Regretfully no, the shell edges are always continuous or fixed to all the objects that connect to them. In cases where you do not want the shells to transfer out-of-plane bending moments to the supports, like an edge beam or wall, there are only a few options:
For shell elements that are primarily subject to out of plane bending, Ram Elements can provide output for the area of steel required in the two principal directions per ACI 318-05.
The shells must be concrete and the model must include design combos to take advantage of this feature fully. Mechanical cover for the shells should also be assigned in the Shells - Materials spreadsheet.
To see the output on screen first select the desired combo at the bottom and then use View - Stresses to see stress contours. In the list of stress options, pick As 1 or As 3 top or bottom.
To get a report of the same information use Output - Analysis - Analysis Results - and select the last option "Flexural reinforcement in shells", Be sure to select the desired combo or combos on the right and then OK.
Obviously having shells with local axes aligned to the direction of the reinforcement is critical for making this output useful.
Also note, this does not take into acount shear design nor the effects of axial tension or compression acting on the shells.
Yes, one of the keys to modeling the mat foundation is to define a regular mesh. Since the nodes of the mesh need to be supported by vertical compression-only springs, we need the mat to be manually meshed. Using a regular size mesh will help so that the tributary area of each node is fairly constant allowing you to use the same spring stiffness.
The assigned spring stiffness should be derived from the soil subgrade modulus and the average shell area. Note, the spring units are in force/length, e.g. kip/in, meaning how many kips of compression at this point does it take to yield 1 inch of deflection. To make the springs compression-only, just check the Compression only - TY box (assuming Y is vertical).
In order to insure stability some soil friction resistance or edge restraint should also be modeled. Modeling the self-weight of the foundation may also be required for stability (this can be modeled using the Gen - Self weight option). If self weight is included as part of the dead load, also make sure that the dead load is used in all of the combos. Because the model includes compression-only springs, iterative analysis is required. It may be that the model is stable for combinations, but not for the lateral load cases alone, but that's OK so long as the combos all produce valid results.
A sample file can be found here.
The method above outlines how to get steel requirements for the shells. Bearing stress results can be derived from the spring reactions.
For something a little more automated, consider Ram Concept.
No, pressure loads on shells are always uniform and normal to the shell surface in Ram Elements. Mesh the shell into smaller pieces in order to create a stepping pressure to approximate a varying force (e.g. due to soil or wall pressure).
You can apply a nodal load anywhere on the surface of the shell and that load will be resisted by the meshed elements of the shell. All that is required is to add a node at the location of the load.
If the load is close to the edge, put the node right at the edge to avoid a tiny mesh between the node and the edge. Ram Elements - Meshing Errors [TN]
Make sure you have the option to segment shells turned on under Process - Analyze - Finite element model tab:
Also make sure you are running version 13.00.03.45, this is a fairly recent enhancement to Ram Elements.
When shells are meshed automatically by the program, a determination about the support for the edge nodes must be made. On the Shells tab of the spreadsheet, the fifth sheet is for "Intermediate Supports", and the first control is a check box labeled "Only at Ends"
If this box is unchecked (which is the default setting) then the program will extrapolate similar supports along the meshed edges of the shell. You can verify the support conditions for the meshed nodes using View - Finite Elements. In the example below, since all 4 corners has some restraint, all four edges are also restrained in 1 or more degrees of freedom.
If you plot or report the nodal reactions for such a wall, you need to report all of the nodes, not just the corner nodes to see the total reaction on the system.
If the option is checked, then only the original corner nodes are restrained, and only those points will have reactions.
Press F1 on the "Intermediate Supports" spreadsheet to get more Help on the "Extent" option.
The "Env" check works for combinations only. If a model has no combinations (only load cases) then the results are zero. If you add one or more combo the envelope will be for all the combinations. (check Max for positive envelope and unckeck Max for negative values).
RAM Elements Masonry Wall FAQ
Ram Elements - Meshing Errors [TN]
Structural Product TechNotes And FAQs
If you manually segment (mesh) a wall you could then delete a few small shells in the field and adjust nodal coordinates of others to create octagonal or other polygonal openings. It's not as convenient as using the opening tool, which is limited to rectangles, and it prevents you from using the design modules, however.
Good explanation on the reactions - that made sense out of some model behavior I was seeing.
Thanks! Any suggestions on how to handle circular openings? Openings with square corners generate pretty high stress concentrations that wouldn't really be there for a circular port in a tank wall, for instance.