What does "engineering accuracy" mean to you?

It all depends upon the size type and material of the boat. However a "baseline" is the IACS minimum requirements for build tolerances:

 

Engineering accuracy by itself is totally meaningless.

There may be certain fields where accuracy is required, but mechanical engineering isn't one of them. You can calculate loads in a construction with nothing more than a slide rule as long as you keep track of the decimals. Once you have determined "accurately" how large or small a force is, you multiply that number by a safety factor, like a number between 2 and 5, depending on rules, regulations or your reputation.

In digital electronics things are often different. A circuit may not work as planned if a signal arrives a few nanoseconds late or early.

 

I have worked in a number of industries, for most land based structures safety of factor numbers are 1.5 to about 2.5, for geotechnical land slide prediction they also use 1.5 or up to 5 or 6 when the uncertainty is high. for light aircraft most structures use 1.5 SOF, but on commercial aircraft the structure is usually around 1.1, with only critical parts like landing gear or engine mounts being up to 1.5. Each of these are reflections on the amount of analysis involved, the reliability of the materials and assembly process used, and the probability of the size of the loads that will be encountered.

Interestingly, in Lloyds rules for pleasure yachts, they used similar numbers for inland waters and even coastal cruisers, but for ocean going deep water cruisers they will use loads that have to be 5 or 6 times what one would expect to encounter. these numbers are based on a long history of experience with ship survivability. This reflects the high degree of uncertainty for boats out in deep water.

So what exactly is "engineering accuracy"? I would say the least amount of accuracy that can reliably meet the need. It is usually a waste of money and effort to try and get a level of accuracy higher than necessary. Usually engineers are not engaged in theoretical investigations, but rather in designing something that will actually have to get built and function within its operating limits. Sometimes high degrees of accuracy are necessary, other times not so much.

 

I believe early WWII code-breaking was also handled in a similar fashion.
A large room of women worked on coded messages, and when one of them
believed they had discovered an actual word, they would relay it to their
co-workers after it was checked by supervisors. I can't remember where I
saw this, possibly in a movie about Turing and the Bletchley Park nerds.

 

Thanks for that.
It will be interesting to see how it lines up with some rowing shells that I
hope to have a look at soon. Their tolerances are much smaller, but irrelevant if
the beast sags when eight 100kg oxen sit in it.

 

I think you have to be in the UK to see it, but the BBC had a programme just last week about Bletchley

http://www.bbc.co.uk/programmes/b016ltm0

I always tell builders that 3mm (1/8th) is good enough accuracy for the boat as a whole. But I hope the gaps between joints is less than that!

I also tell them I cannot see both sides of a boat at the same time

Richard Woods of Woods Designs

www.sailingcatamarans.com

 

For a boat survey, many measurements can be to the nearest 6 inches. For example, the beam is 12'6". For a propeller shaft, it is good enough to measure to the nearest 1/8th. Rigging wire typically gets measured to the nearest 1/64th. A survey of structural damage of a composite laminate or plate may need more accurate measurements in critical areas. On the other hand, I used to loft and it was hard to explain to some customers, that one pixel in the computer screen may become an inch when it got expanded to sixty feet. They asked to loft to the exact points in the table of offsets even though I pointed out they produced an unfair line.

 

Id love to meet a rigger who can measure and cut a 25 meter rod to 1 64 th.

If your building micro chips, Id expect 1 64th to be a bit much.

 

I think gonzo meant the diameter

Anyway...

I'm not sure the "Engineering Accuracy" refers to the "as built condition"...My old saw for yard work was:

Calculate to 0.001"
Measure to 1/8"
Mark it with a piece of chalk
Cut it with a gas axe
Bend and beat to fit
Paint to match

It is kind of hard to expect 1/8" tolerance on a steel item that will change length 2-3" between summer and winter or sun and shade.

 

Reading the thread back I notice jehardiman's cfd comparison. Doing some cfd study's comparing accuracy settimgs think thats a good example. Made me think of local mesh fea settings as perhaps a even clearer sample. Offcourse locating the diagnostic area first and not writing a peppered bill for goofing round like this expert machine shop I visited recently. Allways do a second opinion first and dont have blind fate in experts, beter know whats going on, yes I had a wrong diagnose and really glad I found out. Perhaps even better is I dont need to much accurate enginering, I like it "rough"

btw, wasnt the us going metric¿

 

Since Leo asked for funny stories, My first boss told me one. He was involved in some early radome work for the USAF. They were ordering parts for a prototype. He placed an order for a parabolic reflector he had designed. The mfg asked what surface finish he wanted. He said "as smooth as possible"

He kept calling them to find out what was taking so long. Four months later, he got it. It had taken three days to build, and they had been polishing it round the clock ever since! Knowing what specs actually cost is the perview of job schedulers in a shop. Every engineer should have to sit next to a job guy for a year.

One of my odd jobs was to attempt to reassemble and run a lasagna production line that had been tossed in an Italian junkyard, then picked apart, then the remains were loaded into containers and sent to North Dakota, which has 60Hz 240V power instead of 50Hz etc. The noodles traveled down the machine on a wire screen belt. The belt was driven by gears and chains. The gears were 35mm thick armor plate about 3' in diamenter. The chain looked rather like tank track chain as well. All to move a few noodles at 1ft/min. Leftover tooling from a WWII tank factory was all we could figure. That's what they knew how to make, so that's what they made.

 

All automotive engineering in the US is done in metric.

 

Depends on who and what. In my 03 Dodge, all the power train, except the drain plugs, are metric and the body is SAE. Go figure...

 

The first Airbus A380's were built at factories across Europe, trouble was that when they tried to put them together, the wiring did not line up.

http://www.nytimes.com/2006/12/11/business/worldbusiness/11iht-airbus.3860198.html?pagewanted=all

 

Interesting. I may be wrong about Chrysler. I wonder if only the fasteners are SAE or if Chrysler does some engineering in metric and some in inches/lbs/etc. My guess is the engineerng is all metric.

I know from personal experience that all engineering at General Motors is done in metric and has been for several decades. A few fasteners or other parts such as wheel sizes are nominally inch based due for industry standards or compatibility with widely used parts.