Checking Other Engineers' Calculations 4

[jeg1976]

My previous employer was very lax about formal checking of calculations (specialized industry), and now my current employer has an extremely formal checking process. I am not very comfortable and take longer than I should with checking other engineers' calculations due to rarely if ever been required to do so. Does anyone have any advice or maybe experienced the same thing? I am in the civil (structural specifically) engineering field.

 

[cranky108]

It's not a bad idea, however if you use a spreadsheet then the calculations should be correct. But then you will need to check the formulas.

The more you automate the calculations, the less you will need to check the calculations. But you will need to review the data going in to the machine.

 

[JedClampett]

I was in just the opposite situation. Where I used to work, Nuclear Design, all safety related building calculations were checked, number by number. Where I am now, we check the end product, the drawings.
If you feel you're taking too long, don't worry, they'll tell you. It's tedious and inefficient to review others calculations. You have to get into their head, follow a procedure you might or might not of chose. What I found was that normally the mistakes I found were not significant. But every once in a while....
I feel that time can be better spent by looking at drawings rather than reviewing calculations. But if they're paying you to review calculations, that's what they want.

 

[MintJulep]

Remember to check the calculations, not the math.

It's too easy to fall into the trap of checking that the right arithmetic was performed on the wrong equations, or that the wrong number was correctly carried throughout the entire process.

Often it's better to do your own completely independent calculation first. Then if you come up with a different answer start looking for why.

 

[zelgar]

ONe thing I find when reviewing other people's work is how well they document their calculations. For example, are they providing the formulae before they're doing the calculations. Providing the values AND THEIR UNITS in their calculations. I don't know how many times I've had to search through countless pages of calculations & drawings just to figure out what another engineer was trying to calculate (or where the value was coming from) just because they couldn't take a little bit of time of clearly documenting their work. The lessons you learned in college of showing your work applies in real life too.

 

[ash060]

I agree with Jed. Looking at the final product is the first place to start when reviewing calculations, if something doesn't look right on the drawings than that is probably a good place to start checking the calculations. With spreadsheets I usually have to go through them pretty close because you can't see all the numbers.

 

[SectionIX]

Agree with the above posts...been there..done that. In general, I have found that the calculations are generally right, but the wrong formula or variables have been used. My pet peeve is the inconsistant use of significant figures.

 

[Sparweb]

This is where I built up my own habits of annotating and putting comments into most of the analysis work I do, whether I expect anyone to review it or not. It's most likely going to be me reviewing it years later, when a revision or a change comes up. It's pretty embarrassing to get lost and confused in your OWN work. Writing down the assumptions (with a reason), write the equation in algebra (even in Excel!), note the units of each number, positive direction of the x, y, and z axes, and so on. These things take a bit of time each, but also save time to everyone who reads it after you. Write everything so that it can be read by someone else. These habits make it easier for you to catch your own mistakes, before anyone else can find them!

When reviewing someone else's work, I have a couple of coloured pencils, sticky notes, a pad of paper, whatever is appropriate for the amount of work this will entail. More preparation is needed for a long report than is for a simple one-page change order, of course. Each of these is associated with drawings, references and other documents, so have these handy before you start. Talking to the drafter/writer/analyst helps a lot. That sounds obvious, but I've caught myself puzzling over things that I should have asked the person explain, before I started looking at it.

STF

 

[sirmick]

Checking is good. How expensive is a mistake vs. how expensive is double/triple checking? One sticking point can be ego.