Other than checking/QA/QC, how much hand calcs are you doing? 2

[snipit]

Other than super simple designs such as concrete beam(Mu/4d) and retaining walls level of simplicity. For more complicated stuff you studied in graduate program, did you actually use it? I feel like software will do everything for you. What do you think?

 

[HTURKAK]

The enhanced softwares will do if input data and model is correct otherwise ( GARBAGE IN , GARBAGE OUT ). I remember an old quote (may be 50 yrs ago) Regarding Computers,

I really hate this damned machine; I wish that they would sell it.
It never does quite what I want
But only what I tell it..
(From the fortune database, Berkeley Unix)








Tim was so learned that he could name a
horse in nine languages: so ignorant that he bought a cow to ride on.
(BENJAMIN FRANKLIN )

 

[271828]

I feel like software will do everything for you.

I'm not sure if you're serious. We're nowhere near the point in history where "the software can do everything for you."

Usually, there are quite a few required calculations that are outside the scope of commercial software and custom spreadsheets/Mathcad/Smath. They're one-off calcs that just need done quickly rather than programming, validating, and book-keeping/organizing for later.

Even if the commercial software covers your situations, you need to be able to verify its results.

 

[JLNJ]

I still do quite a few hand calcs. My calcs contain a lot of text in order to tell the story. I'd rather have a few calcs which tell the story than a pile of output which is mostly numbers.

 

[snipit]

The enhanced softwares will do if input data and model is correct otherwise ( GARBAGE IN , GARBAGE OUT ).

EXACTLY. Given you input correct values(you understand how softwares work). Don't softwares do everything for us?

I'm not sure if you're serious. We're nowhere near the point in history where "the software can do everything for you."

Yes, I'm serious, given I'm only 1.5 yoe, which I know nothing, hence, I asked the question. So, since the softwares are nowhere near that point of doing everything for you.

Usually, there are quite a few required calculations that are outside the scope of commercial software and custom spreadsheets/Mathcad/Smath. They're one-off calcs that just need done quickly rather than programming, validating, and book-keeping/organizing for later.

Could you please give some examples of what is it that you use the spreadsheet/Mathcad for?

Even if the commercial software covers your situations, you need to be able to verify its results.

That's why the post title started with "Other than checking/QA/QC"

I still do quite a few hand calcs.

Could you please give me some examples of what they are?



Thanks!

 

[jayrod12]

I do almost 100% of my schematic designs by hand. The reason for this is I don't do my own drafting in our firm, we have a specific drafting department and they're generally fairly busy. Therefore if there are still potentially major layout changes based on the outcome of the SD submissions, we try not to involve the actual draftspeople too early.

Also, software is all well and good once the building shape and layout is locked in. But if there are still potentially big changes, the time it takes to modify a model is at times very significant and therefore I'd rather not waste my time building a 3d model in software that isn't relatively locked in location wise.

I can design a 3 storey steel framed building by hand, with my regular non-graphing calculator, far faster than almost anyone can model it from scratch. At least when we're talking high level design, not the nitty gritty detailed designs.

Lastly, for me at least, I find designing the entire bulk structure by hand initially allows me to build the model with good ideas of what member sizing should work in each location. Then when the model runs for the first time, I can determine modelling errors, or calculation errors, quickly just by looking at the member results. A member that is extremely highly stressed, or vice versa extremely lowly stressed, would be reviewed in more detail to determine whether my hand calcs were incorrect, or there's an error in the model somewhere.

I'm only 10 years since I graduated with my degree (I had 5 years of industry experience before I finished my degree), and I still wish they'd stop pushing the software designs in school. It's not preparing the future STRUCTURAL engineers for what the job actually is like.

 

[phamENG]

Before deciding to trust any software, you should validate it. Often that involves taking a not too simple but not too complex problem and solving it by hand using those "grad school" methods. Then you build it in the software and see what it gives you. If it's close, great! If not, troubleshoot your work AND the model to see who made the error. Sometimes this happens on the individual level (if you're the one sealing the design and taking full responsibility), and sometimes on the institutional level (the business that is on the hook to pay for damages if any of the designs from the software fail).

Once it's been validated, then you have your QC checks after the fact.

But as others said, there is no software that does everything for you. They all have their limitations. What you do by hand is based on the limitations of the software. It may be unique connections, maybe handrail design, maybe you have a threshold for when it's worth it to build a model. If your software takes a minimum of 6 hours to build a model regardless of building/project size just to get the inputs correct and checked and then build and analyze the model, but your proficiency with hand calc methods can get you an acceptable answer in 3 hours, then use hand calcs.

 

[271828]

Could you please give some examples of what is it that you use the spreadsheet/Mathcad for?

Sure.

A steel beam shear connected to an HSS column. The HSS is too narrow for a double angle, and shear tabs aren't allowed for whatever reason. Use a WT welded to the face of the HSS. Connect the beam to the WT stem.

Steel collector beam with slender web. Top flange continuously braced by the deck. Bottom flange free between the supports. Bending plus axial compression. For the Pn calculation, the applicable modes are strong-axis flexural buckling and constrained axis buckling about the top of the beam.

 

[driftLimiter]

Software packages significantly lack detailing ability. Even if you can design the primary elements with software, its far from doing everything. Hand calculations help keep me grounded its easy to recognize when an error is made, and it keeps my focus on the correctness of the output.

The software might do everything needed for a technician, but for an engineer there is much more to consider than what is input or output from a software package.

Its just a tool that we use, anyone can swing a hammer but many will miss the point.

 

[CrabbyT]

Contrarian here. Sure, I calculate stuff by hand sometimes. But I definitely don't do as many hand calcs as the generation that preceded me. It's like my grandpa telling me how he had to look up values in a trig table. I showed him what modern calculators are capable of, and he laughed and called me a cheater.

It's good to know basic formulas like Mu = Fy*Z for beams whose top flange is restrained and where lateral torsional buckling isn't a consideration. But let's get real, when's the last time anyone here broke out a pad and paper and calculated the lateral torsional buckling capacity of a beam? ...well I guess I did it about a month ago, but probably hadn't done it in a decade leading up to that.

About a year ago, I did a bunch of hand calcs to validate the composite bending capacity of a bunch beams.

Usually, after doing hand calcs several times, and if it becomes evident that I'll need to do that calc again, I'll take my hand calcs and turn them into a spreadsheet. I never start off making spreadsheets - I always do the hand calcs first. That way, I can validate my spreadsheet using my calcs.

I also do hand calcs when doing rehab and retrofit. Sometimes I have to check capacity using an outdated formulae from codes that have since been superceded because software doesn't include code checks for things like the 1998 UBC.

Long story short, software is cool. It calculates stuff way faster, and I'm glad I don't have to manually check 47 limit states with 200 load combinations. But in my opinion, we shouldn't rely on software for everything, and we have to exercise good judgement when using software as a tool.

 

[human909]

Yeah I basically don't use hand calcs for most structural calcs. I don't really even use that many spreadsheets. For most structural analysis I use software.

But sure I do use hand calcs for quick and specific decisions. A detailer with a question yesterday. We needed to reduce the edge distance for a bolt, my gut said it would be fine. The hand calc confirmed it.

I do have spreadsheets for some very niche applications that software doesn't do at all.

 

[XR250]

I don't know what a typical SE curriculum looks like now but my sense is that software is pushed pretty hard versus focusing on the fundamentals. I had several classes where we were not even allowed to use calculators. I mean, this was in the mid 80's, but whatever.

 

[human909]

I don't know what a typical SE curriculum looks like now but my sense is that software is pushed pretty hard versus focusing on the fundamentals. I had several classes where we were not even allowed to use calculators. I mean, this was in the mid 80's, but whatever.

I finished my study not that long ago. We barely touched software. IMO more software modelling would have been better. As one learning advantage of software is you can quickly 'play' with more complex structures and learn plenty along the way.

That said I cannot agree more strongly with people warning the dangers of relying too much on software. You NEED to know and understand stand what it is doing inside that black box.


Oh and for what it is worth while I might be a "I don't do hand calcs" guy. I did complete a mathematics degree in years before my engineering degree. No calculators allow in that!

 

[driftLimiter]

It seems like different building materials and building types are better represented by the various softwares available.

If your building is wood, masonry, or CFS. I find it hard to believe that the software out there can handle all of the necessary calcs. Connectors, shear transfer details, diaphragms, anchorage etc. All of these require at least one intermediate step beyond the structural analysis package.

For steel and concrete buildings maybe you could argue the bigger packages for structural analysis can handle more calcs, but still connections (concrete joints!), anchorage details, shear transfer, diaphragm design. These calcs can be done by software or spreadsheets but still require taking results from one package to another.

Often times going through all of the setup necessary to make a structural model / automated design is more time consuming that solving the bending moment by hand and looking up the beam section out of the manual.

For most buildings I work on, you would hard pressed to make a single structural analysis model that designs all the primary framing members let alone detail the connections and lateral system.

But I will give credit to the computer because the code has gotten more complicated and there are many limit states and load combinations that need to get run. Without the software many of these calcs are impractical in today's world. There is a balance, I guess I would just emphasize that prudent engineers aught not to be too dependent on any software.

I think that the industry would agree if you consider that the SE test is a manual test, no computers other than your dumb calculator. It is important that engineers understand the calculations deeply otherwise the software tool is going to be garbage in garbage out. Once a good balance is struck, the software tools become a very powerful option for modern engineers.

 

[human909]

True, quite true.

I'm mostly a steel guy and currently don't do much else. So software works fine for me. I do plenty of complex industrial structures, equipment and vessels. I do lean heavily on FEA for the equipment and vessels. Sure I could hand calc an answer out most of the time but it will be excessively conservative in some areas.

I'm going to be picking up some residential work soon and that will be mostly excel spreadsheets from a colleage. But I'll have to hand calc my way through those spreadsheets to understand what is happening as I am pretty clueless about timber.

 

[centondollar]

I would not praise an engineer who claims to rely heavily on hand-calculations. For anything more complicated than a single or two-storey frame, the process of solving the strength parameters (e.g., bending moment) and deflection for all members is excessively time-consuming and a client will not pay for it - a complete building or bridge is of course completely out of the question.

Another issue is calculation errors, big or small. The FE software (if it is at all reputable and well-established in the industry) does not do them and if the engineer understands what to give as input, how the analysis is performed in the "black box", what element formulations are used (and their possible drawbacks) and how to interpret results, then there is seldom a reason to do structural analysis (in an industrial setting where load combinations are not countable by one hand and the computations may have to be repeated several times to find suitable cross-sections etc.) by hand.

Design, be it by spreadsheets or "hand-calcs" in some symbolic algebra software, is not hand-calculation in my mind. Rather, it is often expected, since there are not yet many software packages on the market with both design capability and good documentation, and the design often involves many types of engineering judgement that a computer program cannot make in an optimal manner.

The computer cannot tell you if you should opt for stiffened girder webs or a thicker web, or if a frame should be braced or rigid, or if you should opt for a truss instead of a PT beam, - you must decide that yourself, and that is engineering these days. The days of number-crunching with 20 colleagues for weeks on end to produce what a FE software does in 10 seconds (after a day or a week of one person creating the model) are over, and we should all be glad for that.

 

[milkshakelake]

I don't know how engineers in the old days did load combinations and still made money, or didn't skip a huge parts of the design. Without software, it's like being in hell. It's fine if there's a few loads, but when getting into wind and seismic, it's nearly impossible. I do wind lateral calcs by hand for wood (I haven't found good software for that yet) and oh boy, it's a workload when it's not obvious which direction will control. Then I start getting into things like uplift and axial loads on a holdown-type jamb with gravity beam attached and it becomes tedious real quick. I just start specifying massive holdowns everywhere.

 

[driftLimiter]

...then there is seldom a reason to do structural analysis (in an industrial setting where load combinations are not countable by one hand and the computations may have to be repeated several times to find suitable cross-sections etc.) by hand.

This implies that a 3d model of the building is being used. For custom homes, even midrise wood buildings its often more trouble than its worth to make a model for it.

Pretty much anything with a flexible diaphragm and good luck making a full 3d model that you can use for both vertical and lateral analysis.

And to your point about documentation, I suppose your right when I am thinking of 'hand-calcs' most of it is just good documentation to support the computer output. I do alot of 'hand work' that is just backup info or detailed info about what went into the software and what is coming.

But back to the OP's point, he says:

I feel like software will do everything for you.

I still think No, absolutely not. Should the software be used for all of the number crunching, yes. But there is a lot more to it that needs a guidance from sound engineering judgement.

 

[milkshakelake]

@driftLimiter I think software can be used for almost anything, but if it's not on the market, you can develop it yourself. That's the hard part, and something I struggle with every day (spreadsheets and VBA stuff). I tried to dip into Python but the learning curve is huge.

I agree 100% that engineering judgment should be used, though I think OP understands that based on the response. OP might be doing pretty standard stuff (like steel) and doesn't usually come across situations not covered in software. That problem arises a lot in wood, not so much steel.

 

[Lomarandil]

Absolutely depends on material, and on industry.

I'm thankful to be in a market niche that's unique enough that I need to perform my own hand calcs (even those automated by MathCAD/Excel) nearly every day. It's the fun of being an engineer. And besides, things like LTB capacity are simple enough (in the basic iteration) that they don't take long at all.

I've previously designed multi-story buildings by hand calcs (excluding the frame analysis). If the building is somewhat regular, not a big deal, and it does help with understanding the detailing process. Granted, when the building is architecturally complex, that quickly becomes infeasible.