Learning Forum

[Jose Mishael C J]

I am performing fatigue crack growth of a 2D specimen with an internal crack. The geometry is a square plate under tension and the internal crack is located at the center of the specimen. I am relying on XFEM for the crack growth study as SMART method is not available in 2D. I am using APDL for the entire simulation including the geometry creation, meshing, solving and post-processing. Before beginning the crack growth study, I performed validation of the SIF at the crack tips. Ansys solver computed the SIF correctly for both crack fronts/tips (Since my crack is fully embedded, there will be two crack fronts/tips). But when I try to perform the crack growth using either Life-cycle method or Cycle-by-cycle method, the solver is failing by showing an error and it reads "Both life-cycle and cycle-by-cycle methods for fatigue crack growth are found in the same model. Only one method is allowed in one simulation model. Please check input data". My code for crack growth is written exactly based on the instructions provided in APDL Fracture Analysis Guide. Key information from the code is as follows.

1. MESH200 elements to discretize the crack, which is modelled as a line. The solver created two crack fronts/tips with names '_XFCRKFREL1' and '_XFCRKFREL2' as expected. 
2. Singularity based approach adopted for XFEM crack growth as mentioned in Fracture analysis guide 
3. Solver successfully computed Level sets during solution phase making sure that crack is properly modelled. 
4. Paris law parameters are specified as suggested in Fracture analysis guide using TB command.
5. Then the error appeared as solver started to perform crack growth. 

The code snippet for fracture parameter calculation and crack growth

! Paris Law Constants

C = 2.29E-13

M = 3 

! Fatigue crack growth Law Specification

TB, CGCR, 2, , , PARIS TBDATA, 1, C, M

!Fracture Parameter calculations

CINT, NEW, 1 ! Crack ID for first crack front

CINT, CXFE, _XFCRKFREL1 ! Element component name assigned by APDL for first crack front elements

CINT, TYPE, SIFS ! Stress intensity factor needs to be computed

CINT, NCON, 8 ! Number of total contours required

CINT, NORM, 11, 2 ! Define the crack plane normal

CINT, NEW, 2 ! Crack ID for second crack front

CINT, CXFE, _XFCRKFREL2 ! Element component name assigned by APDL for second crack front elements

CINT, TYPE, SIFS

CINT, NCON, 8

CINT, NORM, 11, 2 ! Define the crack plane normal

 !Crack growth calculations

CGROW, NEW, 1 ! Crack growth set number for first crack front

 CGROW, CID, 1 ! Crack calculation ID, should be same as above [CINT, NEW, 1]

CGROW, METHOD, XFEM ! XFEM as crack growth method

CGROW, FCOPTION, mtab, 2 ! Fatigue crack growth law, material parameters specified as [tb, cgcr, 2, , , PARIS], if this command is not issued, the crack does not propagate

CGROW, NEW, 2 ! Crack growth set number for second crack front

 CGROW, CID, 2 ! Crack calculation ID, should be same as above [CINT, NEW, 2]

CGROW, METHOD, XFEM

CGROW, FCOPTION, mtab, 2

 !Fatigue related data

CGROW, FCG, METH, LC ! Life cycle method

CGROW, FCG, DAMX, 1.0 ! maximum crack growth increment

CGROW, FCG, SRAT, 0 ! stress-ratio

KBC, 1 ! loads are stepped for fatigue analysis

I have the follwoing questions.

1. What could be the reasons for the error encountered during my simulation?
2. Is XFEM in ANSYS APDL limited to only one crack front/ tip during crack growth? Remember, I can compute SIF for more than one crack fronts/tips using XFEM in ANSYS APDL.
3. Assume that in reality if only one crack tip/front of the internal crack will propagate, then I don't need second set of CGROW commands in my code, and the problem reduced to only one crack. But, how do we know only one crack tip/front will propagate especially for an internal crack located at the centre of the 2D plate? Which crack front will propagate? 
4. Before suggesting a symmetric model, I started with a simplified model. My real model will not be symmetrical. The interal crack can locate anywhere in the plate. Then, how can I solve the issue of multiple crack fronts/tips? 

I already saw that some open source FEM codes are capable to simulate the problem. I want to keep APDL if possible as I am more comfortable in writing APDL codes. Can someone suggest some solutions or their ideas to work out the problem? 

 

[Harshvardhan]

Hi,
It will be helpful to check out the Ansys Help page for XFEM based crack analysis, as it provides some examples for you to go through:
3.8. XFEM-Based Crack Analysis and Crack-Growth Simulation

Regards,
Harshvardhan
Ansys Help
Ansys Learning Forum (Rules & Guidelines)

[Jose Mishael C J]

 

 

Hi,

Thanks for the link.

I already covered almost all the available reference in ANSYS documentation to figure out my problem including the link shared. Unfortunately, there is no clear indication of how to model the multiple crack fronts/ tips during fatigue crack growth in these documents (Some information is available for computing SIF under quasi-static case, which I already implemented in my validation codes). The examples provided under ANSYS help pages also consider only one crack front/tip, which is quite simple compared to my problem. I am looking for something from APDL developement as the error resulted doesn’t make sense as I am specifiying only Life-cycle method in my codes. Also there is no mentioning of XFEM based fatigue crack growth is limited to only one crack front/tips under the limitations/ characteristics of ANSYS help page (https://ansyshelp.ansys.com/public/account/secured?returnurl=/Views/Secured/corp/v242/en/ans_frac/fracfcgxfem.html) or Fracture analysis guide (I know this is the pdf version of the help webpage). 

Thank you.

Regards,

Jose Mishael

 

 

[Harshvardhan]

Hi,
To define parameters for XFEM crack propagation method you need to use the XFENRICH command.
XFENRICH

The examples in the link shared in previous comment all use this command to define the parameters instead of TB, CGCR

I hope this helps.

Regards,
Harshvardhan
Ansys Help
Ansys Learning Forum (Rules & Guidelines)

[Jose Mishael C J]

Hi,

Thanks for the answer.

I haven't provided my entire code in the question. I am using XFENRICH command to create the enrichment zone in my model. The level set values are computed only if the enrichment zone is defined. In my case, the level set values for the initial crack are computing successfully. The problem is only during the fatigue crack growth. This is why I mentioned about the possibility to know the APDL development restrictions regarding multiple crack tips/fronts in XFEM based fatigue crack growth. The snippet of my code which is relevant for XFEM based study can be seen below.

! Define enrichment identification
! XFENRICH,EnrichmentID,CompName,,SING,RADIUS,SNAPTOLERANCE
XFENRICH, ENRICH1, TESTCMP, , SING, 0.0

! Define LSM values for cut elements
! Signed distance functions of the nodes of the intersected base mesh elements
XFCRKMESH, ENRICH1, M200EL1, M200ND_CRACK1

XFLIST

Where TESTCMP is the component representing elements in the enrichment zone, M200EL1 is my component for mesh200 elements of the crack and  M200ND_CRACK1 is the component containing my crack fronts/tips.

Can you please provide any suggestions which are not explained in APDL Fracture analysis guide?

Thank you.

Regards,

Jose Mishael

[Jose Mishael C J]

Hi,

I received support from my ANSYS license provider and found the solution to my problem. I am sharing it with the community.

It was a problem within my code itself. I made mistake in specifying the normal to the crack front/tip plane. The direction of the crack propagation will be in X direction and for two crack fronts/tips, the direction of normals should be opposite to each other. The modified section of the code is as follows.

! New Local coordinate system for CINT definition at left side crack tip

WPSTYLE, , , , , , , , 1

WPROTA, 180

CSWPLA, 11, 0,

CSYS,0

 

! New Local coordinate system for CINT definition at right side crack tip

WPSTYLE, , , , , , , , 1

CSWPLA, 12, 0,

 

!Fracture Parameter calculations

CINT, NEW, 1 ! Crack ID for first crack front

CINT, NORM, 12, 2 ! Define the crack plane normal

 

CINT, NEW, 2 ! Crack ID for second crack front

CINT, NORM, 11, 2 ! Define the crack plane normal

 

Thanks.

Best regards,

Jose Mishael