Forum
Important Notice for New User Registrations
To combat an increasing number of spam and bot registrations, we now manually approve all new user registrations. While this may cause a delay until your account is approved, this step is essential to ensure the quality and security of this forum.
To help us verify your registration as legitimate, please use a clear name as user name or an official email address (such as a work, university, or similar address). If you’re concerned that we may not recognize your registration as non-spam, feel free to email us at with a request to approve your username.
Adding blade flexibility without doing structural analysis?
Hi,
Once you move away from idealized, steady-state, rigid-rotor BEM simulations, the power coefficient (Cp) starts to lose its practical meaning. This is especially true when the turbine operates in a non-uniform wind field and the rotor and shaft are tilted.
In that kind of situation, for an aeroelastic turbine you should evaluate performance using the “Structural Time Graph” outputs. Those channels represent the “measured” structural response (e.g., torque, thrust, bending moments) as seen by the drivetrain and support structure.
By contrast, the values in the “Aerodynamic Time Graph” are derived quantities. They are obtained by projecting aerodynamic forces onto an assumed rigid rotor, so they are not always representative of the actual loads and power transfer in an aeroelastic, tilted, sheared, or yawed inflow case.
——————————————–
Regarding the tower: how you scale it depends on your objective, but I generally would not recommend using full-scale tower properties for a model-scale turbine. Full-scale properties will typically make the tower unrealistically stiff at model scale, which can distort modal behavior and may negatively affect the simulation results. To obtain a scaled tower, you have two common options.
One option is to construct the tower geometry similarly to a blade, using cylindrical (or tapered) profiles for your target tower dimensions, and then generate the corresponding sectional properties.
Another option is to apply established scaling laws to the tower’s mass and stiffness distributions (mass, EI, GJ, etc.) based on your chosen scale factor.
For the tower diameter entries in the structural tower properties table: these are typically used only for visualization and for computing aerodynamic tower drag. You can manually change those diameter values to match your design without affecting the structural dynamics, as long as the mass and stiffness distributions in the properties table remain consistent with the tower you intend to model.
In other words, the simulation’s structural response is driven by the tower’s mass and stiffness distributions, while the diameter field mainly affects how the tower looks and how much tower drag is applied aerodynamically.
Best regards,
David
Hi!
Thank you for your previous answer.
We have some questions about the “structural time graph” outputs. We are not going too deep with structural analysis since we do not have enough knowledge and our thesis’s main focus is how different parameters effects the power coefficient and power. However we are still wondering how to interpret the graphs obtained in the structural time graph. For example the graphs showing torque vs time and aero. power coefficient vs time are oscillating. We are not sure how to interpret this so could you explain? (also is aero. power coefficient the same as power coefficient?)
Another question we have is about the “rotational speed settings”. Previously we have used the always fixed option. However when we tried with the free rotation, it barely rotated. Are these options about the rpm during the ramp up or is it about during the actual simulation?
Best regards
Emma, Vanessa, Belinda, Bjørnar
Hi Emma,
First, just a quick note: the forum is intended for software-related questions and is not meant to serve as a general introduction to wind energy topics. Before posting, we kindly ask that you take a moment to consult the extensive documentation, which also includes a search function to help you quickly find specific terms or topics. You can find the forum and community rules here: https://qblade.org/qblade-forum/topic/forum-and-community-rules/
Now, regarding your questions:
In aero-elastic simulations, oscillations in torque, rotor speed, power, and power coefficient are normal and physically expected. They arise from unsteady aerodynamics (sheared inflow, tilted rotor, etc.) combined with structural inertia, elasticity, and damping. The Structural Time Graphs represent the system response measured at the structural turbine model and are the relevant outputs for this type of simulation.
The “Aero. Power Coefficient” in the structural graph is the measured power coefficient, derived from measured rotor torque and speed. In steady rigid simulations, it converges to a constant value, but in unsteady simulations Cp starts to loose its meaning quickly and should rather be interpreted using time-averaged values.
Regarding the RPM settings please checkout the docs here:
https://docs.qblade.org/src/user/simulation/simulation.html#rotational-speed-settings
With “free rotation,” the rotor speed evolves dynamically based on aerodynamic torque, inertia, damping, and generator or controller settings (if included). Due to the inertia of the rotor and the low aerodynamic torque produced at low RPM the startup of a rotor can take quite a while – thats why the “rampup-only” RPM option exists.
BR,
David
Hi!
We have a question about the downscaling of a fullscale wind turbine. The scaling factor used was 1/140. This was used to scale down with the turbine blades and the turbine tower. All of the tower data was scaled down except the following parameters: EIx_[N.m^2] EIy_[N.m^2] EA_[N] GJ_[N.m^2] GA_[N]. This was done to ensure sufficient tower stiffnes – due to several problems with the tower “melting” during simulation.
Unfortunately the graph data was not displaying similar trends as the fullscale graph data. We have achieved a Cp and Ct value that is corresponding with earlier fullscale simulations.
Is there something else we should be aware about when doing the downscale? What kind of data should we expect with this scaling factor?
Would you also comment whether if this is a good way of scaling down the wind turbine in Qblade?
Attachment that shows the simulation data:
Best Regards, Belinda.
Uploaded files:- You need to login to have access to uploads.
Hello,
when you scale down a turbine both structural and aerodynamic properties change.
The structural properties can be estimated by the geometric scaling factor:
If the geometry is scaled by a factor λ (in your case 1/140), the main structural quantities scale as follows:
- Length scales with λ
- Cross-sectional area scales with λ²
- Volume and mass scale with λ³
For structural stiffness properties:
- Axial stiffness (EA) scales with λ²
- Shear stiffness (GA) scales with λ²
- Bending stiffness (EI) scales with λ⁴
- Torsional stiffness (GJ) scales with λ⁴
The turbine aerodynamics are affected by the change in Reynolds number. Typically, airfoil efficiency is decreased with Reynolds number decreasing.
So I would generally expect the power coefficient to be lower for the scaled turbine compared to the full-scale case, unless the airfoils are specifically designed for low Reynolds numbers. Even if Cp and Ct appear similar, the aerodynamic behavior along the blade may still differ due to the change in Reynolds number.
Regarding the tower stiffness: if parameters such as EI, EA, GJ, and GA are not scaled according to the geometric scaling laws, the structural response will no longer be dynamically similar to the original turbine. This can change tower deflections and natural frequencies and may explain why the trends in the plots differ from the full-scale results.
From your screenshot it appears you are keeping the rpm constant from the full-scale turbine, but lowering the wind speed instead to operate at the target tip speed ratio (TSR). I would suggest using a more realistic wind speed (e.g., around 10 m/s, which is achievable in a wind tunnel or field test) and increasing the rotor rpm instead to reach the desired TSR. This would also help maintain more realistic Reynolds numbers along the blade.
BR,
David
Hi!
Thank you very much for the explaination. We a few more questions to this matter.
- The dimentionless parameters of the tower file:
KSX_[-] KSY_[-] RGX_[-] RGY_[-] XCM_[-] YCM_[-] XCE_[-] YCE_[-] XCS_[-] YCS_[-].
Some of the dimentionless parameters are contributing to stiffness – should any of these be scaled down?
We have chosen to not scalese these down for now, and tried to simulate only considering your recommendations of the downscaling.
- When we tried maintaining normal wind speed (10-12 m/s) and using the same TSR as the fullscale when simulating the labscale model, the turbine flew away. So there is still some difficulities with using the higher windspeeds.
- Also, we have issues with the turbine blades. We tried to scale down the turbine blades in Blade Design Module under Modify Shape –> Scale. In addition to this we scaled down the stiffness of the blade in the Blade file – as we did for the Tower File. This blades was stretch out after ramp up, while the tower during rampup collapsed.
Kind regards, Belinda and Vanessa.
Hi Belinda and Vanessa,
What you are observing is quite typical when working with scaled aeroelastic models.
The dimensionless crosssectionalparameters (KSX, KSY, RGX, RGY, XCM, YCM, XCE, YCE, XCS, YCS) should not be scaled. Since they describe relative positions or ratios within the cross-section, they remain the same regardless of the turbine size.
The instability you observed when using full-scale wind speeds is most probably caused by scaling conflicts. When the geometry and stiffness are reduced but the wind speed remains the same, the aerodynamic loads become too large relative to the scaled structure.
There are different possible scaling approaches, and the choice usually depends on the goal of the scaled model – for example:
- Reynolds scaling keeps the Reynolds number constant so that the aerodynamic flow behaviour around the blades is similar. For small models this typically requires very high wind speeds.
- Froude scaling keeps the ratio between inertial and gravitational forces constant and is often used for aeroelastic or structural studies.
In many studies, scaling approaches are also combined or partially relaxed. For example, one might prioritize Froude similarity for the structural behaviour while accepting Reynolds number differences and compensating by using aerodynamic data from the full-scale turbine.
Best regards,
David
Hi David!
Thank you for your previous answer. We have implemented the scaled numbers in our scaled bladefile and done our simulations for the scaled blade. Now we have design our own wind turbine labscale wind turbine with a design wind speed of 12 m/s and a design TSR of 5.2 using a pretty simple BEM method where we iterated only the blade element at 70% of the total radius. a and a’ was fixed (calculated analytically). We iterated until RE converged and then applied the polars (Cl and Cd at max Cl/Cd) for this Reynolds number to all the other blade elements. The local blade parameters were then calculated for each blade element. When we performed an aeroelastic simulation with this blade, the blades diverged when we simulated with a wind speed of 12 m/s.
Since it was designed with a wind speed of 12 m/s we expected it would simulate with this wind speed. We attempted to use different Timestep sizes, Simulation Lengths, Rump up Times. But it all seemed to depend on the windspeed and RPM. The TSR needed to be 5.2, and therefore there was no change in the TSR from one simulation to another. We ended up with v=5m/s, RPM=517 and TSR=5.2, when simulating with these parameters the simulationed passed the Rampup and did not diverge. However, we are curious of how this is affecting the amount of power that we possibly would achieve if the simulation with v=12m/s and TSR=5.2 did not diverge.
We are using the structural blade data we obtained from QFEM, with TSR=5.2. Is there any aspect of QFEM that may have influenced the blade so it doesn’t work in the optimal windspeed?
Could the tower have an effect too? The tower used was the downscaled NREL_5MW Tower data, downscaled as you suggested in the previous messages.
Would you please give relevant information that would explain why this is happening.
Kind Regards
Hi,
QBlade simply simulates the physical model based on the user-defined structural and aerodynamic properties. These properties need to be physically reasonable, otherwise the structure, and with it the simulation, will diverge.
A pure geometric scaling does not always guarantee that the resulting mass, stiffness, inertia, damping, or eigenfrequencies are still reasonable and realistic. For example, if the model becomes very stiff, the resulting high natural frequencies need to be fully resolved, which may require very small timesteps. If the model is too soft, aeroelastic instabilities may arise.
In your case, I would mainly check whether the scaled blade and tower properties are still physically plausible, rather than the target TSR itself. Could such a turbine actually be built with the structural properties you defined? What kind of material and wall thickness would this require? If the properties are realistic, then a stable simulation at the intended operating point should also be possible.
Best regards,
David
