Transfering Data from I-Deas™ to Ansys™

The idea behind transfering data from I-Deas to Ansys is that we can use the solutions already provided by I-Deas and manipulate them
either in Ansys or with another outside source. Ideally, it would have been easiest to do the work in the I-Deas software package,
but due to it's limitations, we have resolved to extract the data, and work with the necessary data outside of the limitations of I-Deas.
Tranfering data from I-Deas™ to Ansys™ allows for more control of the situation because we are we will extract only the information
that we want or need, manipulate it, and then re-display the data with another tool for further analysis.

All of the manipulation of data in this case is done with a MatLab™.

The first step in this process is to determine the format in which I-Deas™ exports the information and also that format that the data
needs to be in for Ansys™ to properly interperet the all of this information.

First, it was necessary to recreate the cylinder in Ansys. This begins with locating all of the nodes in the I-Deas model and determining
how the information is output from I-Deas.

This data is simply output in the format of:

1st Line : Node Number
2nd Line : X location Y location Z location

However, because the cylinder was built along the Y axis, the X and Z coordinates were grouped in the output file. I know this
because of the documentation and the file header. Knowing that the file was output using CS3 (Coordinate System Number 3)
I know what order the nodal location data is output.

This was the header for the file . . . from this we can decipher the note below
Note . . . nodes are odered according to CS 3 . . . Y - Z - X
2420
5
Cylinder
1 0 8
CS1
1.0000000000000000D+000 0.0000000000000000D+000 0.0000000000000000D+000
0.0000000000000000D+000 1.0000000000000000D+000 0.0000000000000000D+000
0.0000000000000000D+000 0.0000000000000000D+000 1.0000000000000000D+000
0.0000000000000000D+000 0.0000000000000000D+000 0.0000000000000000D+000
3 1 8
CS3
0.0000000000000000D+000 0.0000000000000000D+000 1.0000000000000000D+000
1.0000000000000000D+000 0.0000000000000000D+000 0.0000000000000000D+000
0.0000000000000000D+000 1.0000000000000000D+000 0.0000000000000000D+000
0.0000000000000000D+000 1.3010426069826053D-018 0.0000000000000000D+000
2 0 8
CS2
1.0000000000000000D+000 0.0000000000000000D+000 0.0000000000000000D+000
0.0000000000000000D+000 1.0000000000000000D+000 0.0000000000000000D+000
0.0000000000000000D+000 0.0000000000000000D+000 1.0000000000000000D+000
0.0000000000000000D+000 0.0000000000000000D+000 0.0000000000000000D+000
-1
-1
2411

Secondly, once all of the nodes are in place, it is necesary to mesh the cylinder with the same kind of elements at the same nodal locations.
This set of data has the format as follows:

Element Number Element Type (Shell, bar, spring)
Nodes associated with element

In the case of the shell elements, it is necessary to keep the nodes in the proper order, because they are meshed in a certain order.
For this reason it was necessary to make sure that Ansys meshed these types of elements in the same order. If not we would have
had to re-arrange the elements. Luckily Ansys handled these elements in a similar fashion, so there was no re-arranging necessary.

In order to test the Ansys model and to get some intial results, we looked at making a movie of one of the displacement mode shapes
extracted from the I-Deas model.

The displacement data is output from I-Deas in the form of

Mode Number
Node Number
X_trans Y_trans Z_trans X_rot Y_rot Z_rot
Node Number
X_trans Y_trans Z_trans X_rot Y_rot Z_rot
Node Number
X_trans Y_trans Z_trans X_rot Y_rot Z_rot
Etc . . .

But the ultimate goal for this project is to create a movie of the Hoop or Longitudinal strain that the cylinder experienced during a test.
In order to do this we first exctact the strain data per mode from the I-Deas model. This however comes in the form of strain per
element, and then it breaks down to a tensor of strain values at each element nodal value.

So it was necessary to write a code to break down the data into nodal tesnor values and then tranfrom the tensor from cartesian coordinates
into radial coordinates, so that we could make a more direct comparison to the hoop or longitudinal data gathered from testing.

Finally, adding the modes into a total cylinder response. This is what took the most time and for this reason we decided to cut down the
number of nodes used in the whole process. Instead of using the total 5528 nodes, we cut out all of the midside nodes from the 8 noded shell elements
and used only 4 noded elements. This reduced the numer of nodes to 1835. This still took days on Guinivere to add just two or three modes together
to get the total response of the cylinder.