I made a beam

[SteamyTea]

1 hour ago, Gus Potter said:

The probablitlity bit is to do with how many beams you test

This weeks comic had a large article on looking for sub atomic particles.

CERN has just about reached its design maximum energy limits, and pretty close to the physical maximum anyway, so now they are looking at more sensitive detectors.

The sensitivity is in the number of good detection they get, rather than the accuracy and precision.  They look at millions of data points at a time, rather than millions of separate runs.

I seem to remember that in the old BSI standards there were recommended sample rates i.e. 1 in 20, 1 in 50, 1 in 1000, as the sample rate decreased the confidence in the results decreased.

The mathematics of it are pretty simple.

It is the interpretation of the results that is important.

 

So I could design and build a set of identical beams, some for my shed, and some for a new roof on my house.

They would both do the same job and be subjected to the same loadings.  But the ones on the shed, if they did fail, would have much lower consequences that the ones in a new roof.

So to be safe, SEs allow for this and do what @joe90 does, over engineer.

 

I was watching an extreme engineering Youtube yesterday, the MiG 15 had a loaded mass of ~5000 kg.

The LM F35 has a mass of ~22000 kg.

They both basically do the same job, from different eras, and like buildings, I am amazed how skinny and fragile old roofs look.

[Gus Potter]

20 minutes ago, SteamyTea said:

This weeks comic had a large article on looking for sub atomic particles.

CERN has just about reached its design maximum energy limits, and pretty close to the physical maximum anyway, so now they are looking at more sensitive detectors.

The sensitivity is in the number of good detection they get, rather than the accuracy and precision.  They look at millions of data points at a time, rather than millions of separate runs.

I seem to remember that in the old BSI standards there were recommended sample rates i.e. 1 in 20, 1 in 50, 1 in 1000, as the sample rate decreased the confidence in the results decreased.

The mathematics of it are pretty simple.

It is the interpretation of the results that is important.

 

So I could design and build a set of identical beams, some for my shed, and some for a new roof on my house.

They would both do the same job and be subjected to the same loadings.  But the ones on the shed, if they did fail, would have much lower consequences that the ones in a new roof.

So to be safe, SEs allow for this and do what @joe90 does, over engineer.

 

I was watching an extreme engineering Youtube yesterday, the MiG 15 had a loaded mass of ~5000 kg.

The LM F35 has a mass of ~22000 kg.

They both basically do the same job, from different eras, and like buildings, I am amazed how skinny and fragile old roofs look.

"The mathematics of it are pretty simple."

 

I have a graps of basic statistics and probablility but you have the edge on me here. That said I can follow your first paragraph and agree with the thrust of it.

 

"It is the interpretation of the results that is important." Absolutely.. it's like using finite element analysis.. manure in manure out, and all models must be verified but manual fag packet calculation.

 

"So I could design and build a set of identical beams, some for my shed, and some for a new roof on my house.

They would both do the same job and be subjected to the same loadings.  But the ones on the shed, if they did fail, would have much lower consequences that the ones in a new roof.

So to be safe, SEs allow for this and do what @joe90 does, over engineer."

 

No, SE's are not that blinkered and you mention the consequence of failure correctly.

 

Your shed could be treated as non habitable building. If could be treated as an Agricultural type building with low occupancy and say a 20 year design life. If so then we reduce the loading (quite significantly)  based on say a return period of heavy snow and a reduced roof access load for maintenance. Houses are mostly designed for predicted loading that could occur over 50 years, Agricultural buildings can be designed on a 20 year snow event for example. Also on an Agricultural building we have often slacker deflection limits as we don't have internal  brittle finishes that can crack, doors jamb and so on.

 

If you are going to be designing beams for your house then be careful.. and remember that if you sell someone may ask you to to prove all is ok!

 

If you experiment and gather data on your beams then you could have the same size of beam you have investigated but longer spanning for your shed as the loads are reduced and the deflection limits less onerous.

 

Keep us posted on your experiments..  it lets folk see on BH how these Engineered joists work, the plus and minus points and what you are doing kind of removes this "mystification" where folk are trying to cost up the things.

 

 

 

 

 

 

 

 

 

 

[SteamyTea]

17 minutes ago, Gus Potter said:

No, SE's are not that blinkered

 

17 minutes ago, Gus Potter said:

Your shed could be treated as non habitable building. If could be treated as an Agricultural type building with low occupancy and say a 20 year design life. If so then we reduce the loading (quite significantly)  based on say a return period of heavy snow and a reduced roof access load for maintenance. Houses are mostly designed for predicted loading that could occur over 50 years

That is more like what I wanted to say.

I trying (badly) to point out that the same product may have to be considered differently, in different situations.

 

Tomorrow I am driving up country and back, gives me a good 12 hours to ponder things.

 

[Gus Potter]

13 minutes ago, SteamyTea said:

Tomorrow I am driving up country and back, gives me a good 12 hours to ponder things.

I think what you are doing is great.

 

I love your enthusiasm for this and in general. You are testing my knowledge which is great for me, appreciate that.. a big thanks from me. As primarily an SE it is important for me to be able to distill sometimes complex things into language that folk on BH can find usefull.

 

Here are something other things to ponder on your drive.

 

You might want to explore using solid timber flanges top and bottom.. basically copying an engineered joist.. but they often use laminated timber. This will simplify the design. You router out a slot in the timber and fit the web into that. I think you will get more bang for your buck this way.

 

Next is your glue. Have a look at Cascamite glue which is a proper structural glue.

 

Lastly for now if you can get a hold of an engineered joist off cut and split the web from the flange you'll often see that the web is slightly tapered so it jambs into the routered slot in the flanges. You may get a fright as often the web departs from the flange easier than you think!

 

But your idea of using the timbers to connect the web to the flange also works.. but it is labour intensive and will need a a lot more thought to get something that is reliable and consistent in terms of how it is put together and has no weak spots.