Learning Forum

[longt8]

Hello,

I am trying to calculate the J integral in LS Prepost. I go to "Application"-> "Tool"->"Jintegral" to calculate the number.

[Jim Day]

Make sure you read both d3plot AND the keyword input file into LS-PrePost before proceeding to Applications>Tools>Jintegral.
J-integral values should appear in the message window at the bottom, right hand corner where it now says "You have to read the keyword file ...".

[longt8]

Hello jday I know I need to run a model with a crack and then go to Application->Tools->Jitegral to find the interface as follow but the message box did not give me the J integral value:

Could you let me know "You have to read the keyword file ...". refer to which keyword file ? I have tried to open all the file by LS Prepost but I have find the file with J integral value.

Thank you!
Best Teng

[Jim Day]

By keyword file, I meant the LS-DYNA input file ... probably either Sample.k or SampleJ.k in your case.
With d3plot already open, select File>Open>LS-DYNA Keyword File from the top menu bar to open Sample.k or SampleJ.k. Answer "yes" to the question "Is this the same model as the d3plot file". Then proceed as usual in the operations to calculate the J-integral.

[longt8]

Hello jday Thank you very much for your quick response!
This time I import the d3plot and the keyword input file to LS-PrePost before proceeding to Applications>Tools>Jintegral.


I did it in following steps:
(1) selected the nodes on the crack edge
(2) selected the crack surface node(white dot) where I only selected one node on the 1/4 ellipse surface.
(3) selected the Vector in X direction as the crack open direction1/4 ellipse surface.

Now I have 3 questions:
Question 1 for step(2): the "crack surface node", for this case do I need to select the ellipse surface shown above or the bottom surface plane on the bottom? Is it correct to only pick only one node?
Question 2 for step(3) : If the sample is pull in Y direction, the crack opening direction here is in X?
The crack propagate direction is in X so I am confused with the crack open direction(actually I think it should Y, but the J-integral will become negative) define in the interface.
Question 3 for dimension : dyna is dimensionless, so why the J-integral is N/mm? How the system select the unite with out user defined it. I can adjust it?

Note:
For this 1/8model, I give the symmetric BC:
(1)XZ plane on the bottom
(2)YZ plane on the left
(3)XY on the back.
Finally, I give the displacement on top for XZ plane in Y direction to simulate the a plate with a ellipse hole in the center.

Thank you!
Best Teng

[Jim Day]

This should help ...
https://www2.dynamore.se/public/jernberg/jintegral.gif

[longt8]

Hello jday This link is really helpful!
I think dyna default J integral unite as "ton, N, mm, s, MPa, N-mm" shown in the table, so we need to change our input parameter unit for example geometry and material parameter unites to match "ton, N, mm, s, MPa, N-mm". It that correct?

Thank you!
Best Teng

[Jim Day]

The table shows examples of consistent systems of units.
What gives you the impression that units of "ton, N, mm, s, MPa, N-mm" are required if the J-integral feature of LS-PrePost is used?

[longt8]

Hello jday Yes, do I have to use "ton, N, mm, s, MPa, N-mm" as my unit in order to calculate J-integral where unit is N-mm?
My unite is N, m, s, Pa, N-m, so I want to know the default unit setting for this.
Thank you!
Best Teng

[Jim Day]

You lost me. I don't believe there is a default system of units for J-integral.
The J-integral has units of energy per unit area, which simplifies to units of force/length. In other words, if your model is set up in units of N, m, s, Pa, the J-integral would have units of N/m.

[longt8]

Hello jady I agree with you for "there is a default system of units for J-integral". All the input is dimensionless and they are scaler number so the J-integral value should not show N/mm on the contour . At this point why the lsprepost shows N/mm and do you think it is an error in lsprepost so should we ignore it?
Thank you´╝ü
Best Teng

[Jim Day]

The J-integral is NOT unitless. It has units of force/length. Right now, I don't know why LS-PrePost labels it as "N/mm". That would imply that the unit system is hardwired, but I doubt that's really the case. I'll check with the LS-PrePost developer.

[longt8]

Hello jady Yes, dimensionless system of input ends up with a dimensional unit of J-integral is confusing.
I am looking forward from the feedback from developer.
Thank you very much!
Best Teng

[longt8]

Hello I checked the procedure and contour for J integral https://www2.dynamore.se/public/jernberg/jintegral.gif. It shows that the integral is a contour, but the J integral is fixed number. Could you let me know do I need to pick the largest value in the contour which valueIneed to use?

Thank you!
Best Teng

[Jim Day]

The J-integral reported by LS-PrePost has units of force/length and the value varies both temporally and spatially. Your LS-DYNA input determines the particular units of force, mass, time, and length in your model and in the output. For example, if the nodal coordinates you input with the *NODE keyword are in units of meters, then your model's length unit is meters. It is critical that your input adhere to a consistent system of units.
Previously, LS-PrePost labeled the J-integral as having units of "N/mm", which was incorrect if your model's force units and length units were not N and mm, resp. The LS-PrePost developers recently removed "N/mm" from the label so as not to be misleading. You would need to download a new LS-PrePost v4.8 executable if wanted that update.

[Jim Day]

When I say the J-integral varies spatially, I mean it varies from node to node along the crack tip.