Checking calculations 19

[P205]

I'm writing this thread in regards to what was said in this old thread here:

http://www.eng-tips.com/viewthread.cfm?qid=321928

It leaves me feeling uneasy to say the least. I also feel compelled to ask the following question:

- In the interest of meeting the deadlines AND feeling comfortable, why not be generously conservative in most of your designs?

- Many admit to finding errors often in their own work and in others. It stands to reason that not all errors get discovered. This would lead me to want to be conservative in my design.

- Another thought, given the old triangle (cost-quality-time), if deadlines are very rushed (low time) and the structure simply needs to be satisfactory (low quality/less refined), then being conservative (higher cost) appears to be the way to solve this puzzle.

Thoughts?

---

Personally, I have 8 years of design experience, and spend a lot of time reading books and teaching myself as much as I can. I usually buy a few books a year to refine my skills and knowledge. Honestly, I don't see too many peers doing that. I'm just looking to stimulate some discussion.

 

[hokie66]

I wouldn't call those examples of "over design". Your solutions may be more logical and economical than those proposed by the other engineers, but they represent different constructability issues, not necessarily extravagance.

 

[XR250]

Could also be different serviceability limits. One engineer may be fine with code minimum while another may want double that. I run into this all the time in residential house design.

 

[PEinc]

The original posting said, "In the interest of meeting the deadlines AND feeling comfortable, why not be generously conservative in most of your designs?" What's the difference between generously conservative and excessively overdesigned? Codes specify safety factors, load factors, and resistance factors. Does an owner really want to pay for an engineer to arbitrarily overdesign just because a contractor may not build it right? I don't think so. Yes, mistakes do happen; some contractors are bad; some inspectors are incompetent. Engineers are often sued even though theirs designed are correct and proper, and there will always be some forensics engineer claiming that the design was incorrect. However, in my opinion, that is no reason to excessively or generously overdesign as a preventive measure. I also am not saying that engineers should design every single individual member. Members can be grouped into several different design cases based an the engineer's judgement. I also have not accused anyone here of excessively overdesigning, so let's not go there.

http://www.PeirceEngineering.com

 

[271828]

njlutzwe wrote:

@271828, if you’re off by a factor of 3 and can’t see it without running any numbers, you are most likely designing something you ought not be designing. Or, if it’s newer to you, something you should be having peer-reviewed.

Forget "3." What if he's bumping up designs by 10% but his calculations are off by 40%? He probably wouldn't realize that error, and he has almost eaten the factor of safety.

This reminds me of balancing a checkbook and arbitrarily subtracting $50 from the check register just in case I've forgotten something. What if I forgot something that costs $80? I need better practices than that, or else my checkbook will be in disarray.

An arbitrary bump-up is a horrendous practice.

 

[XR250]

However, in my opinion, that is no reason to excessively or generously overdesign as a preventive measure

I typically over-design metal studs member and connection strengths as it has been my experience that about only about 50% of what I show on the plans gets done correctly (I may be exaggerating). There have been countless times that over-design has saved the contractor's butt when re-work would have been required. Another reason I may over-design some areas is to simply keep the number of details to a minimum

 

[Tomfh]

An arbitrary bump-up is a horrendous practice.

Isn't this what we do when apply any factor of safety?

I remember talking to a professor about it once and he said "NO! Safety factors are for substandard materials, poor workmanship, and extreme overload, NOT design errors!". I always found that a bit arrogant. I'd wager that in practice safety factors have saved many an engineer's backside.

 

[Settingsun]

Not for design errors, rather the 'imponderables' mentioned in the West Gate collapse royal commission report. If in doubt, make it stout.

Important things can be overdesigned but I'd recommend only doing that if not over-expensive. Eg a beam that is repeated 100 times would be refined but a one-off brace that stabilises a large part of the structure could have some fat.

Also think about what failures really occur. To me, that's connections, buckling, fatigue etc (brittle failures mostly). Something like a slab with substantial reserve capacity through redistribution can be on code limit comfortably.

 

[271828]

Tomfh wrote:

Isn't this what we do when apply any factor of safety?

I'd say no. In modern specifications, load and resistance factors are established using rational methods. They're not arbitrary. The methods are laid-out in the literature, in commentaries, etc. They're not established by one guy guessing in his office.

 

[SwinnyGG]

2 cents from a non-SE mechanical engineer who designs all manner of things (but not buildings):

Honestly, as I read some posts on this thread, I read a holier than thou attitude. Mistakes happen.

First.. I echo this sentiment. We all make mistakes, and it's ridiculous to assume that you could work over an entire career, spend 40+ years performing detailed engineering and calculation, on what is more than likely to be hundreds of unique jobs and many thousand unique calculations by the time you are finished, without releasing a single mistake into the world.

It just isn't possible, even in the age of computerized almost-everything. Maybe even less possible than it used to be, since the computerization of engineering has allowed the code to balloon and become more complicated.

With that said. My work very frequently involves looking at existing designs and adapting them, evaluating their capabilities, etc. If, after doing my first pass of analysis, I feel that the structure or widget being evaluated is massively over-designed, my first instinct is ALWAYS to ask the following question:

What did the designer whose work I am evaluating know that I do not? What load, code phrase, environmental condition, material restriction, etc am I missing?

Just because your gut tells you that something is oversize does not mean that the guy who did the first design is an idiot.

In my opinion, assuming that you can subtract from someone else's work without at least considering what you are missing is a dangerous way to operate.

Just my two cents, gents.

 

[DaveAtkins]

For the most part, errors which could lead to failure are fairly obvious. Sizing a beam as a W18X35 when it should have been a W21X44 will not result in a collapse--sizing a beam as a 2X12 when it should have been a W21X44 will result in a collapse.

In the end, a structural engineer can only do his/her best, then get the work reviewed by a colleague (if possible).

DaveAtkins

 

[Archie264]

Years ago I saw an interview with Joe Paterno where he said that he thought the way to reduce injuries in football was to take the facemask off their helmets…then people wouldn’t lead with their face and would return to tacking in lieu of hitting.

In that vein I’ve come to wonder if we wouldn’t be better off if we not only chucked not only the computers but the calculators as well and just stuck with slide rules.

Neither will happen, of course, but a fella can dream…

 

[271828]

njlutzwe wrote:

Honestly, as I read some posts on this thread, I read a holier than thou attitude. Mistakes happen. If you think you don’t make mistakes, then honestly, all i can say is wow. I don’t believe for a second that stilman is trying to say he intentionally overdesigns things. Please re-read his OP

Also, any good structural engineer knows that it is an art as much, or more so than a science. There are so many layers of assumptions it’s hard to even list them all. Choosing to land on the conservative side of that science is what I think Stilman is trying to say.

That's not how I understand the OP, who typed the following.

In the interest of meeting the deadlines AND feeling comfortable, why not be generously conservative in most of your designs?

- Another thought, given the old triangle (cost-quality-time), if deadlines are very rushed (low time) and the structure simply needs to be satisfactory (low quality/less refined), then being conservative (higher cost) appears to be the way to solve this puzzle.

Paraphrase: I don't have adequate time to be reasonably sure that my calculations are correct. I can fix this by bumping up my design by an arbitrary amount. (Even though I don't know the magnitude of the errors I think I might've made.)

Everybody makes mistakes. It's this manner of dealing with potential mistakes that is objectionable.

Example reasonable strategies: peer reviews, comparing program output to established tables, and visually inspect the sizes and see if they look right based on your experience.

 

[drawoh]

Archie264,

I have seen claims that rugby has more injuries than American football. This may reflect that statistically, an injury occurs when the coach opens a first aid kit. Injuries in American football may be more serious because the people hitting you are wearing padding. This is a poor analogy for structural analysis and checking.

I am a mechanical designer, not a structural engineer. I have argued on other forums that the design checker must catch errors that cost more than the design checking process. My background is that I have prepared packages of ten to twenty drawings for one-off prototypes. Two or three parts may require re-work due to errors, which usually is way less expensive than a day or two of design checking and the subsequent squabbling. If, due to size or quantity, the mistakes will cost thousands of dollars, and/or kill people, checking is justified and/or necessary.

In civil/structural engineering, what is the cost of correcting non-fatal mistakes?

As a non-PEng, I have submitted structural calculations for safety purposes, clearly noting my qualifications. I felt way better when someone at the other end, went through my stuff and critiqued it.

--
JHG

 

[Guastavino]

@271828:

You and I disagree, which is no big deal, it's rare two engineers agree on anything. I read his OP as someone coming to the realization of what goes on in the consulting world with unrealistic deadlines. I read the OP to mean, you can't get into the weeds with everything, so why not be conservative with some of those things and move on, because if you sharpen your pencil too much, you may make an error.

As Structural engineers, we all live in a glass house...so I would recommend we don't start throwing rocks...

 

[SwinnyGG]

so I would recommend we don't start throwing rocks...

Especially if the designed safety factor for projectile penetration of the wall diaphragms of said structure is unknown.

 

[JStephen]

To my way of thinking- if there was a consensus among engineers in general that calcs ought to be arbitrarily bumped up just to be conservative, we'd simply see that written into all the codes and standards. The fact that it isn't done that way implies that there is no such agreement.

Although, it's worth noting that some specific fields do use higher safety factors. ASME vessels, lifting and rigging equipment, for example.

One effect not really noted above is that on one-off type items, it is easy to spend $50 worth of time to save $30 worth of material. So on something like that, a quicker but more conservative approach may actually be more economical.

Along with this, if you realize early on that there may be minor adjustments in the calculations prior to the job being finished, a little bit of conservatism there can save a lot of changes later on.

"I have seen claims that rugby has more injuries than American football."- My older brother played rugby in college and then tried out for the football team. His observation was that American football was actually rougher on the players. In rugby, you had no pads, but only occasionally hit someone. In football, you were hitting someone every single play. I don't know if his perception matches statistical injury rates or not.

 

[Archie264]

I have seen claims that rugby has more injuries than American football. This may reflect that statistically, an injury occurs when the coach opens a first aid kit. Injuries in American football may be more serious because the people hitting you are wearing padding. This is a poor analogy for structural analysis and checking.

drawoh,

If that's correct then it would seem to be consistent with the item I posted earlier taken from the article about systems' complexity. That is,

14) Change introduces new forms of failure.
The low rate of overt accidents in reliable systems may encourage changes, especially the use of new technology, to decrease the number of low consequence but high frequency failures. These changes maybe [sic] actually create opportunities for new, low frequency but high consequence failures.

To wit: unpadded rugby players suffer more frequent but presumably less catastrophic injuries than padded football players. So while padded football players might not get as many bloody noses as rugby players, it seems once per season (or so) a professional football player might get paralyzed from a broken neck. To say nothing of the amount of concussions occurring inside those helmets.

I used that analogy to make the point, perhaps unsuccessfully, that with each refinement of the code to correct high-frequency-low-consequence failures there comes the opportunity to introduce the possibility of low-frequency-high-consequence failures. Especially since the new engineer has to learn it all. Mixing metaphors, the seasoned engineer only has to keep up with this years changes and, much like a frog being slowly boiled, doesn't realize just how hot the water is. At some point expecting the new engineer/frog to jump in the water and think it's fine -- or be able to check calculations against the entirety of the code without missing something catastrophic -- might be expecting a bit much.

 

[Tomfh]

I'd say no. In modern specifications, load and resistance factors are established using rational methods. They're not arbitrary.

What I meant was the safety factors, whilst in theory being there to protect against this that and whatever, in reality catch design stuff ups too.

Any safety factor is simply a layer of fat. If an engineer decides he wants a slightly thicker layer of fat, that's his choice as far as I'm concerned.

 

[drawoh]

...

I used that analogy to make the point, perhaps unsuccessfully, that with each refinement of the code to correct high-frequency-low-consequence failures there comes the opportunity to introduce the possibility of low-frequency-high-consequence failures.

...

Do building codes prevent high frequency, low consequence accidents, and cause low frequency, catastrophic accidents?

--
JHG

 

[Archie264]

>>>Do building codes prevent high frequency, low consequence accidents, and cause low frequency, catastrophic accidents?<<<

Who knows? The ending sentence of that point I posted from that article discusses this possibility:

These new forms of failure are difficult to see before the fact; attention is paid mostly to the putative beneficial characteristics of the changes. Because these new, high consequence accidents occur at a low rate, multiple system changes may occur before an accident, making it hard to see the contribution of technology to the failure.