique is generally adopted, which assumes that displacements of the free surface and the floating body away from their mean positions remain relatively small, such that the boundaries may be linearised, and this simplifies the wave-structure interactions significantly. A fully non-linear solution, which considers the instantaneous free surface and body surface as a function of time, is inherently more accurate but also very computationally expensive. There are no immediate plans to develop a fully-coupled non-linear potential solver approach with Flexcom. In any case, for bodies with relatively simple geometry (standard platforms are effectively composed of one or more submerged cylinders), the linear approach is assumed to be adequate, particularly for operational seastates where wave heights are limited.
•Wave induced loads are typically concentrated at a single location, such as the centre of gravity, rather than being distributed over the wetted surface area of the platform. Radiation-diffraction codes readily provide the total loads at the CoG so this is the most convenient means of transferring loads to the structural model. It is possible to apply the loads in a distributed manner, but this would require knowledge of the pressure forces acting over each of the hydrodynamic panels, and mapping the derived forces to relevant locations on the structural model. In summary, it would require significant additional effort on the part of the software user.
•As noted in the section on Floating Platform Loads, second order wave drift loads are computed using QTF inputs. Second order drift loads stem from two different sources, at difference frequency pairs between two different incident wave harmonics (ω1-ω2) and sum frequency pairs between harmonics (ω1+ω2) . Sum frequencies tend to be very high frequency and are not typically of interest, but difference frequencies are very important as they induce slow drift effects at frequencies much lower that that of the incident wave harmonics. Strictly speaking, a full QTF matrix is required for complete accuracy, whereby QTF coefficients are specified for all possible difference frequency pairs, and this may be further complicated if multi-directional wave loading is experienced. Furthermore, the calculation procedure for evaluating the second order loads with this level of precision is extremely computationally expensive. For efficiency, Flexcom uses a well-established approximation (Newman, 1974), which only requires the diagonal terms of the full QTF matrix to be specified. The off-diagonal terms are then approximated using an average of the relevant diagonal entries. This is generally regarded as acceptable given that the difference frequencies of most interest relate to those of similar frequency - i.e. adjacent harmonics, which have similar diagonal QTF terms anyway. Despite its advantages, the accuracy of Newman's approximation has been challenged, particularly for shallow water environments. It is not currently possible to model second order wave drift loads in Flexcom using the full QTF method, but it is hoped to add this functionality in the near future.