SACS - Dummy Structure

[mohammed43]

Hello,

I know that in Dummy structure, we can use only 40 boundary joints.
I had a situation where there are more than 40 joints (cylindrical tower) My dummy joint is centre of the cylindrer. So I made four overlapping joints (4 delete joints) & divided the joints of the cylinder in to four quadrants such that each delete joint has boundary (keep) joints from one of the four quadrants. Total load acting on the centre is made of four parts & applied equally on four delete joints.
Now the problem is when I analyze the model, the deflection doesnt seem right. The whole cylinder is not acting as a single unit under loading. The first quardant of joints are deflecting first then the second quadrant then the third....

Is there any solution for this..

I know that I can model dummy physical members (stiffer beams with zero density ) connecting centre of cylinder to all the boundary joints but i want to try using delete keep....

Thanks

 

[aaroneas]

Any joint that is called out as a "dummy" joint must be only part of the dummy structure only (as in a joint in the center of a primary column is not a dummy joint).

The user manual states:
"Loading incurred by the dummy structure is distributed to the boundary joints assuming that the dummy structure is a rigid body."

Therefore it can only distribute the load acting on the dummy structure to the dummy boundary joints, the purpose of this is to add loading to the structure but not stiffness.

Note that if you add beams with 0 density they will contribute to the global stiffness matrix. The 0 density just means that it will not add to the calculated weight.

Are you tying to "Mesh" the cylinder with the 4 quadrant joints? SACS is a beam element solver, so that means 1 beam = 2 joints. How are you breaking the cylinder into 4 quadrant joints and 1 joint in the center, as this is only one beam and therefore can only have 2 joints. It seems like you are treating SACS like a concept model (like in GeniE) and it just can't work that way (if I understand your question correctly). If you are just trying to induce a moment from the dummy structure I suggest adding a space load and not having the dummy structure at all.

Lastly, you say "the deflection doesn't seem right". What does that mean? Does that mean that the shape seems off or the values seem off? SACS Postvue scales the output to indicate where the large displacement occurs, this may mean that a .01 inch displacement visually appears to be a foot. So, look at the values not so much the shape. If the values are still off, how does it compare to yourself checking hand calcs?

 

[mohammed43]

Thanks for your reply..

I guess i didnt framed my question in the right way. I has a cylindrical meshed plate. And loads (forces & moments ) are to be applied at the centre of the cylinder, which in turn need to be distributed to the top nodes of the cylinder.

Deflection: I mean to refer to shape not the value

Space loads: I tried using this, but the deflection shape doesnt seem right after analysis..

Due to the limitation of only 40 boundary joints, I created 4 joints at the same location (absolute cordinates) & each joint is a dummy joint, which distributes to boundary joints..

Suppouse I Have say 120 boundary joints , I created dummy joints 1,2,3,4.(same absolute cordinates)
Out of total 120 boundary joints, first 40 joints are the boundary joints for dummy joint1, second set of 40 joints are boundary joints to dummy joint 2 and so on..
note: i divided the loads(force & moment) in to 4 parts and applied on each of the 4 dummy joints..

 

[mohammed43]

Please see the attchment for refrence..

 

[mohammed43]

Image showing the cylindrical plate..

The members i modelled to undersatnd the distribution of load from the centre to all those joints..

 

[aaroneas]

SACS is not very good at dealing with this type of situation (as far as my experience). SACS is good at dealing with beam elements (where it uses "Timoshenko Beam Theory" to solve its beams).

SACS does offer this information about plates:

"Plate offsets can be used to model transition points between plate and beam elements. Any two adjacent plate nodes can be specified as the same joint name. Plate offsets specified at each plate node can then be used to separate the nodes and place them in different spatial positions. This will result in one edge if the plate being described by the motion of one joint which can be connected to a beam element. For example, when modeling a tubular member with a finite element mesh, there is usually a transition point where beam element theory becomes sufficiently accurate. At this point, all of the plate elements must be attached to a single central joint which is the beginning joint of the beam element. The plate elements are connected to the central Joint with offsets such that the ends of the plates are located at the surface of the tubular. The transition joint will define the complete displacement of the cross section at that point and will assure proper internal load transfer. Also, the cross section of the tubular at the transition will remain plane during deformation which is a constraint of normal beam theory."

In essence if you are trying to run a FEA analysis where the loads are applied eccentrically and you want the mesh to properly interact it might be better to use a package like ABAQUS or ANSYS (where the program is good at working with a mesh + load transitions + shell interactions).

But if you still want to try this in SACS, you might try the Master/Slave node option.

Also, the image you posted doesn't appear to have the "green" elements evenly spaced. If the load is being transfer through those members it could result in some incorrect deflections.

I hope this helps.

 

[mohammed43]

The green elements, I showed are for explaning that load transfers from the centre to those nodes.
I had tried Master/Slave but its not working for the moments. Its good only for axial loads. Not working in the sense, deformation shape is not as expected.
Following are my observations:
Dummy Structure option: Deformed shape is correct but limitaion of only 40 boundary joints
Space Load: Defomation shape varies lot from the expected
Modelling Stiffer Beam elements (simlar to green elements as shown in the above image) Deformation shape is similar to the expected
Master Slave: Deformation shape varies
Space Load: Deformation Varies a lot.

I am trying to analyse a pedestal foundation, I had made the structure & meshed in Ansys & imported to SACS.

Thanks