Abaqus compared to risa, staad, ram 2

[delagina]

What is the difference if I use plate or solid function of risa or staad for connection check or individual beam design as compared to abaqus?

I haven't used abaqus but interested to know what their application in structural design.

 

[Erik Panos Kostson]

These software are not really comparable since the first two software are structural design packages, while abaqus is a general purpose fea software, and not a design package, used predominantly in the automotive,aerospace,mechanical engineering and other sectors. Also very popular within research and academia (e.g., material and bioengineering) since it allows one to build user defined material behaviours, and other advanced functions. Abaqus code strength is like Marc, their advanced nonlinear modeling capabilities (2D and 3D models mostly). So they could be used say for a very nonlinear crash, blast or impact analysis where components get severely damaged and that includes geometrical, material and contact nonlinearities, and not so much for linear or geometric nonlinear analysis of beam and plates that is very often used in structural design software (that might be based on FEM or not).

So in structural engineering abaqus could be used for advanced nonlinear and buckling/post buckling
collapse of 3D structures, perhaps an elastoplastic analysis of steel connections that also include contacts, or for advanced modelling of concrete and soil (material nonlinearities are important here), perhaps also blast analysis (abaqus explicit) so something design packages are not able to do that well (advanced nonlinear anlayis of shell and/or 3D hex/tet/wedge meshed models).

 

[WARose]

I've sat through several demos of Abaqus. It's primary purpose (at least as was presented to me) was for elastic stress wave propagation analysis. (It may look at plastic wave fronts as well, but I didn't get into that.) It is popular for impact and crack/damage detection applications in a variety of industries. For structural engineers, it's typically not useful unless you do some sort of forensics.

Personally, I wasn't particularly impressed with it. For the kind of prices we were talking.....I would have thought it could generate a variety of waves (internally) based on a single contact force. Instead you kind of had to generate them yourself. It also (IIRC) didn't have the Spectral Element Method available. (Which means you are in for some heavy meshing.)

 

[Erik Panos Kostson]

Ultrasonic wave propagation in e.g., solids is a very niche field for FEA, hence it is not really a good representation what abaqus is used for and the main user base. Abaqus was developed by engineers that were developing Marc that is another software that specialises and is known for like abaqus for advanced nonlinear analysis capabilities. I do not know every user but abaqus is used by automotive companies (Link), but also for aerospace structures (Link), plus many other sectors (any structure really could be a ship, where one wants to do structural analysis), and within civil/structural it is used e.g., but not only within oil and gas (Link), and the nuclear industry (Link). So being a general purpose FEA software it can be used for many different things/structures (Link), and probably very often if one wants to get the full benefit when performing structural analysis, within advanced nonlinear analysis (static, dynamic/implicit/explicit like for blasts and explosions).

 

[WARose]

Whatever Abaqus is primarily used for.....I have a hard time envisioning it being cost effective for someone doing day to day structural design. The quote I got from them was around $30,000 a year. By comparison, STAAD cost me 6k for the first year and about $900 a year since. And I use STAAD for dynamic loading and nonlinear analysis all the time.

 

[ThomasH]

Hi

I have worked with STAAD in the past, mid 90s, for some years. The last few decades I have worked with a general software (Nastran), same type of software as ABAQUS.

If I go back to the original question. If you model a connection in Risa, STAAD and ABAQUS. I don't know Risa but if I remenber correct, in STAAD, you have plate elements, that's it. In ABAQUS you will have different 2d elements so you have to make a choice regarding what element is suitable for your analysis. And then it is modelling the geometry.

Once you have your model there will be results, stresses. Depending of the geometry of the connection the stresses may or may not exceed yield. If they exceed, then what? Can you include yield and perhaps also contact conditions in STAAD? In Nastran I do it all the time and I know ABAQUS has the same capacity. That may be neccessary for realistic results.

The point here is not that one is better than the other. But I would say that comparing a structural design package to a general fem software isn't quite fair or perhaps even relevant. For somebody doing normal (whatever that is ) day-to-day structural design work STAAD is probably the better choice. But that statement is because the code (US/Eurocode/?) is included in the software, the required analysis is usually within the softwares capabilities and the price is nicer than for the general code.

But if I compare STAAD and Nastran and skip the code requirement I don't think there is anything STAAD can analyze that I can't do in Nastran. But I know several projects that I have made in Nastran that STAAD could not do.

If you plan to buy something you need to understand what you are buying. But testing just to learn or because it is fun. Sure, why not?

I think ABAQUS primary application from the start was non-linear analysis. I have collegues that use it and that is for bridges, not wave propagadation. That is sort of the idea, it can probably do both.

My 2 cents

Thomas

 

[WARose]

I don't know Risa but if I remenber correct, in STAAD, you have plate elements, that's it. In ABAQUS you will have different 2d elements so you have to make a choice regarding what element is suitable for your analysis. And then it is modelling the geometry.

You have the option of solids in STAAD. And their "brick" elements have gotten much better over the years. (They use to overestimate stiffness.)

Once you have your model there will be results, stresses. Depending of the geometry of the connection the stresses may or may not exceed yield. If they exceed, then what? Can you include yield and perhaps also contact conditions in STAAD?

You can model contact in STAAD. It's going to depend heavily on user input.....but then again, so does Abaquis. As the saying goes: garbage in.....garbage out.

No matter what FEA software you have for connections, one of the biggest obstacles to overcome is stress concentrations. If you model connections that work by the code....you'll get concentrations that exceed allowable principal/shear stresses all over the place. (No matter what the software.)

But if I compare STAAD and Nastran and skip the code requirement I don't think there is anything STAAD can analyze that I can't do in Nastran. But I know several projects that I have made in Nastran that STAAD could not do.

In structural design (i.e. buildings, not aerospace)? I'd be surprised.

 

[Erik Panos Kostson]

The main thing is that these are (Staad and many other) design package, and abaqus , nastran are general purpose fea packages not design, and so a comparison is not appropriate, since they do not have for instance steel or concrete design capabilities say according to EC3 or EC2, like you have in design codes.

In the same way design packages cannot do advanced nonlinear analysis (fracture mechanics, or steel plasticity with iso.and kin. or combined hardening for cyclic loads,....). As far as I can see in staad, material nonlinearities are not supported.

A simple example of advanced fea in civil that would not be possible in many design packages (it would be interesting to see if this can be modelled in a design package, if it does it would be very good)Link

Thus for design according to say eurocodes a design package is the appropriate tool, while for advanced nonlinear analysis (predicting crack propagation on welds in welded brackets with existing cracks, often an analysis that needs to be taken on e.g., Offshore structure for life time extension) as in the link above, a general purpose fea tool like abaqus is more the appropriate tool.

 

[ThomasH]

WARose

I was not aware of that you can model surface-to-surface contact in STAAD. If it also is possible to include a yield criteria to model the local yield effects then connections may be solvable with STAAD. There may be other issues, like modelling bolts, but they should be solvable.

Thomas

 

[JoshPlumSE]

And I use STAAD for dynamic loading and nonlinear analysis all the time.

I think the move towards heavy duty FEA applications (Abacus, Algor, Marc...) has to do with the type of non-linearity that you're dealing with. I tend to separate my thinking into 3 levels of non-linearity:

1)Simplified Non-linearity
a) Geometric: The STAADs and RISAs of the world can easily handle P-Delta analysis. This means your structure still obeys small deflection theory (mostly) and you merely use a simple correction to account for te P-Delta effect.
b) Tension only / Compression only elements: Again, the STAAD's and RISA's of the world can adequately handle compression only soil springs and tension only bracing most of the time. You don't want to model a structure that is entirely dependent on tension only members for stability in those programs. Because they don't (at least not that I know of) contain a true non-linear solver. RISA, at least, iterates a linear solver to try and converge on a solution for tension only members. Now, we had talked internally about moving towards more of a non-linear solver. But, I don't think it was ever considered a high priority item.

2) SAP2000 level of non-linearity:
There are other programs as well (GTStrudl?), I'm just not very familiar with them. By this I mean a better non-linear solver that uses multiple load steps during the solution and adjusts the stiffness along the way. True Gap and hook elements. The ability to consider plastic hinging and such. Honestly, this should be about as far as structural engineers working on buildings should have to go for their FEA. This should do well for building and bridges even those with reasonably complex cable supports.

3) Highly non-linear: These programs don't lend themselves very well to structural applications. They'll do fluid flow and heat transfer and buckling and all kinds of material non-linearity. They load position may change with the deformation of the structure. But, chances are you can only analyze it for one load condition at a time.

I've seen these programs used by structural engineers for very specialized situations. Maybe you've got a cable / fabric structure with a complex configuration. Maybe it's something associated with an aircraft where the deflection changes the lift / loading and there are dynamic / flutter aspects to it. Maybe you're analyzing an atypical material (glass?) in a structure or piece of art for cracking or fracture propagation or something. But, very rarely would you use these programs for the analysis of a typical building.