A Question for Civil Engineers-Testing Steel Beams?

Many years ago, my uncle was a VP of a local supermarket chain. One of his responsibilities was supervisng the construction of new stores, and he told me an interesting story. Supposedly, once the steel was delivered from the fabricator shop, it had to be tested for strength (resistance to bending). This was usually done by a testing company; in this case, the store was under s deadline for construction. My uncle made sme inquiries, and wound up testing the steel at a nearby sand and gravel pit (the steel beams were tested by loading them with bags of sand).
Is this true? Was steel so variable (in quality) that it had to be tested (before being used in construction)?
This was in the early 1960’s , I believe.

Just asked my dad, the architect (he built buildings from the mid-60s and continues to consult on buildings).

Steel and concrete were and are tested. But, samples are selected randomly from each batch or pouring and sent off to labs for testing. You then get a report back certifying the steel or concrete and telling you its physical properties. You don’t do any testing on site. And self-testing would be worthless. Without the lab reports the architect or structural engineer cannot sign off on a build.

It’s possible your uncle got a load of beams that turned out to be discrepant somehow. Maybe the paperwork (material certs) got lost. But it doesn’t sound credible that he would be able to do the testing with sandbags. Also seems like he could have returned and replaced the beams in less time than it would have taken to do the testing.

It’s unlikely anything useful could be achieved with sandbags, unless you don’t mind using bent beams.

The elastic deflection of a steel beam is a matter of its dimensions. Various strengths of steel have more or less the same elastic stiffness, although this seems counter-intuitive.

Where steels differ is in their yield point, the load and deflection at which they no longer spring back. With steel beams, you cut samples and perform a load-elongation tensile test to detect the stress at which the deflection becomes non elastic. Then you certify the batch.

If you lose your certs, or lose track of your paperwork, it can be expensive. If you’ve kept track of the batch you can sacrifice a beam or two to get them re-certified, but if not, you have to sample every one. Which may or may not be possible, depending on whether they are useable after you’ve cut pieces out of them. I’ve known some very expensive items be scrapped because their certification was lost and the process of sampling for mechanical testing, and then repairing the sampled area and certifying the repair, was just too tortuous.

Not quite. For concrete, you will often do testing on site. They do something called a ‘slump test’, which I understand consists of taking a sample of the concrete mix, putting it in a foot-high cone, then removing the cone and measuring how far down the concrete will slump. My neighbor did this as part of his job during runway construction at the airport. They tested nearly every load of concrete, and he said he rejected a fair number of them.

I must have misunderstood what my dad was saying…thanks for the clarification.

More than that for highway construction (and I’d bet runway construction as well). A sample of each batch is poured into a cylindrical mold, about four inches in diameter and eight inches in diameter, and allowed to cure. The cylinder is then subjected to a compression test and if it fails – whoopsie, that section of roadway has to be torn out and redone. Our neighbor has been slowly lining the edges of his driveway with these cylinders half embedded vertically. The test, for some reason, involves a molten sulfur cap about an inch bigger in diameter being molded onto both ends of the cylinder. The neighbor gives them a few licks with a hammer to make them break up and fall off.

the reason is that concrete, unlike steel, is not a uniform substance, produced under controlled conditions in a factory.
Concrete is made of crushed rocks, sand, water, and cement powder that glues it all together. The mix in each truckload can be very different; the size of the gravel, or the amount of water added to the mixture radically changes the strength of the concrete.
So often a Slump test is done at the truck, before pouring the wet concrete, and a test cylinder is made at the site. The cylinder is sent to a lab where it is submitted to high pressure until it breaks, to verify the strength of the specific mix that was poured.

I disagree. I am not saying how it is done in practice but I disagree with the concept of what you are saying.

There are non destructive tests for metals but the basic concept of testing a beam by “testing” it (if you know what I mean) is perfectly good and is done all the time.

If you have a beam which is in practice suppossed to bear a stress of X and was calculated with a safety factor of K then you can test the beam by subjecting it to a stress of something less than K*X. If the beam withstands the stress without permanent deformation then you know it has passed your test. If the beam suffers permanent deformation then you have a bad beam which you do not use. Why would you worry about this? That is the whole point: discovering with the test that the beam does not meet its specs and not in real life when it collapses unde a truck.

Ropes (nylon, wire, etc.) are tested this way. I would rather end up with an unusable, broken, rope than to find out it was defective when it breaks and I fall from 5 stories high. If a rope designed to hold a person is designed to (say) hold 600 pounds with a safety factor of 10 then a test subjecting it to 5000 pounds would be useful.

In fact bridges are tested by placing loaded trucks on them all the time. In August 2003 seven bridges in Guangzhou were tested for safety. Nine freight cars carrying 30 tons of goods were driven across the Haizhu Bridge, which is 70 years old, to test its safety. The other six bridges tested are Renmin Bridge, Jiefang Bridge, Haiyin Bridge, Jiangwan Bridge, Guangzhou Bridge and Hedong Bridge.

Not only that, a beam is designed to have a certain deflection under certain loads. If you start loading it and the deflection is greater than expected then you also know you have a bad beam which will probably deform sooner than designed also.

There are also other tests which detect microcracks and other weaknesses.

In other words, beams are designed to meet certain specs and they can be tested to see that they actually do meet them. (Again, I am not saying this is routinely done at construction sites but that tests are possible and actually done in practice.)

Regarding concrete beam I would think pre-cast beams are also tested.

And a page and photos of the concrete slump test: http://en.wikipedia.org/wiki/Concrete_slump_test

That’s eight inches in height. :smack:

So, Calvin’s dad (They drive trucks over the bridge until it fails, then rebuild it using the last truck’s weight as the capacity) wasn’t so far off after all.

For its 50th anniversary, the Golden Gate was closed to traffic and hordes of pedestrians released. It apparently was the biggest load on the bridge ever – the deck’s slight upward arch was flattened; the vertical wires from the suspension cables to the deck were slack at the ends and singing in the middle. The engineers were kicking themselves for not putting strain gauges all over.

Absolutely! It depends what kind of certification we’re talking about.

In the UK, lifting gear such as hoists and slings are designed with a safe working load that’s six times smaller than the calculated failure load, and they are required to be regularly “proof tested” to twice safe working load. So from that point of view you’re quite right.

OTOH, many items are constructed from a material with specified properties, e.g. the American Petroleum Institute species pipe steels of grades X52, X65 etc. What is notable about these grades is that they have the same alloy composition, and their other properties (yield point, notch toughess etc.) are determined by the manufacturing process. If you lose track of which serial numbers corrospond to which grade, the only way to re-certify your pipes as X65 is to chop bits out and test them.

I take your point that you could in theory look at the design requirements of your pipeline or whatever and test to the safety factor and beyond, but that is not always convenient. Pressurising a pipe up to 200 bar at zero degrees C can be tricky!

But yes, regarding the OP you have a good point. I was thinking from a limited POV.