Help wanted - EN 959 testing

[ChrisB]

my query is to assist with interpretation of the results. I'm looking for context of other anchors and anchor testing methods.

Ah, I understand now. Something to be aware of in that context is the difference between different types of testing and different levels of confidence in stated capacities.

A manufacturer will test to establish conformance with the standard they are certifying to. For anchors, that will typically include shear and pull-out. They will only be concerned with the capacity as established by the test set-up in the standard.

A specifier (in this context, such as BCA considering the suitability of DMM, PECO, IC, etc) will test for the conditions in which they intend to use the anchors, which might include static and dynamic loads, pull-out and shear, and different substrates (rock or concrete types). So their numbers might be different.

An installer will want to know that the anchors have been installed in a way that will achieve the performance expected by the specifier. To confirm the competence of an individual installer, the tests will typically be destructive. For anchors in cave, they will normally be non-destructive to anchors that are acceptable. It's not necessarily important to test every aspect, for example, if you do a pull-out test to verify that the hole has been cleaned and the resin has been injected and cured adequately, you can be confident that the anchor will also perform in shear.

Further up, 'characteristic load' was mentioned - apologies if you know all this, but that has a very specific meaning in the EN1990 series. It's effectively the typical failure load, so when using it in design of structures (bridges and buildings, etc) it should be multiplied by a factor to represent the variability of the load, typically 1.4 or 1.6, and divided by something like 0.87 to represent the variability in the materials (the 0.87 might be lower for typical cave rock).

Hope that helps in understanding what you find in different standards, or different parts of the same standard (eg, EN 959 section 4.5 vs 5.3)

 

[Bob Mehew]

I have tested them - in a fashion - my query is to assist with interpretation of the results. I'm looking for context of other anchors and anchor testing methods.

I understand this. EN959 and the BCA scheme use static (or at least slow pull) tests for their anchor testing which gives me confidence it is an widely accepted compromise in this case.

That being said, I have been considering some dynamic (drop) testing and how/where would be best to do this.

To go back to one of your early questions about how many anchors should one test. The 2018 standard only makes reference to "the anchor". I have found in my records that a proposal was made to clarify the need for a quality management system during the consultations prior to the issue of the EN 959 2018 version, but it was not accepted. I would put it to you the key question is how do you ensure that every anchor you make does meet the standard. Given several tests required by the standard are destructive, that means you can't test and then use each and every one. So you are looking at a quality management system. That is what we did when we looked at manufacturing our own anchors and advised Simon to do when he developed the IC anchor.

Interpreting results in part relies upon applying statistics. The simple approach is to just take the average of a number of measurements. That can fall down if there are two different failure modes as we (BCA E&T) found to our surprise when testing other anchors after DMM decided to cease manufacturing there P hanger / eco anchor. (Some failed with the metal / resin bond going, other failed by metal failure - not DMMs.) So we went to the more complicated method of checking the results fitted a normal distribution and then determine the 95% percentile value such that one had reasonable confidence that only 5% of the population of anchors would fail below a specified value. (That was to align with civil engineering stajndards.) If you want more detail then PM me and we can chat it through.

I would add that BCA do have an anchor puller which could possibly be loaned out. But please please do not try testing placed anchors as the head will distort at low pull out values as well as fracturing the resin. (That would probably trigger a demand to replace the anchor and replacing anchors is still not yet well developed.)

I did do some trials with a glue in anchor connected via a 1 m long lanyard (i.e. sewn loops) to a 100 kg mass and doing a FF 1.0 drop. Unfortunately I can't find the results. My memory was that the lanyard broke each time but the anchor head was distorted. It is also worth noting that I don't recall the peak dynamic force in the ten years plus of rope testing I have been involved with, getting near 20 kN (as that was the limit of the load cell). Typically it varied from 8 to perhaps 15 kN. And given that is using a steel mass, replacing the steel mass with a body would substantially reduce the peak force as much of the energy would go into distorting the body, not the rope.

There is a bit on dynamic loading in a Hiti document . Re theory of static v dynamic, one also needs to take into account the energy dissipation and the speed at which it does so. Fortunately, the speed is not supersonic as in an explosion, where the behaviour is different. I wonder if this is overlooked when we are only discussing the dynamics of a person being arrested from a fall by an anchor?

 

[Fjell]

Out of interest what is the expected elastic limit of P-bolts et al?

 

[Bob Mehew]

Out of interest what is the expected elastic limit of P-bolts et al?

Most of our (BCA E&T) work involved finding the pull out strength, so we did not stop when the head of an anchor started to show movement to see if it was a permanent deformation. (But frequently, we had already observed cracking of the resin where it coated the metal of the head of the anchor.) However in the early 2010s we did purchase several anchor testers designed to provide an axial pull up to 20 kN. This was done on the basis of EN 795 (Personal fall protection equipment - Anchor devices) which does include a requirement that the "installation should be verified appropriately, e.g. by calculation or testing". We soon had reports of deformed heads. My guess is deformation probably occurs above 5 kN. One person claimed to be able to place a force on a P anchor head at 12 kN without deformation but I never saw how he applied the force. My suspicion was that it was not in "pure" axial mode. BS 8539:2012 provides more information but I don't have the 2021 update to reliably quote from it.

Oh and scanning the EN 795 standard, Annex B comments in its introduction that "the static strength test is based on a minimum factor of safety of two".

 

[Fjell]

Falling on a rope is not that dynamic, the rate of change would be similar to something like an engine component I would think. So (for the metal) the static measurements should be fine for dynamic calcs. The rope is a different matter.

You have different criteria when doing things like blast modelling on steel structures, rather a specialist area in my experience. And you don’t get the interesting effects you see in armour design (people get killed by spalling, not actual penetration, UK tanks use a round that doesn’t generally penetrate but pancakes - charming as that sounds).

 

[andrewmcleod]

While playing around with a pull tester on the recent anchor installer workshop hosted by the CNCC, I think we got to ~10kN on a BP anchor where the small boulder the anchor had been placed in had cracked in half and you could see all the way down the length of the anchor...

I think I also got to ~10kN on a different BP anchor placed without any resin in (which I don't think I was able to fully hammer into the hole). I then removed the anchor from the rock by hand (by appropriate wiggling).

PS 10kN was the limit of the tester, not the point of failure.

 

[Bob Mehew]

While playing around with a pull tester on the recent anchor installer workshop hosted by the CNCC, I think we got to ~10kN on a BP anchor where the small boulder the anchor had been placed in had cracked in half and you could see all the way down the length of the anchor...

I think I also got to ~10kN on a different BP anchor placed without any resin in (which I don't think I was able to fully hammer into the hole). I then removed the anchor from the rock by hand (by appropriate wiggling).

One problem we (BCA E&T) found is weakening of the rock due to microfracturing from prior use of explosives. (Do not use Fairy Cave Quarry as a test site!) So I suspect your bolder was weak to start with. As an aside, Simon Wilson did most of his testing in natural exposures of limestone compared to E&T's work which was mostly done in quarries. I think that explains why his results were far less spread (maximum to minimum) and achieved a higher mean value.

I recall tales of installers using DMM anchors just placed in the drilled hole as a suspension point whilst they glued in other anchors. A significant feature of the strength of glued in anchors is the irregular surface of the drilled hole. That reinforces the chemical bond between the rock and resin by creating a mechanical bond, just as you found between the 'fat' legs of a BP anchor and the shape of the hole. Which is why you use a percussive drill rather than a diamond core drill to make the holes.

Presumably you did not notice of there was any permanent distortion of these anchors after being subject to around 10 kN?

 

[Chocolate fireguard]

There is a bit on dynamic loading in a Hiti document . Re theory of static v dynamic, one also needs to take into account the energy dissipation and the speed at which it does so. Fortunately, the speed is not supersonic as in an explosion, where the behaviour is different. I wonder if this is overlooked when we are only discussing the dynamics of a person being arrested from a fall by an anchor?

That Hilti reference was more interesting than I expected (although they ought to employ better proof readers).

A couple of quotes:

“breaking loads are of roughly the same magnitude during static loading and shockloading

tests.” – this for dynamic loading times of milliseconds (rather than 10s or 100s of milliseconds in the case of a mass on a rope).


“If only elastic deformations are allowed (no permanent deformations) after the shock

incident, the static resistances of the anchor are also suitable for shock. This leads often

to a non-economic anchor selection. To avoid this, different regulations allow

plastic deformations on condition that the anchors are replaced after the shock incident.”

Presumably an anchor in a cave would be replaced after it had been plastically deformed (by what???)



Regarding the rest of that paragraph, the energy involved in straining the rock is not more than a few tens of Joules when the force applied is 10kN (as opposed to perhaps several million Joules in explosions) and the speed involved is certainly not supersonic.

But it is sonic – the energy propagates through the material at the speed of sound, which is over 3km/s in concrete and probably similar in limestone. During the 100ms or so that the dynamic force pulse takes to rise from zero to 10kN these waves could traverse a typical limestone block many dozens of times.

So I completely fail to understand how it could be possible for all the energy to remain in the vicinity of the anchor and build up a stress pattern that differed in any way from the static one.



Anthropomorphising the situation, what we have is humans looking at the drop test and saying “😆Yeehaw, that was fast – take that, rock” and the rock saying “🥱Sorry pal, you’ll have to do better than that to take me by surprise”

 

[mikem]

Presumably auto-translated from German (or written by someone who doesn't have English as their first language)

 

[Fjell]

That Hilti reference was more interesting than I expected (although they ought to employ better proof readers).

A couple of quotes:

“breaking loads are of roughly the same magnitude during static loading and shockloading

tests.” – this for dynamic loading times of milliseconds (rather than 10s or 100s of milliseconds in the case of a mass on a rope).


“If only elastic deformations are allowed (no permanent deformations) after the shock

incident, the static resistances of the anchor are also suitable for shock. This leads often

to a non-economic anchor selection. To avoid this, different regulations allow

plastic deformations on condition that the anchors are replaced after the shock incident.”

Presumably an anchor in a cave would be replaced after it had been plastically deformed (by what???)



Regarding the rest of that paragraph, the energy involved in straining the rock is not more than a few tens of Joules when the force applied is 10kN (as opposed to perhaps several million Joules in explosions) and the speed involved is certainly not supersonic.

But it is sonic – the energy propagates through the material at the speed of sound, which is over 3km/s in concrete and probably similar in limestone. During the 100ms or so that the dynamic force pulse takes to rise from zero to 10kN these waves could traverse a typical limestone block many dozens of times.

So I completely fail to understand how it could be possible for all the energy to remain in the vicinity of the anchor and build up a stress pattern that differed in any way from the static one.



Anthropomorphising the situation, what we have is humans looking at the drop test and saying “😆Yeehaw, that was fast – take that, rock” and the rock saying “🥱Sorry pal, you’ll have to do better than that to take me by surprise”

All they are saying is you need to decide what you are designing for. In any sort of construction you would obviously be designing below the yield point. If you are designed for a theoretical emergency load, then you can go above the yield point but below the ultimate strength if you are OK with deformation or will replace it.

In the case of P bolts we operate within the elastic limit normally, but a shock load may exceed that resulting in deformation, but it will not break. If the shock load had to result in no deformation then you would need a much more heavy duty bolt. I can well imagine the quoted 16mm bolt to be such a thing in our case.

 

[mikem]

However, a less forgiving bolt will mean more of the force is transferred to the load (our falling caver)

 

[Bob Mehew]

In the case of P bolts we operate within the elastic limit normally, but a shock load may exceed that resulting in deformation, but it will not break.

Sorry but to be pedantic, I would suggest that the "will" should be a "should". Little is absolutely certain in this world but I would accept that the likelihood of its breaking under a shock load from a falling body (or two) seems remote. However we did get several anchors fail at loads well under 20 kN whilst testing them. And to reassure you, we did not not use any more anchors from those suppliers in caves. There is also the threat of chloride stress corrosion cracking weakening the anchor which we think we have taken adequate steps to avoid but one can not rule it out.

 

[mikem]

Generally it has been the rock that has failed rather than the bolt

 

[ian.p]

The thing to watch out for with concrete screws is that it’s very easy to put too much torque across the bolt heads during installation and this introduces stress fractures which can lead to failure. If the holes aren’t cleaned out well or the concrete screw has been used too many times then they can jam and the temptation is to jump on the spanner or get a bigger spanner. I have seen a number of bolt heads sheared off during installation because of this approach. It would be a good idea to find out the maximum safe torque installation value for the screw you intend to use and then test that you can get the screws into a limestone substrate reliably without exceeding that value.
The other thing I have seen happen a lot especially with worn screws is that the screw strips the thread it’s cut sometimes completely and the screw reaches the end of the hole and spins or partially and then the anchor can apear sound but the pull out strength is significantly diminished compared to a well installed bolt.

The main takeaway from the above is I suspect the reliability of installation will be less good than with an IC anchor. Concrete screws get used quite a bit as roped access anchors but they are always individually subjected to a pull test before use after installation.

 

[ian.p]

Worth also having a read of this: https://www.thebmc.co.uk/bolt-failures-on-north-wales-limestone#:~:text=A%20%22Thunderbolt%22%20climbing%20anchor%20in%20place%20on%20a%20route.&text=Following%20the%20catastrophic%20failure%20of,'Thunderbolt'%20anchors%20with%20caution.

 

[georgenorth]

The thing to watch out for with concrete screws is that it’s very easy to put too much torque across the bolt heads during installation and this introduces stress fractures which can lead to failure. If the holes aren’t cleaned out well or the concrete screw has been used too many times then they can jam and the temptation is to jump on the spanner or get a bigger spanner. I have seen a number of bolt heads sheared off during installation because of this approach. It would be a good idea to find out the maximum safe torque installation value for the screw you intend to use and then test that you can get the screws into a limestone substrate reliably without exceeding that value.
The other thing I have seen happen a lot especially with worn screws is that the screw strips the thread it’s cut sometimes completely and the screw reaches the end of the hole and spins or partially and then the anchor can apear sound but the pull out strength is significantly diminished compared to a well installed bolt.

The main takeaway from the above is I suspect the reliability of installation will be less good than with an IC anchor. Concrete screws get used quite a bit as roped access anchors but they are always individually subjected to a pull test before use after installation.

This is also my experience of concrete screws. In softer rock (e.g. flowstone in a caving context) it’s really very easy to strip the ‘thread’ from the hole.
The company I work for uses them a lot as fixings for various uses. If removable rope access anchors are required then shield type anchors seem far more foolproof (e.g. https://www.fischer.co.uk/en-gb/pro...fwb-wallbolt/internal-thread-wall-bolt-metric)