Structural Loads and Analysis Programs 1

[rodan2000]

I'm relatively new to structural engineering; just took my PE exam. I'm also really interested in programming and computers in general.

When you come across a simple one-story building, just regular old design by hand methods seem to do the trick: satisfy time constraints and produce a good design. When you come across a two- to three-story+ building with multiple irregularities, when do hand methods become inefficient? I feel like with programs like Revit, Risa, etc. there should be a way to utilize computers to reduce the time for design. Whether it be generating loads and distributing them to actual design with a 3D model, I feel like there is always a better/faster/accurate way to do things.

Am I alone in this feeling? Are there programs out there that are capable of doing this for relatively difficult buildings (not talking high-rises) in a reasonable manner? If there are programs, what are some examples? Which ones do you find that are best?

Follow up questions, when does a building become difficult enough to be efficient to model something in 3D (for example, not using a sledge hammer to build a dog house)? Or vice versa, when is a building simple enough to do just by hand?

These are a lot of questions. I apologize. I'm just looking for some insight from other structural engineers. Thank you in advance!

 

[ToadJones]

If they keep adding BS to the codes, soon we will be designing dog boxes with FEM !

 

[msquared48]

Big dog houses take big dog sledge hammers.

I use computer programs for all frames, trusses, and multi-span beams loaded other than uniformly.

Unless it is a high rise, everything else is pretty much by hand or gut feel.

Mike McCann
MMC Engineering

 

[MiketheEngineer]

I use RISA a lot. But building the model can sometimes take almost as long as hand calcs. - I have been doing hand calcs for forty years!! But then again if there are any changes or different loads you might want to check - once you have the basic model built - IT IS MUCH QUICKER.

Kind of like CAD - I can draw as fast as ANY CAD guy - but when it comes to changes - I lose big time.

Also have built some rather complex Excel programs that do a real good job one one story buildings - but that is all. Suppose you could expand that to two or three stories. I like it because I put the formulas in it and know how they behave and why!!

 

[DST148]

Although manual methods may be approximate in some cases and time consuming in others, I feel the designer should be well versed with both - the manual methods and the computer programs. But either way it takes time to gain familiarity and confidence level.
Spreadsheets may be written for most of the engineering tasks that we as designers perform on a daily basis. They are also available commercially and for free downloads. In 2008, we had purchased a package of more than 100 spreadsheets from Daniel Li International for about $1200 and found them to be very useful for our small design office.
We do lot of load bearing masonry + plank building structures which are ten story and higher. I wrote a spreadsheet for approximate analysis of masonry shear-wall framed buildings including design of masonry shear walls. Although we do have commercially available software, this spreadsheet has served more than its purpose.
You mentioned that you are a recent graduate and interested in computers. You may also write stand alone applications for in-house use if you are well versed in any of the programming languages.

Earlier we used to make simplifications of the structures to suit the only available in-house FORTRAN Program and limited computer capabilities. These days with a PC as powerful as the main frames of the 80's at each desk and so many software and spreadsheets available commercially and for free downloads, choosing between manual methods and an appropriate computer program requires a bit of engineering judgement. But I am sure you will learn these tricks over a period of time.
Good luck!

 

[ToadJones]

most of my work is industrial and heavy duty in nature.

For industrial buildings with cranes, I do all my frame analysis with FEM software. It is a huge time saver when it comes to checking different loading conditions. It is also a big help with moving loads.

I will also sometimes use FEM for rather simple beams because the code has made them a joke to check by hand.

Anyone out there find Section H3.6 in the Black (AISC 360-05) book a major regression in torsion analysis?

Obvious hi-jack, sorry.

 

[SAIL3]

I use the computer as a tool...where and when is a judgement call.
Don't expect the computer program to do everything for you...it will always give you an answer, right or wrong, it is up to the engineer to determine if it is correct or not.
It is defintely a useful tool if used judiciously.
As for the AISC code(360-05), I find it burdensome as a practicing engineer in private practice and just leave it at that for the sake of brevity.

 

[TXStructural]

Revit is not an analysis tool, it is a way to speed up interchange of drafting, but not structural models, not yet. Bentley does have a new system to allow CAD models talk to structural models. It is still first generation, but the system works with many vendors' software, not just other Bentley proucts. Autodesk is trying to do the same, but they do not play well with others, preferring to try to force our industry to conform to their software.
You will find that most production offices use things like RAM for structural analysis, member sizing, and computing reaction forces. RISA works too, but requires significantly more input because it is more versatile (fewer internal assumptions.) Going to a pure FEA program, like Algor, is the next step in versatility, but requires another leap in complexity to setup a simple structural model. StructurePoint (formerly PCA) has several useful tools for component analysis and design, but not for design of multi-component building structures.

Remember that a computer only computes, and the real engineering is in layout, loading, understanding behavior, checking for sanity of the output. Do not trust the software unless you know what it's doing and could replicate the method (given sufficient time and resources.) Also, the software output must be evaluated for constructibility. Do not simply take the output and print it as a set of CDs. Level member sizes and/or reinforcement. Avoid excessive optimization where it complicates repetitive members (similarly-situated members should be the same, not almost the same.) This holds true for any building material.

 

[rodan2000]

msquared48: That gut feel doesn't come easily to me unless I've done it several times like wood beams/shearwalls. Steel and concrete aren't at 'gut feeling' yet. We'll get there someday.

MiketheEngineer: That's another thing that leans me towards using RISA, etc. Time savings in the end.

DST148: I believe we have some of those spreadsheets from Daniel Li but some are limited in their use. I've wanted to expand them to be more useful but one of two things happens: company doesn't want to invest time or I think there is something out there that already does the task (like Risa-foundation, etc.) We'll see how that goes. Thanks for the input!

SAIL3: I totally understand the engineering side of things. I even found an error in one of our spreadsheets. I would just like to have tools that are useful and make my engineering design more fluid, accurate, and faster.

TXStructural: I know that Revit is not an analysis tool. It's just for modeling and plan work. It does however transfer between Risa and Robot Structural Analysis, etc. Which makes me think that I might be able to kill two birds with one stone by modeling the structure then exporting it for analysis. I know it's not perfect from what I've heard... but it could be a start.


Overall, I think after this I feel like putting some effort into RISA and RAM. Becoming proficient in those seems like my best option. Do you know if any of other modules like RisaFloor, RisaFoundation, etc are worth it?

 

[ishvaaag]

In terms of productivity when talking of 3D Reinforced Concrete structures I haven't seen anything like CYPECAD. It houses tens of codes inside for a wild variety of countries. This praise coming from one that doesn't like its being quite a blackbox input-output, says enough.

The only trick is to have dxf or dwg plans that you import to the program. Then everything flows easily.

 

[DST148]

rodan2000 - I agree that spreadsheets have limited use. But then even the commercially available software have some or the other limitations, no matter what the vendors proclaim.
I have mostly worked in small design offices which neither have time nor money to invest in resources and employee training but still expect the output at the end of the day.
I don't mind spending about 2% of my salary on books, seminars, workshops,professional memberships etc. over and above my personal time in writing some of the spreadsheets and stand alone design programs to speed up the work. But again this is strictly my point of view.

 

[nicam]

TX Structural i think you are a bit hush on software, sure the garbage in garbage out saying is true, but we engineers have had many years of training to understand the garbage in garbage out theory. The don't trust software argument is valid for an untest new software or untrained people.
How can you replicate a structure with DL + Ws + moving loads with vertical & lateral loads by hand methods? It takes 1 hour for my pc to do that calc, what would the boss say when your spending 3 weeks analysing a structure like that!

 

[JoshPlumSE]

Rodan -

RISAFloor is similar in concept to RAMSteel. It specializes in the optimization of composite steel beam or steel joist systems. Optimization for deflection, stress. Partial composite action.... et cetera. In theory, the calculations should be pretty similar to what RAMSteel gives. There are a few key issues affecting a RISAFloor vs. RAMSteel decision.

1) What interface are you more familiar with or do you prefer RISA-3D or RAMSteel? If you are not very familiar with RAM, but like RISA's interface, then you'll probably very much prefer RISAFloor. If you are already familiar with RAM and use it constantly, then the interface differences in RISAFloor could be frustrating. At first we always tend to prefer what we are most familiar with. I think that's especially true with software.

2) Designing the lateral frame. RISAFloor interacts directly with RISA-3D for the lateral analysis. Therefore, the user should have a lot more control than in RAMFrame. That's both for adding/ modifying input and reviewing output. And, if you're already familiar with the 3D interface then this is something you may really like.

3) Interacting with Revit. RISAFloor / RISA-3D have put a lot of effort into their Revit link. That's a commitment that is not likely to change. Not sure if the RAM / Bentley guys have the same level of commitment. That's not a knock on them, just an assumption that their main focus will be linking with the Bentley BIM program.... not putting a tremendous effort to link to one of Bentley's main competitors.