Dear Jason:
By way of reminder my research project is to model a wind turbine in the Jamaica Caribbean Sea. I have selected the DeepCwind floater as one of the floater types to assess. The water depth at the site location is ~100 m. However, the water depth that is set for the DeepCwind model in OpenFAST is 200m. This is applies to both the HydroDyn and MoorDyn modules. I changed the water depth in both modules to 100 m. It is a simple input in HyroDyn. However, in MoorDyn it is a bit more involved process. In MoorDyn, under Connection Properties, I changed the z values to -100; left the y coordinates as the default value and the x coordinates were reduced to half the value, using the ratio of the length of 100/200. So for x, instead of 418.8,-837.6,418.8 I changed to 209.4, -418.8, 209.4. I left the unstretched length as the default value, 835.35m. When I ran the simulation for the 100 m deep water I got this error for each of the 3 mooring lines:
“Catenary solver failed for one or more mooring lines. Using linear node spacing.”
1)Can you please confirm if the water depth can be changed in the models? If it is can, what are the required changes in the MoorDyn module?
I also have a few more queries. I ran the same DeepCwind case but this time using turbulent wind of 3mps to 25mps with 2mps time steps. The results were okay for 3mps to 11mps but between 13mps and 17mps I got the following message:
“FAST_Solution:CalcOutputs_And_SolveForInputs:SolveOption2:AD_CalcOutput BEMT_CalcOutput(node 19, blade 1):UA_CalcOutput:Mach number exceeds 0.3. Theory is invalid.”
2) Can you say what the problem might be here? Not sure.
For the same case between 19mps and 25mps the same message as in 2) is shown with this additional message:
“Warning: Turning off Unsteady Aerodynamics due to high angle of attack (47.643 deg). (node 5,
blade 1)
Warning: Turning off Unsteady Aerodynamics due to high angle of attack (48.052 deg). (node 5,
blade 3)
Warning: Turning off Unsteady Aerodynamics due to high angle of attack (46.165 deg). (node 6,
Time: 1 of 900 seconds. Estimated final completion at 13:08:14.
Warning: Turning off Unsteady Aerodynamics due to high angle of attack (45.022 deg). (node 5,
FAST_Solution:FAST_AdvanceStates:ED_ABM4:ED_CalcContStateDeriv:SetCoordSy:Small angle assumption
violated in SUBROUTINE SmllRotTrans() due to a large blade deflection (ElastoDyn SetCoordSy). The
solution may be inaccurate. Simulation continuing, but future warnings from SmllRotTrans() will
be suppressed.
Additional debugging message from SUBROUTINE SmllRotTrans(): 3.85 s”
3) I believe this might be due to the blade having to adjust due to the high turbulent wind, not really sure. Can you please advise?
Finally, general query.
When I ran the simulation using the 5MW_OC4Semi_WSt_WavesWN.FST original case I note the following message was shown at the beginning for the ElastoDyn and SubDyn modules respectively:
“Nodal outputs section of ElastoDyn input file not found or properly formatted”
“Nodal outputs section of ElastoDyn input file not found or properly formatted”
The same message appeared for all other cases that were ran above.
4)Can you please say why this occurs and if it is something that needs to be resloved?
Thank you.
Regards,
AOAW
Dear Andre,
Here are my answers to your questions:
- Yes, the water depth can be changed, but I’m not following your reasoning for halving X while keeping Y and the unstretched lengths the same. The mooring system would likely need to be redesigned for the shallower 100-m water depth.
2 and 3) Are you initializing the rotor speed and blade-pitch angle to their respective mean values based on each mean wind speed you are simulating? If not, I can see the start-up transients being unrealistic (with large blade deflection, high angles of attack, etc.) and triggering all sorts of warnings near the beginning of the simulation.
- The “nodal output” warning will not effect the simulation results; you can ignore the warning.
Best regards,
Jason:
Thanks for the feedback. I have modified the mooring system based on the 100 m water depth and that solves the problem. I have used steady wind of 8mps as per the original case to check initially. However I am having another problem.
“The BEM solution is being turned off due to low TSR (TSR=-10196).”
I have tried to change the rotor speed to a low value say between 0 to 2 rpm but that results in unsteady aerodynamics. Not sure what the problem is as I thought the adjusting the rotor speed either up or down would have solved but. Neither approach works.
Wit regard to the following response:
2 and 3) Are you initializing the rotor speed and blade-pitch angle to their respective mean values based on each mean wind speed you are simulating? If not, I can see the start-up transients being unrealistic (with large blade deflection, high angles of attack, etc.) and triggering all sorts of warnings near the beginning of the simulation. Am I to calculate an initial pitch angle and rotor speed based on the wind speed I am using? I had used the default values of 0 initial pitch angle and initial rotor speed of 9 rpm
In general I am not sure how to use the pitch angle or rotor speed in analysis/design. Is it that in real life you can use the controller to set the initial blade pitch angle and rotor speed, based on the existing wind conditions, such that the initial pitch angle does not have to be zero but can be varied accordingly? To make it clearer, if I am designing a steel column for example for a particular bending moment I can adjust the column size to suit the moment that I am designing for. How does this similar design process apply for the pitch angle and the rotor speed?
Regards,
AOAW
Dear Andre,
The strong negative value of TSR suggests that your model has gone numerically unstable. Did you change anything in the DeepCwind semi input files other than the water depth in HydroDyn and mooring properties in MoorDyn?
If the rotor speed and pitch angle are not initialized near their proper operational conditions, the controller may not be robust enough to recover, resulting in a controller-induced instability. The initial rotor speeds and blade-pitch angles for the NREL 5-MW baseline wind turbine can be found in Figure 9-1 of its specifications report: nrel.gov/docs/fy09osti/38060.pdf. At 8 m/s wind speed, a rotor speed of 9 rpm and pitch angle of 0 deg should be fine.
Best regards,
Jason:
Thanks. Let me see if I understand fully what you are saying and also that I reading the graph correctly.
By proper operational conditions you mean rotor speed and pitch angles that are acceptable for the particular wind speed being investigated, in this case 8 m/s. If I understand the graph correctly at 8 m/s the rotor speed is 9 rpm and the blade pitch angle is still at 0 degree. The rated rotor speed is about 12 rpm, which could be considered as operating within the normal operating range. After about 11m/s rated wind speed, the pitch angle increases until it gets to a max of 24 degrees, at the cut-out wind speed of 25 m/s. Now I also see that the initial rotor speed is about 7.5 rpm and the intial pitch angle is 0. Meaning that the initial values selected should be close to 7.5rpm and 0 respectively. Is my understanding correct?
What I find interesting about the graph is that on the y- axis there a different variables with different units but they are all represented on the same axis. Please clarify, to make sure my understanding here is also correct.
With regard to HydroDyn in addition to the water depth I also changed Hs to 10m and Tp to 12.5s, for 8mps steady wind, sorry about leaving out that important bit of information. This is the case that created the instabilities. The default case of 1.2646m for Hs and 10s for Tp worked fine without any errors based on a check that I just did. However, when I used 15mps turbulent with 10mHs and 12.5s Tp, I still have this error:
“FAST_Solution:CalcOutputs_And_SolveForInputs:SolveOption2:AD_CalcOutput:BEMT_CalcOutput(node 19,
blade 2):UA_CalcOutput:Mach number exceeds 0.3. Theory is invalid. This warning will not be
repeated though the condition may persist”
Not sure how to solve this. if this means that 15mps turbulent wind might be too great for the system.
Finally I ran the linerization case and the output that I got is not what I was expecting. I got the platform motions and forces at the fairleads and anchors. I am not sure how to go from there to get the natural frequency or period of the system. Please explain?.
Thank you.
Regards,
AOAW
Dear Andre,
Your interpretation of the graph is correct. You should set the initial conditions by taking the rotor speed and pitch angle associated with the mean wind speed you are simulating, e.g., RotSpeed = 9 rpm, Pitch = 0 at 8 m/s; RotSpeed = 12.1 rpm, Pitch = 14.9deg at 18 m/s.
I would not expect a severe sea state to trigger an instability, but it may induce large motion, which may temporarily cause the rotor speed to be large and Mach near the blade tip to exceed 0.3.
The linearization process has been discussed many times on this forum. The natural frequencies, damping, and mode shapes can be derived through eigenanalysis on the state matrix, A.
Best regards,
Thanks Jason. I will check it out.
Regards,
AOAW
Hi Jason:
I have since discovered that the crux of the problem with the changing of the water depth was not related to the mooring line. I started to look at it again and made incremental changes, starting with the original DeepCwind case. The model runs satisfactorily for 100m water depth and the related mooring line adjustments. However, I made a change to Hs from original value of 1.2646m to 6.71m and got a warning message that stated that chi was too great and I should set skewmod to 1 instead of 2. It also mentioned that I had a low TSR. I then decided to change the value of Hs incrementally starting with a value of 1.2646 and incrementing by 0.1m for each simulation. It ran okay for Hs of 1.2646 m and 1.3646 m however at 1.4646 m I got this message:
Warning: SkewedWakeCorrection encountered a large value of chi (90.108 deg), so the yaw
correction will be limited. This warning will not be repeated though the condition may persist.
See the AD15 chi output channels, and consider turning off the Pitt/Peters skew model (set
SkewMod=1) if this condition persists
I proceeded by setting SkewMod= 1.0 and incrementally increasing the value of Hs until I got to 3.3646 m and get this message:
The BEM solution is being turned off due to low TSR. (TSR = -6509)
I am not sure what is going on here but it seems like for more extreme sea states the model is having some stability issues. Would be grateful for your feedback to find out if this problem can be resolved. In addition:
Does setting SkewMod to 1 instead of 2 affect the credibility of the output results in any way?
Thank you.
Regards,
AOAW
Dear Andre,
These large skew angle and low TSR errors imply to me that the motion of the rotor is large enough that it meets or exceeds the wind velocity, which sounds extreme, but would mean the flow through the rotor can drop to near zero and the skew angle can be large. Can you confirm that the turbine motions are reasonable for the conditions you are simulating–what do the platform motions, blade/tower deflections, rotor speed, and blade-pitch time series look like up until these errors are triggered? Do you see signs of unstable behavior (exponential growth in motion)?
Can you run the original OC4-DeepCwind model (at the original water depth) with the higher wave heights without errors?
What does your water depth / mooring system change do to the natural frequencies of the floater in six DOF, which you can assess, e.g., through free-decay simulations without wind or wave excitation?
Best regards,
Thanks for the feedback Jason.
I had a thought this morning that I should check the original case for greater wave heights and see how it behaves. I will check it and let you know.
Regards,
AOAW
Jason:
I have a query about the mooring line. The mass density is given as 113.35 kg/m.
Is this the same as the weight/m length?
I was checking through a mooring line catalogue (nylon material) and the weight/m length for 80mm dia mooring line is 3.97 kg/m. So just want to confirm that this property is the same property given as mass density in OpenFAST.
Regards,
AOAW
Jason:
I have now confirmed it’s the same unit. I believe the reason for the large difference in values is because OpenFAST uses the properties for a chain instead of a nylon rope.
Please confirm.
Thank you.
Regards,
AOAW
Dear Andre,
The MassDen in MoorDyn is the mass (not weight) per unit length of the mooring line. This mass density should be expressed independent from external buoyancy. The line weight in water will be calculated by subtracting the line buoyancy from the mass times gravity.
I hope that helps.
Best regards,
Jason:
Following up on two of the queries that you had put forward on April 14, beginning with:
“Can you run the original OC4-DeepCwind model (at the original water depth) with the higher wave heights without errors?”
The original OC4-DeepCwind runs satisfactorily from 3mps to 23mps. For 25mps it runs successfully with the following message:
“FAST_Solution:CalcOutputs_And_SolveForInputs:SolveOption2:AD_CalcOutput:BEMT_CalcOutput(node 19,blade 1):UA_CalcOutput:Mach number exceeds 0.3. Theory is invalid.”
This rules out the extreme sea state being the source of the instabilities as I had previously concluded.
I have amended the mooring line dimensions using some results that I found in a paper in which FOWT was in 90m of water. It runs satisfactorily from speeds of 3mps to 13mps but I used SkewMod=1 as was recommend by OpenFAST, since SkewMod=2.0 caused a large skew angle. After 15mps, the same error as for 25mps for original DeepCwind model is shown. Between 17mps and 25mps it becomes unstable during startup but still runs successfully. The typical message given is (this was at 25mps):
“Warning: Turning off Unsteady Aerodynamics due to high angle of attack (54.283 deg). (node 5,
blade 1); Warning: Turning off Unsteady Aerodynamics due to high angle of attack (58.041 deg). (node 5,
blade 3); Warning: Turning off Unsteady Aerodynamics due to high angle of attack (52.493 deg). (node 6,
blade 3); Warning: Turning off Unsteady Aerodynamics due to high angle of attack (45.231 deg). (node 6,
blade 2); Warning: Turning off Unsteady Aerodynamics due to high angle of attack (45.188 deg). (node 5,
blade 2)
FAST_Solution:FAST_AdvanceStates:ED_ABM4:ED_CalcContStateDeriv:SetCoordSy:Small angle assumption
violated in SUBROUTINE SmllRotTrans() due to a large blade deflection (ElastoDyn SetCoordSy). The
solution may be inaccurate. Simulation continuing, but future warnings from SmllRotTrans() will
be suppressed.
Additional debugging message from SUBROUTINE SmllRotTrans(): 0.75 s
Time: 3 of 900 seconds. Estimated final completion at 17:09:36.
FAST_Solution:CalcOutputs_And_SolveForInputs:SolveOption2:AD_CalcOutput:BEMT_CalcOutput(node 18, blade 3):UA_CalcOutput:Mach number exceeds 0.3. Theory is invalid.”
Having given the update, I have some specific queries:
1) Does the message about mach number exceeds 0.3 (25mps and original OC4-DeepCwind model) mean that the model is not satisfactory given that it occurs at only one node?
2) Does the unsteady aerodynamics shown between 17mps and 25mps mean that the model is unsatisfactory at these wind speeds and needs tweaking, notwithstanding that it runs successfully?
3) AeroDyn theory manual sates, “Set SkewMod to 1 to use the uncoupled BEM solution technique without an additional skewed-wake correction. SkewModFactor is used only when SkewMod = 1.” Does this mean that mean that SkewMod=1 is one way of assigning a skewed-wake correction and SkewMod=2.0 is another way of allowing for skewed-wake correction? Want to confirm that it is satisfactory to use SkewMod=1
4) The mooring line properties shown in the MoorDyn theory manual defines “Can and Cat” as the added mass coefficients in the transverse and tangential directions respectively. Do these have to be changed when the properties of the mooring line is changed (e.g. diameter, mass density)?
I have on final query and it is about another one of the other questions you had raised on April 14:
“What does your water depth / mooring system change do to the natural frequencies of the floater in six DOF, which you can assess, e.g., through free-decay simulations without wind or wave excitation?”
“Is the free-decay simulation the same as carrying out linearization and finding the eigen values of the A matrix?”
Thank you.
Regards,
AOAW
Dear Andre,
Here are my answers to your questions:
- The high Mach number warning has been discussed on our forum before, e.g. see: Mach Number exceeds 0.3 for large wind turbines (15MW) - #4 by Garrett.Barter.
- I’m not sure I can comment based on only on the information you’ve provided me. How does the system response (platform motion, blade deflection, loads, etc.) compare when you change the water depth? Could you redesign the mooring system in the shallower water condition so that the water depth is not impacting the results as much as it is now?
- AeroDyn always uses the inflow local to each aerodynamic analysis node, so, the aeordynamic loads will always be effected by the skewed inflow. SkewMod = 2 adds the Pitt/Peters skewed wake correction to the local induction. SkewMod = 2 is generally more accurate than SkewMod = 1, but SkewMod = 2 is invalid for very large yaw errors (much greater than 45 degrees). It seems odd to me that you’d get high yaw errors with SkewMod = 2, not seen at the original water depth, by changing the water depth / mooring design.
- I would expect Can and Cat in MoorDyn to change when changing the line type (e.g., changing from synthetic line to chain), but not by simply changing the diameter or length of a given line type.
A linearization analysis and free-decay simulations can both be used to identify system natural frequencies and dampings, but they are not the same thing. By “free-decay”, I mean provide a nonzero initial condition (e.g., initial platform-surge displacement) and run a (nonlinear) time-domain simulation without incident waves (likely also without wind); this can be used to identify the the platform natural frequencies, dampings, and quadratic drag. It is common to run six separate free-decay simulations (perturbing each platform DOF one at a time) to identify the natural frequencies and damping of the six floating platform modes (surge, sway, heave, roll, pitch, yaw).
Best regards,
Jason:
Thanks for the prompt response.
Regards,
AOAW
Jason:
You said that:
“It is common to run six separate free-decay simulations (perturbing each platform DOF one at a time) to identify the natural frequencies and damping of the six floating platform modes (surge, sway, heave, roll, pitch, yaw)”
Are the six separate simulations for the free-decay test done with different random seeds?
Thank you.
Regards,
AOAW
Jason:
Also does, "without incident waves" mean set WaveMod=0 in HydroDyn or should HyroDyn be turned off altogether?
Regards,
AOAW
Dear Andre,
Free-decay simulations of floating offshore wind turbines are typically run without incident waves. This means HydroDyn is enabled and WaveMod = 0, in which case the WaveSeed(s) are not used by HydroDyn.
Best regards,