Calculation Packs

[Trenno]

In the UK, for every new structure that is erected it needs to be reviewed and approved by Building Control which is a local government authority. This will include reviewing the drawings, specs and a calculation pack.

Going through this process on a recent project I had a few thoughts that I'm keen to get others opinion on:

1) Similar to the emerging trend of BIM-only deliverables (ie no 2D drawings), do you think we'll reach the stage where all of our calcs will be sat within an analysis model file rather than printing out 100s of pages of output? In my opinion it's much easier to whiz around an analysis model to check a design than sifting through endless reams of paper.

2) How do go about striking the right balance between too much calculation and not enough. For example, are you providing calculation output for every single column on the job or are you providing an indicative calculation for one column and summarising the rest?

3) Does your company have templates for how a calculation pack should look, similar to a pro-forma? For example, got a shear wall lateral system? Ok good - present this output XYZ.

4) I think a thorough basis of design report and solid set of drawings should be able to tell you much more about the adequacy of the structure than a long-winded calculation.

 

[IDS]

It's a long time since I worked in the UK, but in Australia it has not been standard practice to provide detailed design information at that level for many years (at least in my experience, in infrastructure design).

Standard practice information for detailed design review is a design report + construction drawings. The design report consisting of a summary of design inputs (with sufficient detail for the reviewer to extract all information required for an independent analysis), a summary of the design procedures used, and design outputs with sufficient detail to confirm that the drawings are consistent with the analysis and design carried out. The reviewers will then have sufficient information to carry out their own analysis and detailed review.

Reports (and drawings) are usually supplied just in pdf format these days. To my mind providing everything in BIM format would not be a good move, since it would be more likely that invalid assumptions in the design would be buried in the computer data, and would not be picked up by the reviewer.

Doug Jenkins
Interactive Design Services

http://newtonexcelbach.wordpress.com/

 

[skeletron]

1. Yes and No. I think the purpose of calculation packages is to document the final design results. Printing your output would accomplish this. But I also see the model as being the single piece of documentation and know that connection design has been heading toward automating that and incorporating it into the model as much as possible. So, yes and no.

2. My structures are limited in complexity, so, I document worst cases and then explain how others fit under that threshold. Personally, that is what engineering is to me: learning how to threshold design in multi-variable (cost, construction, fabrication, behaviour, historical context, etc.). I do appreciate that for higher importance structures that a validation of every piece of material coming through the door is required.

3. I'm a lone wolf, so my calculation template is pretty dynamic based on the expectation of the client. I tell the story with my calculations because I'm often filling in the blanks for someone that doesn't have the time to dig into it. In that respect, I'm documenting design decisions and telling the story which is arguably more important than anything else. The numbers should (at this point in a career) pencil out.

4. Agreed. Printing out ASCII output and putting it in a binder doesn't tell the story and is realistically not supposed to verified by a human. A design report, a good set of drawings, some calculations for the rational design components, and a conversation is really what should be provided. AHJ's in Canada haven't resorted to calculation package submittals which is kind of a blessing in disguise, I've never had to deal with that scrutinization.

 

[NalaCap0ne]

Wanted to toss in my 2cents as a structural engineer currently working in the UK. First off on your opening bit, I'm not sure how much building control actually "reviews" the calculations. In my experience (which is limited to facade works), they typically expect a calc package to be submitted but dont really go through the calculations...they just sort of tick it off the list t hat you've submitted them. This in my experience has led to a lot of issues where the client says things like "oh but we have building control sign off!", when in reality that means nothing when a cladding panel falls off the building.

I digress....back to your questions:

1) I personally think one day we will get to this point, however I dont think it will be any time soon. Several things are going to need to happen before this occurs in my opinion:

[ul]
[li]affordable computers will need to get extremely more powerful to run such massive models. Especially when you start to consider connection design and things like that. I may be a bit biased because in facade engineering we deal with alot of small connection design between rails and other elements (this may be a bit more streamlined with steel structures for example). [/li]
[li]software standardization - sort of how we have a "standard" 2D drafting software/file type in AutoCAD, one of the many many many FEA analysis tools will need to sort of take reign before the collaboration required to review and share large calculation files is possible. When youre thinking in 2 dimensions, this isnt as much of a problem because at the end of the day you can take any CAD software and print to PDF...but a large 3D model as we are talking about would be considerably more difficult to share considering the size, file type, etc.[/li]
[li]old people need to die - While I'm not particularly young, its no secret that the older generation of engineers can be a little stuck in their ways. We all know an engineer or two who still dont like using CAD software. Unfortunately, alot of times its these old foggies who are in charge of reviewing these calc packages...so until there is a wide spread acceptance of 3D modelling, I think 2D drawings and calculations will still be around[/li]

I've had experience working on a paperless job, and I can confirm that it stayed paperless for about 20 minutes. The other thing we need to consider is how the information we engineers output is communicated with the builders and checkers on site...they dont like carrying around expensive tablets etc...they prefer paper drawings (as do I).

2) I typically work with fairly repetitive constructions, so I typically concentrate on the worst case areas and sort of pass all others by inspection based on that. When you get into more complex geometries and things like that, the amount of calculations you need to produce get larger. Like you, I do think there is a happy medium somewhere in the middle. Perhaps creating a large model to cover the repetitive "standard" bits, then a set of hand calculations to cover the fiddly areas. But as of right now in our 2D world, Our calculations are done as cleverly as possible to reduce the amount of calculations needed.

3) We dont really. Because thing change so much from job to job, the calculation packs can look quite different. Again, I think a template can usually get you 90% of the way there, then the extra 10% requires some bespoke work. In the world of facades, if you need a rain screen calculation for a well known system...that can pretty much be completely done in a standard spread sheet. However, as soon as you introduce some weird returns and differing fixing conditions...things go sideways.

4) I definitely agree. Especially considering again that I dont believe people at building control even really go through the pages and pages of output. Our reports can be pretty long, but they are typically summarized in a few pages at the end, and I bet this is really the only thing people reviewing our calcs looks at most of the time. Unless the person reviewing the calculations is an engineer and has the time to go through them with a fine toothed comb, then they are just going to look for the summary and tick a box.

At the end of the day, having a paper trail and submitting all of these documents is typically to cover our asses really. The mistakes that lead to failure are typically made somewhere between the final drawings/calculations being completed and the actual structure being completed. How the completed information is shared is, in my opinion, the most critical aspect of actually constructing a structure. As engineers, we can create the most detailed and accurate model known to man...but if the guy on site says sod it because his tablet froze when they were referencing the drawings and just does what he thinks is best...then the model doesn't mean anything anymore.

so TL;DR, one day we will have a utopia of 3D analysis models...but we still have a long way to go until that utopia becomes a reality.

 

[cliff234]

Given that most structural failures are connection failures, ask yourself this, "Does the software we use today to design building structures analyze and design every single connection?" Until the answer is "yes", manual calculations (and review of all connections by an experienced engineer - not a computer) will be required.

I suspect that most building code officials who request calculations want to see the calculations to verify that there are calculations - and hopefully those calculations will be more than just thousands of pages of computer output.

 

[spieng89]

1) Similar to the emerging trend of BIM-only deliverables (ie no 2D drawings), do you think we'll reach the stage where all of our calcs will be sat within an analysis model file rather than printing out 100s of pages of output? In my opinion it's much easier to whiz around an analysis model to check a design than sifting through endless reams of paper.
That'd be tough logistically. The reviewer would need every engineering software available.

2) How do go about striking the right balance between too much calculation and not enough. For example, are you providing calculation output for every single column on the job or are you providing an indicative calculation for one column and summarising the rest?
I always try to be on the lighter side and summarize calcs where possible. The easier it is for reviewer to digest packet, the less jerk questions they will ask me.

3) Does your company have templates for how a calculation pack should look, similar to a pro-forma? For example, got a shear wall lateral system? Ok good - present this output XYZ.
I wish

4) I think a thorough basis of design report and solid set of drawings should be able to tell you much more about the adequacy of the structure than a long-winded calculation.
I agree, but most drawings are far from great given the time constraints on many engineers. The calcs help paint a picture of methodology and load path that may not be apparent from the drawings.

 

[rowingengineer]

How would the reviewer be able to measure the thickness and weight of the paper to determine if it is appropriate submission.

 

[WARose]

2) How do go about striking the right balance between too much calculation and not enough. For example, are you providing calculation output for every single column on the job or are you providing an indicative calculation for one column and summarising the rest?

Every single one. Usually as per STAAD's output. (And you can control how much output there is.)

If you are looking to save space.....one thing I just don't get in a lot of calculations packages I see today is all the reproduction of code in it. You'll see just about all of ASCE 7 & ACI 318 stuffed in there.

In my day, we just noted what table the value came from and went from there. (And that's the way it was and we liked it! )

 

[Slurgi]

1) Similar to the emerging trend of BIM-only deliverables (ie no 2D drawings), do you think we'll reach the stage where all of our calcs will be sat within an analysis model file rather than printing out 100s of pages of output? In my opinion it's much easier to whiz around an analysis model to check a design than sifting through endless reams of paper.
I think there's too many ways/tools to use to be able to standardize everything.

2) How do go about striking the right balance between too much calculation and not enough. For example, are you providing calculation output for every single column on the job or are you providing an indicative calculation for one column and summarizing the rest?
Provide calc for most heavily loaded member for a given reinforcement layout.
Summarize the rest.

3) Does your company have templates for how a calculation pack should look, similar to a pro-forma? For example, got a shear wall lateral system? Ok good - present this output XYZ.
People design things in different ways, so no.

4) I think a thorough basis of design report and solid set of drawings should be able to tell you much more about the adequacy of the structure than a long-winded calculation.
Depends on the quality of the drawings I guess..

 

[KootK]

With regard to #1, I'm just going to go ahead and dream a big, unlikely dream...

1) I do think that model handoff for review will eventually be the way of things. You know, in a galaxy far, far away...

2) I think that, in combination with the design basis and the drawings, the model probably is the most meaningful thing to review. When you look around this forum and acknowledge how young designers are designing their structures and the kinds of errors that they're making, modelling stuff predominates.

3) I feel that this will require a technological advancement. Namely, that all FEM packages will need to be able to output a locked, read only model to some kind of common, 3rd party, open source FEM package that anyone can use to review a model and the designs within it.

4) This ties into my opinion that, eventually, all projects ought to be reviewed by other, independent, anonymous, practicing structural engineers acting as the reviewer. That, either instead of, or in addition to, AHJ review.

What's the point of dreaming if you can't dream big...

 

[TLHS]

Koot, as far as the read only thing, you can already stamp electronic files and models using the electronic stamps in Canadian jurisdictions. I haven't done it yet, but you can use a pdf/a 3.0 file as a container and then apply the stamp to it.

As far as the thrust of the discussion, I think we're a fair way from just handing over a model. The model itself is only definitive in the most straightforward of structures. The model is a tool in the analysis, not the analysis.

The big thing for improving the checkability of packages in my company has been our push to actually document design basis, intent and judgements done based on the calculations. The narrative of the calculations is where you have the best chance of finding mistakes. A model doesn't have that.

 

[bugbus]

I have always found it difficult to summarise within a fairly succinct design report some of the more complex/nuanced areas of a design, including:

Moment redistribution within indeterminate, ductile structures; judicious allowance of slightly under-capacity members taking into account the redundancy of the structure, ductile behaviour, tolerance for localised damage or permanent deformations, or some combination of all the above; tabulating which is the critical design load combination (when there could be several different ones that have been considered and have influenced the design of a particular member or connection); attempting to explain in simple terms why (based on MCFT, which is stipulated here) the shear strength of a concrete beam or slab is not a fixed number but depends on the accompanying bending and axial components, which probably differ wildly for the many different members and load cases; etc.

Unless an independent reviewer is going to review in detail the entire design, the design report and calculations are almost useless in my opinion, other than to provide some pretty broad context.