Calculation setup

Articles on this page:

• Using a Generator DLL
• Decreasing simulation time
• Simulation speed
• Running a simulation from the command line
• Running a simulation without the RNA
• Can I model a partially constructed turbine?
• What is the relaxation factor?

Problem

We have chosen the co-simulation approach which forces us to use a predefined step-size for the generator-solver. Since Bladed uses variable steps we need to align the applications in time. How do we do this?

Solution

To run Bladed as a fixed step integrator you need to set the minimum and maximum timesteps to be the same in the simulation control settings. Choosing the step size could be difficult as if you make it too small you will find that your simulations run very slowly, if you make it too large you will not capture the high frequency dynamics and in the worst case you may have an unstable simulation.

Keywords

Generator; DLL; Fixed step

Problem

What are some practical steps to reduce simulation time?

Solution

Here are some practical steps to reduce simulation time. Each method should be used with caution as simulation fidelity will be reduced.

• Reduce the number of modes in the model.
• Increase the minimum time step in “Simulation control”.
• In earlier versions of bladed you can use the “Blade station economiser” in “Simulation control”.
• The time step will also be constrained by things like the controller time step and dynamic wake time step. You can try increasing these however care must be taken if the controller has been design for a particular discrete time step. Also this will not help if the time step is usually smaller than this during the simulation.
• The “Refine deflections” option in outputs will save some simulation time if unchecked but please look in the theory manual to understand what this is doing before removing it.
• You can reduce the number of members in the tower or blade. If the blade is particularly detailed this could help as aerodynamics are calculated at every blade station.

For more detailed guidance on simplifying the model, especially in the context of Bladed Hardware Test, view the "Model simplification for realtime execution BHTM" document.

Keywords

Simulation speed

Problem

What determines the simulation speed in a time domain simulation?

Solution

The Bladed integrator uses a variable step Runge-Kutta algorithm. This means that the step-size is changed from one step to the next based on an integration error estimate relative to the integrator tolerance. If there are high frequency or rapidly varying states, the integrator takes short steps; if all the states are low frequency or unchanging then the integrator takes larger steps. If on a particular step the integrator calculates the error of any state to be larger than the integrator tolerance, then the integrator repeats the step with a smaller step-size. This is called a step reduction.

If the calculation outputs option “Software performance” is selected, there are additional output groups in the results of dynamic simulations. In the output group called “Software performance”, it shows the number of step reductions caused by each state. If one state has a particularly large number of step reductions, then it is worth checking whether:

• The frequency of the mode related to this state is high.
• The related mode has been excited due to an instability.

In versions 4.4 and later, there is also a “Simulation step sizes” output group. Here, you can plot simulation time against step number. Each step can be assumed to take a similar amount of calculation time, which means that the steeper the line, the quicker the simulation. If the curve goes particularly flat at one point in the simulation, it’s worth taking a look in the results at what has happened at this time, which may in fact uncover real issues.

Keywords

Slow; Speed; Integrator

Problem

How can I run a simulation from the command line?

Solution

To run a calculation from the command line you first need to create a dtbladed.in file. To do this you first need to enable the option “Warn when starting calculation” in Tools -> Preferences.

You then need to run your calculation using the “Run Now” button. A “Starting Calculation” info box will pop up (don’t click either "Start" or "Abort" at this stage). You should then navigate to the Bladed installation folder where you can find a sub-folder which name starts with "$$$$" (similar to the one shown below). In this folder you will find the dtbladed.in file. Take a copy of this dtbladed.in file to a folder location of your choice.

Open the command prompt.

There are two ways to run the simulation from the command line:

OPTION 1 (all versions of Bladed):

• Navigate to the location of your dtbladed.in file, e.g. cd C:\test_folder
• Call dtbladed using the installation folder path, e.g. C:\DNV\Bladed 4.18\dtbladed
  NOTE: windnd can also be run using the same approach, by first navigating to the location of the windnd.in file, and then running the windnd executable.

OPTION 2 (version 4.10 or newer)

• Navigate to the location of dtbladed.exe file, e.g. cd "C:\DNV\Bladed 4.18"
• Call dtbladed and specify the path of the dtbladed.in file, using the "i" command line switch, e.g. dtbladed -i C:\test_folder\dtbladed.in

Keywords

Command line; Calculation

Problem

Is it possible to simulate in Bladed without the RNA e.g. for modelling of a tower and support structure during installation?

Simulation without RNA can be useful to e.g. compare tower frequencies to other software, without the influence of the RNA.

Solution

There is not an explicit option in Bladed to ignore the rotor-nacelle assembly, however this can be achieved with the following steps. The basic idea is to make all RNA components rigid and (almost) massless, and to remove any aerodynamic loads on the RNA components.

Simulation type

• Choose a "Parked" simulation type, or a Campbell diagram in "parked" configuration

Blade screen

• Delete outer blade stations so that there are only 3 stations (the minimum allowed).
• On the "mass and stiffness" tab, set the blade mass and inertia to a small value, e.g. 0.01.
• Delete any point masses on the blade.
  
  

Nacelle screen

• Set the nacelle drag coefficient to zero, and mass and inertia properties as shown below.
  
  

Rotor: Hub screen

• For a geared turbine: set the properties shown in red below.
• For direct drive turbine: set the properties shown in red and green below.
  
  

Rotor screen

• Set the overhang to 0.01m.
  
  

Imbalances

• Set the rotor imbalance mass to zero.
  
  

Yaw screen

• Set the yaw type to "None".
  
  

Power train screen

• Set generator inertia to 0.01 and disable all power train flexibility options.
  
  

Brake screen

• Define a brake with a high torque value (e.g. 1e10).

Flexibility modeller screen

• Disable the blade flexibility.
  
  

Aerodynamics screen

• Disable the dynamic wake and dynamic stall models.
  
  

Keywords

Without RNA; Rotor nacelle assembly

Problem

Is it possible to model a turbine in Bladed that is in process of being assembled? Possible scenarios:

• Tower only
• Nacelle and hub but no blades
• One or two blades installed

Solution

For the "Tower Only" scenario you should follow the steps in the "Running a simulation without the RNA" article above. For the "RNA without blades", the same article applies but just ignore the three steps labelled "Nacelle screen", "Hub screen" and "Rotor screen".

To model a partially-constructed turbine with one or two blades installed: You can specify fewer than 3 blades in the Rotor screen and the simulation will run ok. This is easy for one blade; you need to specify a counterweight but you can make it negligibly small. For two blades it will assume that you want a true two-bladed rotor with 180-degree blade separation, but you can get around this by setting the "Error in Blade Azimuth" for either of the blades to be +/-60 degrees. You can do this on the Imbalances tab of Calculation Parameters.

Keywords

Partially constructed; Partially assembled; Construction; RNA

Problem

The "Initial conditions relaxation factor" under "Additional Items" may be optionally set to a value less than 1 - this can often help in improving the initial state of the system, reducing problems such as large blade motion in the early part of a simulation. But what exactly does it do?

Solution

The relaxation factor is used to slow down the convergence in the initial condition finding algorithm, which tends to lead to better results. The algorithm uses a numerical gradient descent type approach to find a state of the system in which there is zero acceleration of the rotor, i.e. steady (equilibrium) state. With relaxation factor (R) = 1, the "next" value found by the numerical solver is used as the starting point for the following iteration. If R is less than 1 however, the starting point for next iteration is a weighted sum of the "next" value and the previous value. The lower the value of R, the larger the weight of the previous value. In the block chart below, X_n+1 is the starting point for iteration n+1, X_new is the (hopefully) improved value found by the solver, R is the relaxation factor and X_n is the previous state. The iterations continue until zero accelerations are reached (within some tolerance).

Keywords

Relaxation factor; Initial conditions; Stability