In Situ Guard Rail Testing (I summon Ron telepathically) 2

[KootK]

I have the opportunity to be involved in a substantial amount of guard rail testing on some mid-rise, Canadian, PT buildings. I've not done this before but, as a budding entrepreneur, I don't consider myself to have the luxury of sticking to things that I already know how to do. For what it's worth, I'm in the proposal stage at the moment. I also have a couple of good resources (ASTM E935-13 an CSA A500-16). I've scanned them in search of my answers but I'm afraid that I don't have time this week to do any deep reading. I'd be mighty grateful for any help that anyone can offer with respect to the following questions:

1) When looking at a bunch of guardrails, what percentage needs to be tested in order to establish statistical significance? A500 seems to say "depends on how many" but then seems to imply two or three per installation configuration.

2) Where does one get the equipment for the test and how much does is cost to buy or rent?

3) What's a reasonable time allotment for the execution of a single guard test in the field?

4) Obviously, any other relevant standards or reference material would be appreciated.

In addition to the standards that I mentioned, I've got this which is decent: (Link). I know there's at least on good thread out there with Ron contributing but I'm at a loss to track it down so far.

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.

 

[JoelTXCive]

The testing rig above is impressive. I have also seen other postings on Eng-Tips discussing testing of guard rails.

I think my questions are loosely related to Koot's above...

What I have always wondered is: Are these tests 100% non-destructive?

Don't you get some concrete cracking, or minor plastic deformation of the anchors, or anchor plate bending during the test?

As the years go by....If you test the same rail multiple times; then could you theoretically take a good rail, and make it unsafe just due to the testing?

 

[CANPRO]

A couple notes:

The hydraulic equipment is pretty cheap to rent. There are a few places to rent around here (I'm in a medium size market) and you can usually get the pump and the jack for around $100 or less for a day or two. If this is going to be something you do regularly I'd suggest buying the equipment (probably looking in the $1000-$1500 range depending on what you need).

Are the units currently occupied? If they are, you should have a plan in place in case a rail fails. You might want to put that on the owner - maybe mention in the proposal that they're responsible for restricting access to failed rails. Depending on the style of guardrail, a failure could mean that its going to fall over the edge - unlikely for most rails, but worth taking a look at.

I've done quite a bit of on site testing over the years, not specifically for guardrails, but a variety of other anchors. For estimating the time (this is especially true if the units are occupied) consider that something will almost always restrict your access for a couple of locations. There is a pile of building materials over an anchor, a tenant won't let you through their unit, one of the anchors isn't installed yet, etc... this list of reasons I've had to spend an extra day on site because of BS that got in my way is almost endless. I strongly suggest making it clear you need free and continuous access to the test areas or there will be additional charges.

A couple other tips and tricks: You should have something on site to measure deflections before, during, and after the test. Bring spare parts for your test equipment. Bring whatever tools you might need to change these spare parts on site. Don't go too cheap on your fees - the big name testing companies are almost surely going to be charging much more. These jobs can be very profitable if well planned with no major issues.

 

[StrucDesignPE]

Is this the thread you are looking for?

https://www.eng-tips.com/viewthread.cfm?qid=426182

 

[KootK]

Is this the thread you are looking for?

Indeed it is, thank you.

Thanks to you too CANPRO, some sage advice there.

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.

 

[Agent666]

The number to test can depend on the coefficient of variation of your population, the problem is you don't know this unless you are testing to destruction (because you never know how close you are to failure just testing to a static load.

Our loading code (AS/NZS1170.0) has the following table for 'prototype testing'. This is as the name suggests intended for establishing the validity of the design, usually in pre-production though. The K_t factor is essentially a load factor. The alternative way to read this in reverse is if you have a loadfactor of 1.5 say, you test to 1.5 x working load and if you were testing 10+ handrails you could handle a coefficient of variation of 20% which may or may not seem unreasonable. This might give you an indication on numbers, sometimes you start the tesitng with a number in mind but details take 5 times more load than you originally considered so you might deem detail adequate much earlier in the testing regime and start testing every second installation or something. You might specify an initial number and depending on results require further testing, offering a staged fee might might also make it more palitable to those paying the bills. Depends if you are going into it knowing they will not be adequate or if its expected that they will pass with flying colours.

Usually for 'prototypes' you might test to destruction and continue to test enough units to establish the coefficient of variation at which your test samples achieve the given k_t load factor on working loads and this proves the design is adequate. Once you achieve this benchmark you don't need to test any further units.

It's important to note that you obviously don't have the luxury of testing to destruction (I'm assuming?!). Our loading code also has a procedure for 'proof testing' also, this is intended to be applied to items that might have been built and need to achieve a specified ultimate strength and/or servicaeability requirement.
This alternative procedure simply states the item must be tested up to the ultimate factored load and there are acceptance criteria for ultimate and serviceability limit states as per following screen capture. Simply put it must be capable of resisting the design load as you would expect. It can fail 1% over this if you like but you've done your job as the designer.

It's quite important to agree acceptance criteria & testing methodology beforehand (how long it might need to maintain load without further deformation, and the cutoff for acceptance, etc), as one you have the results it can alter your view (i.e. lowering acceptance criteria based on all details only achieving 90% of ultimate load for example).

Hopefully the above is of some help in answering your first point at least.

 

[racookpe1978]

Bookmarked.

One comment, perhaps more later.
Do NOT assume you will have a "perfect and flat" platform (sidewalk, porch, concrete pad, concrete floor, deck, asphalt pavement, etc. The posts could be in a curb, no overhead to brace against, on top of a wall or kneeboard or sitting bench. You may need to brace back against a structural wall (if a porch, that could be a ways back), or if a balcony - you might have all three (back wall, flat solid floor, and an overhead. Think of three (or four) different ways to anchor your test device.

What is "acceptable" permanent and temporary deformation, if the 50 lbf/foot criteria is met, but deflection does occur?

What is limit if a railing has 2 posts, but the posts are 6 or possibly even 7 or 8 feet apart? (Comparable Canadian rules for metric). We only "average" standard posts 4 feet apart, many other configurations are possible.

How will you certify (calibrate/re-calibrate) your force gage? (The ram will give you reliable and repeatable "psi" readings, and you can get a cal sticker for the psig gage, but that by itself does not assure anybody what the "force" is on the end of the ram - particularly if the object is deflecting (motion distance of the ram) does not equal final force.