I am simulating an offshore wind turbine using the FAST 8. I designed a controller to control the platform pitch angle. I am wondering what the values should I use to compare the performance of the controller.
- I calculate the root-mean-square of the pitch angle with the 0 degree as the mean value.
- I calculate the standard deviations of the pitch angles with the mean pitch angle calculated by available data.
As for the tower base fatigue load, I am wondering if the value obtained from situation 1 is more suitable to describe the tower base load or the value obtained from situation 2? I know I can calculate the tower base load using MLife. I am just wondering if the situation 1 is more meaningful or situation 2 in terms comparing the tower base load.
The root-mean-square with zero mean and the standard deviation about the mean are both approximations of what is more commonly used to assess fatigue i.e. the damage-equivalent load (DEL). In MLife, the effect of the nonzero mean can be accounted for in the calculation of the DEL, using the Goodman correction.
Please note that assessing the performance of a controller is rarely straightforward, as it is not typically suitable to focus on only one load metric. In addition to the tower-base load, it is likely also important to assess blade loads, drivetrain loads, tower-top loads, tower-top accelerations, actuator duty cycle, power output, etc.
May I ask how to calculate the DEL of tower base fatigue load based on the fast results? Should I derive it simply by sqrt(DEL(TwrBsMxt)^2+DEL(TwrBsMyt)^2) or any other methods?
For fatigue, the load variation at each point in a cross section may be unique, so, typically fatigue loads are processed using a load rose approach, e.g., in 10-degree increments around the structure. NREL’s post-processing tools like MLife have a load rose option that you can enable.
The vector sum approach you describe (without the DEL operator) can be used for post-processing ultimate loads.