Fatigue Analysis of OC4 JACKET

[Krishnaraj18]

I have a doubt regarding OC4 jacket simulation in openfast.How can i do fatigue analysis of OC4 jacket in openfast?should i get the stress concentration factor of each joint of the jacket? or not?

[jjonkman]

Dear @Krishnaraj18,

I'm not sure I fully understand your question. Certainly in OpenFAST you can calculate the member-level loads (sectional forces and bending moments) in members of jacket support structures for a range of conditions (load cases), including for the NREL 5-MW baseline wind turbine atop the OC4 jacket. However, OpenFAST does not compute stress or strain directly, and as such, does not directly consider stress concentration factors.

Perhaps the following OpenFAST discussion is of use to you: #1747?

Best regards,

[jjonkman]

Dear @Krishnaraj18,

I would say that tower deflection is driven by the aerodynamic loads on the rotor and tower and the blade deflection is driven by the aerodynamic loads on the rotor and gravity, but there is not a 1-to-1 correlation because of directionality of the forces, the pitch angle of the blades, and the different blade stiffness in flap and edge.

I agree that for parked simulations it best to set WakeMod = 0 and AFAeroMod = 1 in AeroDyn.

I would say that in most cases of variable-pitch wind turbines, the blades are pitch to 90 degrees in any parked or idling condition, regardless of the wind speed.

Regarding fragility, see a similar question discussed on our forum: https://forums.nrel.gov/t/deflection-of-elastodyn/2233.

I'm not sure I understand the rest of your questions.

Best regards,

[Krishnaraj18]

In parked only i have to change WakeMod and AFAeroMod?what about the other changes i have to edit ?I found n one journal like in intact parked mode only changing the angle 90 degree.

[jjonkman]

It sounds like you already know how, but to fix the blade-pitch angles to 90degrees, you set BlPitch(1-3) = 90 in ElastoDyn and disable pitch control in ServoDyn (PCMode = 0).

"Parking" a rotor implies that the brake is applied and generator will not spin, which can be achieved by disabling the generator DOF in ElastoDyn (GenDOF = False). But it is more common to idle a rotor in storm conditions, which can be achieved by keeping GenDOF = TRUE but not engaging the generator by setting GenTiStr = TRUE with TimGenOn > TMax in ServoDyn.

Best regards,

[Krishnaraj18]

How can I account corrosion in OpenFast?Is there any option in OpenFast that i can given corrosion rate? And what about PSI,I found an option in PSI.But my doubt is how can i account PSI for a particular location? That i have to find in other software?

[jjonkman]

Dear @Krishnaraj18,

Corrosion is not directly accounted for in OpenFAST. Modeling corrosive damage would imply a change in structural properties (e.g. mass and/or stiffness) over time. This damage would not occur instantaneously such that changes in mass and/or stiffness would need to be considered over the course of a given simulation, which is typically 10 minutes or so in length. Instead, the effect of mass and/or stiffness changes can be realized by setting up different input files characteristic of each damaged state to be considered.

By "PSI", are you referring to local stress? OpenFAST does not directly output stress; rather, OpenFAST outputs reaction loads at cross sections of flexible members, including shear and axial forces and bending and torsion moments. Computing stress or strain would require post-processing in a sectional analysis tool or structural FEA tool.

Best regards,

[Krishnaraj18]

Hii jjonkman,
I want to do soil structure effect on the response of the blade,tower and all the response of OC4 jacket wind turbine.I have to design the soil layer and planning to give distributed spring.My doubt is how can do it in openFast.In subDyn i found a SSI input file provision.But how to do it for distributed spring model for flexible foundation.Can you help me to figure it out.

[jjonkman]

Dear @Krishnaraj18,

SubDyn itself supports an apparent fixity (AF) representation of the foundation, as well as a coupled springs (CS) representation of the foundation; with the later available through the SSIfile. But a distributed springs (DS) model is not currently supported.

The new SoilDyn module of OpenFAST (to be released soon through the following OpenFAST pull request: #986) has a placeholder for future implementation of a distributed springs representation of the foundation.

Best regards,

[Krishnaraj18]

Thank you,
i have one more doubt,in some journal they mentioned about FounDyn module to consider SSI.Where is that module actually in OpenFast?How it's works?

[Krishnaraj18]

If i am considering Coupled spring model,What should be the changes i have to do in OpenFast modules?

[jjonkman]

Dear @Krishnaraj18,

I recall seeing a paper where FounDyn was mentioned, but this module was neither developed by, nor shared with, NREL.

To model a coupled springs (CS) foundation in SubDyn, set the reaction flags (RctTDXss through RctRDZss) to zero for each DOF of each reaction node you want stiffness to be considered, and set the stiffness matrix of the CS foundation via SSIfile. See the online SubDyn documentation for more information:
https://openfast.readthedocs.io/en/main/source/user/subdyn/input_files.html#base-reaction-joints
https://openfast.readthedocs.io/en/main/source/user/subdyn/input_files.html#ssi-input-file

Best regards,

[Krishnaraj18]

Dear Jjonkman,
I have one doubt regarding SSI file in OpenFast.For considering pile soil interaction i am planning to give stiffness matrix as SSI file .But my doubt is how can i generate stiffness matrx,and which stiffness i have derive,pile head stiffness or what?And while giving SSI file what are the changes i have to do in dat files.And is there any chances of increasing the responses in blade ,tower and substructure?

[jjonkman]

Dear @Krishnaraj18,

The SubDyn documentation I linked above explains the input file settings to define a coupled springs (CS) model of the foundation within SubDyn. As far as how to derive the stiffness, in a CS model, the entire stiffness of the pile + soil within the foundation is lumped and represented as a stiffness matrix at the mudline at the point where the pile pierces the soil. NREL has not supplied software to compute this CS stiffness, but you can read related papers on CS methods to understand the process.

Best regards,

[Krishnaraj18]

Dear Jjonkman,
I have on doubt regarding coupled spring model.In my project for considering Pile soil interaction, i used coupled spring model.I given SSI stiffness matrix as given
4.96133E+07 Kxx
0.00000E+00 Kxy
4.94332E+07 Kyy
0.00000E+00 Kxz
0.00000E+00 Kyz
1.40810E+08 Kzz
0.00000E+00 Kxtx
1.24989E+07 Kytx
0.00000E+00 Kztx
2.93747E+10 Ktxtx
-1.24989E+07 Kxty
0.00000E+00 Kyty
0.00000E+00 Kzty
0.00000E+00 Ktxty
2.93747E+10 Ktyty
0.00000E+00 Kxtz
0.00000E+00 Kytz
0.00000E+00 Kztz
0.00000E+00 Ktxtz
0.00000E+00 Ktytz
7.56098E+08 Ktztz
my soil type is sand,Is this values are reliable or not?And in the result i found no changes in blade deflection.But found a variation in tower top deflection and transition piece displacement.The transition piece displacement i found a large increase,but while looking to the Mudline bending moment,(BMx,BMy,BMz) i found no variation.Is it possible ?Why it's coming like this?Can you explain it.

[jjonkman]

Dear @Krishnaraj18,

I can't comment on the specific values you are using for the coupled spring (CS) stiffness matrix. These will depend on many factors.

I'm not too surprised that adding soil stiffness will lead to increased motion of the tower-top and transition piece displacement. I would guess that the external wind and wave excitation, which dominate the mudline loads, would not be much effected by the increased displacement, but I would think the increased displacement would have some effect on, e.g., the impact of RNA weight on the mudline loads.

Best regards,

[Krishnaraj18]

Dear Jjonkman,
I have done the dynamic simulation of the OC4 jacket wind turbine by providing the stiffness matrix mentioned in the previous discussion as an SSI.txt file in the Subdyn module. I found significant differences in the response of the tower and substructure. However, I want to find the new natural frequency of the entire structure, including the substructure, tower, and blades, due to the soil-structure interaction. How can I get the natural frequency?

[jjonkman]

Dear @Krishnaraj18,

You can calculate full-system natural frequencies of OpenFAST models, including those with SubDyn enabled, through an OpenFAST linearization analysis, followed by eigenanalysis, as has been discussed in many topics here on OpenFAST GitHub, as well as on our forum (https://forums.nrel.gov/).

Best regards,