Communicating structural calculations 13

[spaderess]

There is a disregard for the handicraft of writing and performing legible sequences of calculations in my immediate context. Trying to check or re-use even my own, let alone others, calculations is a pain.

I have been looking into literature or standards in an attempt to better this. Yet, most of the stuff I found is centered on general mathematical writing and therefore not as applicable. The only books I found directly dealing with the specific problems are Robert Motes books The Engineers Word, and The engineers Tables. Standards, on the other hand, seems to be specific to firms and not accesible from the outside. This lack is surprising, coming from an architectural background, where all aspects of drawings are highly standardized.

Therefore, my questions are these.
1. Are there any literature or widely available standards dealing with presentation and/or performing of structural calculations?
2. Do you know of any examples of sets of calculations available, to be used as a best-practices-example?
3. What do you consider to be essential qualities in regards to clarity when checking others calculations?

 

[spaderess]

Thank you all for the great responses. This is truly helpful.

Hokie66 - I will try to implement that. Your advice is similiar to that given for general mathematics, which ill reconsider. My previous objection were that the plain amount of calculations + meta-text makes the whole document balloon. I guess its an experience call, knowing what is essential to communicate.

WARose - Thank you for the tip. I bought it at the spot, looks like a sound investment.

Motorcity - Do you treat the references as footnotes, or is there any specific way of referencing? We work mainly from the Eurocodes, so theres not that many obscure equations. I can imagine thorough referencing also help in standardizing the calculations between people, so that the use and context is easier to read up on later.

Drawoh - Word seems ubiquitous in this context, although Ive never understood its uses over Latex. Would you consider equations written in plain-text to be readable, or is the equation engine of word/latex critical? Ive been experimenting at home with a Markdown workflow recently, primarily using the fantastic software Calca and .css to standardize the pdf output. Its similiar to Mathcad, but with plain text at the cost of a cup of coffee. Problem is, it just doesnt look as pretty as a true equation engine does, although using Latex for the math and markdown for the rest is a possibility.

Kootk - Do I get your #2 point correctly as that the software which allows easier combinations of calculations with images support actual understanding, and therefore help prevent the large mistakes? That sounds reasonable, considering how useful it is to diagram next to hand calculations.

IRstuff - Yeah, its a world of difference between the readability of excel and mathcad. Ill try to emphasize that aspect a bit more.

 

[hokie66]

I like calculation sheets, or computer output, which have separate ruled columns, usually on the right, for references and explanatory notes.

 

[MotorCity]

spaderess - we only formalize calcs if they are submitted to the municipality or client for review. It varies by engineer how references are made. Most people simply provide a list of codes and references since anyone checking calcs should be familiar with your equations and where to find them. However, there are some people who reference each section of the code for each equation they use. I see that as being tedious and unnecessary. With regard to everyone using the same software or method for calcs, that will never happen in a large office. We have several software packages that essentially do the same thing. Some are more user friendly and robust than others, but engineers are free to use any software or hand calcs they choose. I don't think its reasonable to expect everyone to be proficient at using the same software. Engineers come from various backgrounds and have different technical abilities (or lack of). As long as you produce a sound structure at the end of the day using reasonable techniques and resources, you good to go.

 

[drawoh]

spaderess,

Microsoft Word has a perfectly functional equation editor. Click on the Insert tab.

I use Linux at home. I have learned LaTeX and I like a lot of stuff about it, including its equation display. This is what TeX/LaTeX was originally written to do. I can send the LaTeX files out in PDF format. The equations do not display well in HTML, and I don't think they display at all in RTF. I have not tried to do the actual calculations in Octave, to be integrated with LaTeX. This would be nifty and it should be possible. Octave calculations can be processed into graphs using GNUplot, and integrated with the LaTeX file. LaTeX can be run from a GNU Makefile, which is convenient.

Microsoft Word does an adequate job of formatting and displaying equations. I can send the results out as a readable and writable file, to be reviewed by engineers. I have done a number of safety calculations recently, and the results either were vetted (thread507-409684), or they ought to have been vetted.

--
JHG

 

[IRstuff]

There are indeed a number of programs that can create equation graphics, but aside from Mathcad, TKsolver and a couple of others, none can use those beautiful graphics in an actual calculation. Therein lies the dilemma; documentation by those graphical approaches leaves open the possibility that the documented equations are possibly incorrect, or that the actual calculations are different from the graphics. And if the actual calculation is "tweaked" for whatever reason, you now have to go back and correct the graphics.

While Microsoft Word Equation Editor is essentially a free add-on, Mathcad is, unfortunately, VERY expensive. TKsolver is more affordable:

https://www.uts.com/ItemDetails.asp?ItemID=0100-50-0010-00

There's an even cheaper option, which is Mathlook for Excel:

https://www.uts.com/ItemDetails.asp?ItemID=1100-40-0000-00

I've never used it, since I have Mathcad, but it might at least get you a pseudo-self-documenting worksheet.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[KootK]

I don't think its reasonable to expect everyone to be proficient at using the same software. Engineers come from various backgrounds and have different technical abilities (or lack of). As long as you produce a sound structure at the end of the day using reasonable techniques and resources, you good to go.

I agree. I wasn't suggesting that everybody be proficient at the same software. And I recognize that 100% standardization is wishful thinking. What I'm talking about isn't ETABS/RAM stuff, it's the dozen or so company vetted spreadsheets that most offices maintain for doing common things. You know: spread footings, openings in steel beams, column load take-down, yada yada. It is my strongly held opinion that, where such internally vetted tools are available, they should be used with very few exceptions. In my experience, at least 80% percent of EIT, one off spreadsheet will contain an error. Seriously.

Kootk - Do I get your #2 point correctly as that the software which allows easier combinations of calculations with images support actual understanding, and therefore help prevent the large mistakes? That sounds reasonable, considering how useful it is to diagram next to hand calculations.

That's precisely what I was getting at. Doing a meaningful review on a endless, streaming effluent of numbers is painful and time consuming at best. The accompanying sketches are the most valuable part I feel. How did you decide you're trib area? Unfortunately, because hand sketches take time and force folks to go non-digital, they seem to get used too sparingly. Clipping and pasting sketches into some kind of digital scratch bad seems to produce excellent results. As an example, a rockstar EIT that I used to work with would give me calcs like this:

- Load casing done in matrix form so that you could easily see what when into determining max/min without having to wade through the results of every combo.
- Pasted snippets of the code provisions where it wasn't completely obvious what code provisions had been drawn upon. Folks mostly have access to digital codes now so it's easy.
- Snippets of the framing plans with hatching showing the trib areas for beams. Same for columns with snips of the architectural sections to show unbraced column heights.
- Snippets of SAP2000 results. Moment diagrams, deflections etc.

It's nice, as you can imagine. In this day and age, it's all to often the case that "checking" means getting directed to some folder full of models and half baked "spreadsheets". Easier just to redo the design myself. And I'll do that for some spot checking at the end. Mid-project though, I need to be able to see that things are being done correctly in order to keep the project on track.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[dik]

Haynewp: That's pretty much the same that I do...

Dik

 

[WARose]

[blue](spaderess)[/blue]

WARose - Thank you for the tip. I bought it at the spot, looks like a sound investment.

The version I've got is from 1991. I think there is a later one. But I love this one because all the solutions are hand calcs. It's like looking at a good set of calcs from a older engineer.

 

[MonsieurR]

For normal stuff the reviewers accept spreadsheets and/or mathcad files, not really very tidy and clean.

For fancy stuff I use LaTeX. For a faster typing I use mathcad and then use a mathcad to LaTeX parser I found on the internet a while ago. This helps make calculations faster. And the document is pretty.

When I work with people unfamiliar with LaTeX, then Word is the way to go, but it is a pain in the ****. When I learned LaTeX I forgot everything about Word.

About the completeness of the calculations, when doing obscure works (e.g., complex shells) where I work it's standard to make the calculations a little bit less legible such that only the reviewers (people who know the calculations) are able to understand it, so people unfamiliar with the subject feel a bit lost. This typically involves removing steps and using obscure symbols. For handwritten calculations just copy your doctor's handwriting.

 

[hokie66]

MonsieurR,
If your complex shell collapses, your undecipherable calculations won't help you.

 

[MonsieurR]

I disagree, they are not undecipherable, just harder to understand. BTW I didn't come up with the idea!

 

[NorthCivil]

My first job, the supervising engineers were flexible on format, but my calculations were gone through with a fine tooth comb. I started using mathcad - after some practice, it was just as fast (and later on faster) as doing numbers by hand. Huge bonus is that human calculation errors were eliminated. Going between metric and imperial, a huge time waster, was automated (Canadian here).

In the process I ended up reusing huge chunks of spreadsheets for similar projects. By the end of my tenure there I had automated 75% of my day. I could show up and work 2 hours a day, spend another 2 developing increasingly complex mathcad sheets, and read/study/relax the other 4. The supervising engineers were hesitant to give me any more work because I was already pumping out so much.

Refusing to use mathcad because it takes too much time is akin to refusing an offer of 100$, to be given next month, to receive $10 today. You will save time, and the benefits will be realized very soon.

 

[hokie66]

Monsieur,
Whatever...it is a poor idea to intentionally obfuscate.

 

[EDub24]

I personally try to tell a story. First I state the background info. Then I move into the design standards, assumptions and specified data. Then I go through my calcs following the load path for tge most part. For example, for gravity loads I would start from the roof and work my way down to the foundation. For lateral loads I’ll start with my diaphragm loads and go down from there.

 

[bootlegend]

I started using mathcad - after some practice, it was just as fast (and later on faster) as doing numbers by hand.

I am trying to get to this point with Mathcad and have made a few sheets for repetitive stuff that I do. The biggest aggravation for me is that I do a lot of one-off custom calculations that rely on hand sketches and free body diagrams to explain. I'd like a quick and easy way to import a hand sketch to the Mathcad sheet alongside the calculation. It seems by the time I sketch it, scan it, import it, format it after it throws off the sheet layout that it's quicker to just do the calculations by hand.

 

[tempeng]

I'm another MathCAD convert. I like the easy integration between math and text which can make it an easier task on the reviewer by explaining what you're analysing, why and the method you're using. It's easy to paste screenshots from AutoCAD too if it starts getting a bit "wordy". Having the math presented very similar to hand calcs makes the sums easy to follow too as all steps to the solutions are visible. Don't get me wrong, this wasn't an overnight process to jump into but well worth it in the end.

 

[IRstuff]

@bootlegend -- one option is to get a digitizer, ala touchscreen or pad, a stylus, and Windows Ink or equivalent. There are many approaches to freehand drawing directly on a computer these days.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[rn14]

All those that like or are considering Mathcad should take a look at SMath Studio. I like Mathcad and all but at $1000 a seat I feel it is way to expensive considering you are basically buying a program language. They are also awful about breaking old sheets with new versions.

Those of you using LaTex and/or Word may want to look at using python along with Jupyter notebooks or SageMath.

 

[KootK]

I don't function well in a highly structured environment... I'm happiest when I'm working independently...

I hear 'ya on that. I've always thought it a bit cruel how things work out for our ilk. Being good at doing production requires a degree of attention to detail bordering on autism. Being good at managing the production requires one to be a "more than one way to skin a cat", 80/20 kind of guy/gal. It's rare that both skill sets come naturally to the same engineer.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[OldBldgGuy]

Okay, I'm response #39 & you guys (as usual) are mostly way above my head. I'm going to have to check out Mathcad, because I generally do hand calcs and my notes include code or design book references noted in the margins because even though I usually know where they are it saves time later, especially if I have to rewrite them for someone else. And that is always my big issue, occurring much more frequently these days: when asked to submit calcs, I have to rewrite them to make them presentable. I really can't conceive of getting out of that without using more design programmes (I do have some, I'm just not that comfortable with their invisible parts) but on the other hand I find that anyone asking for calcs is automatically relieved and immediately accepts something if it's computer generated. It seems to me that it's usually a corporate (read bureaucratic) response to something they don't understand to ask for calculations, & to be immediately reassured if it looks like the responder uses a computer.