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.

 

[drawoh]

stiman86,

This is why the factor of safety sometimes is called the factor of ignorance.

--
JHG

 

[PEinc]

Overdesigning to compensate for your expected but unfound design errors is wrong and costly. Don't make errors. Check your work.

http://www.PeirceEngineering.com

 

[JLNJ]

Errors are common. More common than we'd like to admit. I've made my share.

Big firms have some kind of checking policy (usually). One man shops? Rarely, unless it's big enough to require a peer review.

I found an error today from a vendor who was reporting a load of P. It looked to me like he forgot that his columns were loaded from two sides. When I asked, he told me, "Thanks, they should have been two-sided." Plus some additional equipment loads. The total was 2.4P. I'll probably add 15% to that and use 2.8P.

 

[Guastavino]

so what PEinc says is what most everyone will say on the record...and in a perfect world yes...but overdesigning is what happens with unrealistic deadlines...is it the goal? No, but do you have time to sharpen the pencil on everything? Sometimes no. And sometimes, sharpening the pencil costs the client more if they really need the building more quickly. Especially industrial clients where production losses can be huge.

For me this boils down to judgment and value systems...I'm not saying go crazy with overdesign, but if you use the same size columns to avoid having to design every single one to the nth degree, I say go for it. Repetition also has a way to pay for itself...and overdesign leads to future flexibility.

As for errors, every engineer makes mistakes, and that should be expected. Otherwise you expect perfection, which is not realistic. Building in conservative design to accommodate for errors isn't a PC way to put it, but it does make one tend to want to lean that way when they design things. And, frankly, I believe an argument can be made for being conservative for this reason. For one thing, the cost to "fix" things later would be significant versus minor cost increases in material up front...

I have about 9 years experience now and I used to sweat this a lot more than I do now. The reason? I worked for another firm and the things I was worried about in terms of design weren't even considered by some engineers. When I asked certain questions, people didn't know what I was talking about. What you'll find on this board is engineers that care enough to ask. A lot of engineer don't care, it's a paycheck. So being on this board and asking questions means you care enough that you're less likely to make a big mistake.

I'm a pragmatic realist and I call it like I see it. I've also been accused of being cynical. I don't like the negative connotation with that. The truth and cynicism are often confused, at least as I see it.

 

[271828]

What if your error causes you to be off by a factor of three?

 

[Archie264]

My understanding is that there is a field of study related to the complexity of systems and as the system’s complexity increases at some point a threshold is crossed where there is virtually guaranteed to be a failure. I hope the code writers bear that in mind.

I did a quick internet search on the topic and this was one of the articles that popped up:

http://web.mit.edu/2.75/resources/random/How Complex Systems Fail.pdf

I thought point #14 was particularly relevant:

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. When new technologies are used to eliminate well understood system failures or to gain high precision performance they often introduce new pathways to large scale, catastrophic failures. Not uncommonly, these new, rare catastrophes have even greater impact than those eliminated by the new technology. 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.

 

[Guastavino]

@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.

Case and point is a question i had on this board earlier this week on an embed plate. The initial design was off by a factor of 4ish, but I knew i needed to study the issue more, so I came to this board after doing my own homework and came up with a better solution.

 

[PEinc]

I agree with a lot of what njlutzwe said. Too many different design cases makes for false economy. There are savings in repetition and simplification. However, I have heard many engineers say that they overdesign because they cannot be sure that the contractor won't make mistakes. I disagree. That's an inspection problem. The designer usually isn't responsible for inspection. Design it correctly and let someone else worry about building it correctly. You can't design conservatively enough to prevent stupidity in the field.

http://www.PeirceEngineering.com

 

[Guastavino]

And to PEinc’s point, if it’s too conservative and the contractor knows it they are less willing to take you seriously and they may change something somewhere that is critical because they think you way over-designed EVERYTHING...

A good engineer once told me that I don’t have to design everything to a 1.0 unity ratio, but i needed to have a REASON for why it’s sized the way it is. As drawoh said, this is where the “factor of ignorance” can come in handy...

Also, imagine a situation where you have a 1,000,000 piece of equipment supported on a platform. Let’s say bare minimum with 1.0 unity all the way around, the platform will cost $15,000 to build. Or for $20,000, you can build a platform that will be stout as all get out. So for 0.5% cost of the project we greatly exceed the MINIMUM which is the code. This is a value judgment. I don’t think there is a wrong answer, but i lean towards the $20,000 job and move on.

 

[P205]

Overdesigning to compensate for your expected but unfound design errors is wrong and costly. Don't make errors. Check your work.

Let me give you a scenario so that you can understand what I'm talking about.

Say that, in reading a textbook I bought on steel design, I find out that under certain seismic system, a moment frame structure must have columns of a certain size because of X reason that isn't in the code, but is for some other reason based on tested assemblies. Had I not read this book, I wouldn't know this, and other engineers around me don't know about this. This is what I'm really talking about, not just checking your wl^2/8 calcs and column run-downs.

Another example is, lots of engineers I've met don't consider that a simply-supported beam on one side of a column, causes a moment in the column ( V x e = M).

Building in conservative design to accommodate for errors isn't a PC way to put it, but it does make one tend to want to lean that way when they design things. And, frankly, I believe an argument can be made for being conservative for this reason. For one thing, the cost to "fix" things later would be significant versus minor cost increases in material up front...

I agree with this. I think I should employ this way of thinking more often.

the things I was worried about in terms of design weren't even considered by some engineers.

This is what I'm talking about. This stresses me out! Your typical older engineer who stopped learning and still uses his codes from 20 years ago will just give you the "here, I can design this moment frame building no problem, by hand. See how fast I did that." These types don't even consider that there might be stuff they are missing from continuing education. I just gave one example above, but there are countless things like this that come up the more you read.

@Archie264
Interesting read. Thank you for contributing that.

overdesign because they cannot be sure that the contractor won't make mistakes. I disagree. That's an inspection problem. The designer usually isn't responsible for inspection. Design it correctly and let someone else worry about building it correctly. You can't design conservatively enough to prevent stupidity in the field.

All I'm going to say is, it's often more complicated than you make it seem. You can't simply "let someone else worry about building it correctly" because it's always your problem. You also need to be comfortable enough to sign that letter of conformance at the end of a project. So you absolutely want to have it built right. Some contractor's make zero effort.

Also, imagine a situation where you have a 1,000,000 piece of equipment supported on a platform. Let’s say bare minimum with 1.0 unity all the way around, the platform will cost $15,000 to build. Or for $20,000, you can build a platform that will be stout as all get out. So for 0.5% cost of the project we greatly exceed the MINIMUM which is the code. This is a value judgment. I don’t think there is a wrong answer, but i lean towards the $20,000 job and move on.

Excellent example. I believe this is the right mindset to have for most projects. At least for these fast-paced projects (which is all of them theses days!).

 

[P205]

My understanding is that there is a field of study related to the complexity of systems and as the system’s complexity increases at some point a threshold is crossed where there is virtually guaranteed to be a failure. I hope the code writers bear that in mind.

On this point, I attended a concrete workshop recently, held my cement.ca regarding the upcoming updates to the code for the new NBCC/OBC and I was not impressed. These PhD types who perform the research and write the codes are making thing more and more complicated. There was a portion of the session dedicated to breaking down certain code clauses so that normal people can understand them! I'm not kidding. At one point the presenter had to bring in his colleague to help him clarify something.

It's just plain obvious that if the people writing these codes can barely make sense of them, how is your typical engineer supposed to? How are the people who don't attend these workshops supposed to? And shouldn't we be making things simple enough that the bottom 50% of designers can implement them correctly?

 

[TLHS]

stiman86,

I disagree heavily with the examples you picked.

I agree that there are many mistakes made in design, but your examples don't seem to actually touch on that. Your examples relate to assumptions, models and maybe approximations rather than errors. I can tell you that a hand designed moment frame is likely to be as safe as something done via a complicated finite element analysis as long as the limitations of the analysis are understood. Additionally, treating a simply supported beam on a column as a pure axial load isn't going to fail the structure and is generally going to be negligible. It's acceptability will depend on connection design, detailing, your bracing scheme and other similar things. When Alexander Newman and Charles Carter, two very good engineers, are pretty handwave-y on the topic, I'm going to be careful calling any reasonable set of assumptions wrong if applied appropriately (

https://www.aisc.org/globalassets/modern-steel/steel-interchange/2002/2002v11_si_web.pdf

). The code you're using may have explicit directions on the topic, but not all do and there's more than one way to skin a cat.

There are mistakes and then there are disagreements in methodology. There are a wide variety of choices of models that will give different answers and still be right within an acceptable margin.

More detailed analysis is not always the right path forward. It can easily result in more confusion and errors than using straightforward analysis appropriately. You're significantly more likely to lose the forest for the trees when you have added complexity. There are definitely applications where you want to utilize every tool in the box but you don't want to overcomplicate things if it's not necessary. Being overly precise can also give a false sense of security. Failures aren't in the few percentage points you generally get from refined analysis methodology. They're not even in the 20% error you might get from screwing up the math on one item. They're in an inability to recognize load paths, properly define loads or account for local and global stability issues. Alternately, they're in combined errors or complete fuckups that you don't catch because you don't check that your inputs and outputs actually line up and make sense. This is significantly mitigated if you keep your analysis straightforward or verify with a straightforward analysis.

This is the opinion of someone in the earlier portion of his career who is very much in favour of keeping up with new technique and continuing education.

 

[PEinc]

stiman86, it is the design engineer's responsibility to properly design (not excessively overdesign) a structure. It is the contractor's and inspector's responsibility to see that the structure is built properly. If the designer is also responsible for inspection and/or certifying the construction, then it is the designer's responsibility to first design it properly AND secondly to inspect it properly. These should be two separate responsibilities. The owner is paying a designer, a contractor, and an inspector. It isn't the designer's responsibility to overdesign a structure in case the contractor and inspector improperly perform their work. I'm sure that the owner expects each to do his work properly and would not be happy to know that the designer is inflating the owner's project cost in order to compensate for potential, shoddy construction and inspection by others.

http://www.PeirceEngineering.com

 

[JoshPlumSE]

Some quick thoughts on this subject:
1) A really good, experienced engineer working on a type of project that he has done many times before, can probably come up with a reasonable design without doing much in the way of calculations. Is this wrong? Maybe, maybe not. If the structure does calc out when the calculations are run and it is not excessively conservative, then this is just an example of engineering experience paying dividends to the person who hired that engineer.

2) Now, a less experienced engineer or one working on a type of project he is not as familiar with would not have as good of a feel for what the final design would be. In that case, this is not a good idea and more thorough calculations have to be run. This takes time. You can reduce that time a bit by making conservative simplifying assumptions and such. But, this time reduction can have a major cost to the owner. For me, that's what the OP's question is about. And, there just isn't one solution for all projects.

3) Ultimately, it's the owner's decision. Does the reduction in project completion time justify the increase in construction cost associated with a conservative design. I've seen fire re-build projects for industrial facilities where the lost revenue associated with two weeks of lost operation exceeded the total construction cost of the project. In that case, conservative and fast is much, much better for the owner. In fact, we built a "temporary" facility really, really fast just to stop the losses. But, knowing they would replace the majority of the equipment which was salvaged from another site with new equipment manufactured specifically for that sites specifications. So, in essence it was better / cheaper for them to build the facility twice in a year that it would have been to wait for everything to be economically designed.

 

[P205]

TLHS, I appreciate your input. You do bring up a good point that about disagreements in methodology. The article you posted in which Newman tackles the question of unbalanced load on HSS column is interesting. I have a textbook that says you should consider the moment. I agree with you that it comes does to judgement and that there is more than one way to skin the cat.

PEinc, no one is saying we should excessively overdesign. In general, I don’t feel like I’m arguing to do the things that you’re arguing against doing.

JoshPlum, good points.

 

[P205]

Twice in the last few months I was retained to re-design projects that had been completed by another engineer. The owners had a sense that it was over designed, and the engineers stuck to their design. I was able to save the owner enough construction and material cost to more than offset the fees they were paying to redesign.

All that to say, some people do excessively over design, and I don't consider myself one of those people.

 

[hokie66]

stiman86,
Would you mind sharing with us what type structures you found to be "over designed", and in simple terms how you arrived at that decision?

 

[P205]

The first was a pre-eng building that is to be constructed over an existing ice pad. The first engineer required boring underneath to tie each opposing footing together to take care of the outward thrust that resulted from the frames due to snow loads. By designing a suitable (albeit larger) footing that could resist the sliding, uplift and overturning, I eliminated the need to bore under the ice pad.

TLHS will be happy to know that I made use of Newman's textbook on light metal building foundation design.

The second one was a sign foundation where the contractor had done some funny things to the reinforcement and anchor rods. The original engineer wanted them to rip it all out and start over. When I looked at it, the sign base had about 3 times as much overturning resistance than needed. So I didn't feel it necessary to rip everything out and was able to make it work with a couple small alterations.

 

[Guastavino]

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.

Lastly, people talking about inspection and design responsibility...please...a lawyer will pick that apart faster than you can blink. And even if the argument they have is ridiculous, your insurance provider will be scared anyway.