MES vs RIKS on buckling load

[EngineerMickeyMouse]

Hi Support,

Could you please confirm my thoughts on your software, am I right to say that if I am interested about estimation of maximal critical load which will result in a structure buckling, user shall have exact (or at least very similar) results using Mechanical Event Simulation or Riks Buckling Analysis? Please note I am fully aware, that postbuckling phase may be assessed only in Riks, not MES.

[John_Holtz]

Hi E.M.M.,

I would agree with that, but there are a few points that may make the comparison difficult:

1. The inertia that is considered in an MES analysis might affect the buckling load, depending on how quickly the buckling occurs.
2. I am not sure how easy it is to determine "the" buckling load from either analysis type. In other words, does N displacement indicate the buckling load, or 2*N, or some other value? So trying to get "the same" buckling load from two different types of analyses is subject to some interpretation.

Have you tried the two analyses yet? What were your results?

[EngineerMickeyMouse]

I am in the middle of the try, however I have issues with software. I would like to stop nonlinear analysis at particular point, but once solver reach autotm limit it will drop me off from results tab and I cannot review gathered data. That is why I tried to tick on constant step size at advanced parameters, but then after 2nd step of analysis error message pops up " Run time error 91, object variable or with block variable not set". Any idea how to stop analysis at particular step and be able to review output at results tab?

[AstroJohnPE]

Hi,

I do not understand specifically what is happening when the analysis reaches AutoTM and you cannot view your results, but my guess is that there are no results to view because it did not get to the first converged time step. If there are any converged results, you should be able to view them in the Results environment either while the analysis is running or after the analysis has stopped (regardless of why it stopped).

For a Riks analysis, I am not sure if there is a way to stop the analysis at a specific step. (For an MES analysis, you would change the duration of the event.)

There may be times when the analysis had produced some results, but for one reason or another the Results tab becomes disabled (grayed out). If you exit from Sim Mech and open he model again, the Results tab may be available. If not, you should be able to do a "Check Model" to get back into the Results environment to see the existing results.

[EngineerMickeyMouse]

 Thanks, I will try this soon.

Please tell me, is there a way to plot a graph of all steps where at X will be situated deformation, while at Y axis will be RIKS load factor?

[John_Holtz]

It may be possible to do that (change the X axis of the graph from Time to some other result, such as displacement). This works for graphing other types of results, but I do not know if it will work for the load in a Riks analysis. Give it a try.

1. Display the loads and constraints ("View > Visibility > Loads and Constraints")
2. Select the load to graph ("Selection > Select > Loads and Constraints")
3. Create the graph (right-click in the graphics area > Graph Value(s))
4. To change the horizontal axis, "Results Inquire > Graphs > Graph Options > Axis Control".
5. Change the X Axis from "Default" to "Value", and then click the button to choose type of result for the X axis.

If that does not work, what you can do is create the graph of the load and export the data. Create another graph of the other result (displacement) versus time, and export that data. Then use a spreadsheet to combine the two sets of data and plot what you want (load versus displacement.) This article describes how to export the data from a graph: How to export graph values in Simulation Mechanical for use in third party program

 

[EngineerMickeyMouse]

John, I have followed your steps and managed to plot graph below, however I noticed some inconsistency. Please see below.

Looking at Fig.A you can see the load was applied only up to 1s (while all analysis took 10s), however 100% load was applied (value of load entered at FEA Editor was 1000 N). At Fig.B we see that used Initial Load Factor Increment was 3, as per Autodesk Help Definition (http://help.autodesk.com/view/ASMECH/2015/ENU/?guid=GUID-0B77F4DE-B0CC-4CB1-B067-E0B2E3A11847) one may say that on the Fig.A we should see load being incremented along all time of analysis, not just first second. To double check it, Fig.3 depict Load Level, which means on first converged step there was 3*Load applied, next 3.12*Load etc, while on Fig.A plot we can see at first step 0.1*Load, next 0.2*Load. What is going on? I shall expect to have the same results on Graph as it was set in MES RIKS Analysis parameter. Please let me know the answer.

[John_Holtz]

You are right! Sorry, but I missed the fact that the graph was wrong when I did a test model. I reported this problem to the developers.

If possible, use the reaction force ("Results Contours > Other Results > Reactions > Reaction Force (Negative)") to create a graph of the load versus time. In this case, you would select the nodes that are constrained and graph a sum of the reaction force (in whatever direction is appropriate).

If this is not possible, the load level from the log file (.ALG file) can be used to create a graph in a spreadsheet. The steps below are specific to Excel and may be different for other programs.

1. Select the convergence history text in the log file and copy it.
2. Start Excel and create a blank spreadsheet.
3. Paste the convergence history into the spreadsheet using the text import wizard ("Home > Paste > Use Text Import Wizard")
4. Select "Fixed width" in step 1 for the file type, then click the "Next" button.
5. On step 2, click on the "Data preview" to add a break between the "Load Level" and the "c" that appear on some lines. (The other breaks should be okay to split the data into separate columns.)
   
6. Click "Finish" to import the data.
7. Optionally, add a column and enter the original line number. This will make it easier to re-sort the data if desired.
8. Sort the convergence history by the "c" column, then by the "Time" column ("Data > Sort & Filter > Sort"). 
9. Graph the columns "Time" and "Load Level". (Optionally, add a column where you calculate the "Load Level" multiplied by the magnitude of the applied load, and graph time versus the load.)

If I remember correctly, you wanted to graph the displacement versus the load. So instead of step 9, you would add the displacement result to the table in Excel, and then graph the load versus the displacement.

[EngineerMickeyMouse]

I will try Excel in a moment, first what I did I have selected all nodes where support is provided and plotted graph, than edited curve to sum multiple nodes and I have arrived with nice graph of Reaction vs Time, however now it is not possible to use axis control to use displacement of DIFFERENT (it is obvious displacement at support is 0) node instead of Time at X axis. Is there any way to do this in the software?

[John_Holtz · Accepted solution]

It looks as if the answer to the question "Is there any way to do this in the software (plot result at node A versus the result at node B)" is ... Yes. I'm impressed!

After you create the graph of the sum of the reaction force at the desired nodes, you would do the following:

1. Switch back to the results window. (Pick the appropriate "Presentation" in the browser.)
2. Change the result type to displacement (or whatever result you want on the X axis).
3. Select the node that you want to use for the X axis.
4. Right-click > "Add Curve(s) to Graph" and pick the existing graph. This gives a plot of the reaction forces versus time, and the displacement versus time, on the same graph.
5. Change the result on the X axis from time to displacement. ("Results Inquire > Graph Options > Axis Control > X Axis Existing Curve" and choose the displacement curve for the X axis.)

[EngineerMickeyMouse]

John, your assistance was amazing! Thanks, much appreciated!

Below you can see final results, utilizing Sim Mech and Excel method to plot the most important graph of the RIKS algorithm.

Thanks!