Fall Arrest Load Cases 1

[JohnnnyBoy]

I am looking into a fall arrest anchor design and I could not find any documents speaking to load cases for fall arrest. Lets say under 1.0DL + 1.5 LL + 0.4 WL a roof truss is at 95% utilization. Given the additional 5000lbs arrest load added to the load case the truss would be overstressed and fail. Now the chances of a full live load, dead load and partial wind load being perfectly timed with a fall is extremely unlikely. Does anyone have any documents or experience with what load cases should be used for an existing building?

 

[EZBuilding]

One of my old bosses argued the use of ASCE 7-16 2.5 load combinations for extraordinary events for a similar application. I have not given it much thought since, but it could be a starting point for you.

 

[phamENG]

Now the chances of a full live load, dead load and partial wind load being perfectly timed with a fall is extremely unlikely.

I disagree. Think about it: on a roof, the most likely time for people to be up there is when you're doing a bunch of work, like reroofing. So you have some workers and a big pile of material (your dead load and live load). A storm picks up and the workers rush to wrap up what they're doing and secure their materials before it hits. So you have people in a hurry, with wind picking up. Can you think of a more likely time for somebody to fall off the roof?

That said, I'd look a little closer at the requirements. I don't practice in Canada, so maybe it's different, but in the US the 5,000lb load is really more for the connections and components. Each fall arrest component needs to be tested to 5000lbs. When it comes to the actual structural response, we need to be engineers and do a more detailed analysis. Let's face it, if the fall arrest system is only holding one person and there's a 5,000lb load on it, that person is now going to be in at least 3 pieces (only 2 if it's a belt harness). So you have the required test loads to ensure reliability of components, and you have the engineered load path. Now if you can make the engineered load path work for 5,000lbs - that's fantastic and you should do it. But sometimes it doesn't, and you need to be a little less conservative and a little more realistic. Of course the less conservative you are the more accurate your analysis needs to be - which can be difficult to do. (I know there will be a firestorm of decent on this - but OSHA does provide a caveat for engineered design carried out by a competent individual in lieu of the prescriptive 5,000lbs requirement.)

 

[Aesur]

Are we talking roof or floor live load? I'm no longer familiar with Canadian load combinations as I haven't touched them in 5+ years, however is there a distinction between live load and roof live loading that could change the load combination? Assuming the fall protection is being used for construction purposes, maybe you don't have full live load, but a construction live loading? I believe there may be provisions for construction live loading versus final live loading.

 

[MotorCity]

You will need separate load cases for roof live load and live load. I would consider the 20 psf uniform load as roof live load and the fall arrest load as a live load since it is more of a localized load as compared to roof live load which is more of a global live load. If your truss does not pass, I would definitely take advantage of live load reduction if your code allows it.

 

[Once20036]

IBC `18 section 1607.10.4 defines this as a live load, and it would need to be applied in the load combinations like every other live load (here in the states).

 

[XR250]

Wouldn't this be considered an impact load?

 

[JLNJ]

I vote for the extraordinary event load case.

 

[le99]

For building structures, I suggest the fall arrest load should be treated as an extreme case of live load with a load factor of 1.0 (no add'l safety factor). For the connection of the fall arrest system and the building element, the load shall be an individual load case with OSHA specified safety factors.

 

[human909]

It is funny how often this pops up on this forum.

The simplest, clearest answer given by le99 IMO.

For building structures, I suggest the fall arrest load should be treated as an extreme case of live load with a load factor of 1.0 (no add'l safety factor). For the connection of the fall arrest system and the building element, the load shall be an individual load case with OSHA specified safety factors.

IBC `18 section 1607.10.4 defines this as a live load, and it would need to be applied in the load combinations like every other live load (here in the states).

While this might be true it is not in line with the limit state design approaches. The 5000lb figure is not a limit state figure it is a prescribed figure that already has 'safety factors' built into it. Combining it with live load and wind load makes zero sends in my opinion. That said the 1.5 LL + 0.4 WL is a combination that doesn't exist in my code and besides most of the time 1.0WL would still be greater.

Personally I'd make sure that all connections are suitable for the load an and done with it. Also remember that this is a peak and extremely brief load, using tensile strength rather than yield strength would be appropriate in my book.

All that said even with those cavets I still found myself designing crazy strong mountings for this device:

https://www.3m.com/3M/en_US/p/d/b10512888/

 

[skeletron]

This article deals more with load testing than design, but perhaps it helps:

https://www.structuremag.org/?p=21129

There is very little guidance in the CSA Z standards which deal with FA explicitly, but not explicitly enough because there are still some misunderstandings and voodoo magic out there.

My take for the design of the truss/joists: use the extraordinary load event of ASCE 7. Intuitively, you'll have 100% of the dead load and 50-100% of the live load on the member when the 5000 lbs. (ultimate load). You'll probably always need reinforcement on the section.

 

[Once20036]

For building structures, I suggest the fall arrest load should be treated as an extreme case of live load with a load factor of 1.0 (no add'l safety factor). For the connection of the fall arrest system and the building element, the load shall be an individual load case with OSHA specified safety factors.

I agree logically that this is an extreme load case and intuitively this argument rings true, however, if IBC wanted it handled this way, they`d have defined it this way. They say that it's a live load, and it should be handled in accordance with requirements for every other live load. There was recently a thread about ignoring results that you don't like, and nearly everybody said "you`ve got to follow the code", even if you disagree with them. Fascinating that a week or two later there's a different take on the topic.

While this might be true it is not in line with the limit state design approaches. The 5000lb figure is not a limit state figure it is a prescribed figure that already has 'safety factors' built into it.

Agreed about he safety factors, I think that the 5000# force is a 300# person, falling 6', with a 6" stop distance, and FS=2.0 (or something like that). IBC defines this as a ~3100# live load, that gets a 1.6 LRFD factor up to ~5000#

 

[human909]

Agreed about he safety factors, I think that the 5000# force is a 300# person, falling 6', with a 6" stop distance, and FS=2.0 (or something like that). IBC defines this as a ~3100# live load, that gets a 1.6 LRFD factor up to ~5000#

I think it is a nice round figure that somebody or organisation once chose and ran with it. Everything that has come afterwards has been making the calculations suit the 5000lb figure rather than the other way around. Somewhere along the way somebody converted the same number to kN, rounded it and ran with it using the sensible measurement system.

For context it is pretty easy to get loads between 5-10x body weight during a decent factor 2 fall. (A factor 2 fall means falling twice the length of the restraint.) Suitable energy absorbing devices should limit that to 5x body weight. To get over 10x body weigh you really need something weird going on with much longer lanyard and no energy absorbing devices. At those sorts of levels the user becomes the energy absorbing device and it would get quite painful.

For what it is worth I've taken falls that have had loads in excess of 5x my body weight.

 

[phamENG]

A few points to consider:

Fall arrest and lifeline anchorages.
In addition to any other applicable live loads, fall arrest and lifeline anchorages and structural elements that support these anchorages shall be esigned for a live load of not less than 3,100 pounds (13.8 kN) for each attached lifeline, in every direction that a fall arrest load can be applied.

Self-retracting lifelines and lanyards that automatically limit free fall distance to 2 feet (0.61 m) or less must have components capable of sustaining a minimum tensile load of 3,000 pounds (13.3 kN) applied to the device with the lifeline or lanyard in the fully extended position.

D-rings, snaphooks, and carabiners must be capable of sustaining a minimum tensile load of 5,000 pounds (22.2 kN).

D-rings, snaphooks, and carabiners must be proof tested to a minimum tensile load of 3,600 pounds (16 kN) without cracking, breaking, or incurring permanent deformation. The gate strength of snaphooks and carabiners must be capable of withstanding a minimum load of 3,600 pounds (16 kN) without the gate separating from the nose of the snaphook or carabiner body by more than 0.125 inches (3.175 mm).

Anchorages, except window cleaners' belt anchors covered by paragraph (e) of this section, must be:
1910.140(c)(13)(i)
Capable of supporting at least 5,000 pounds (22.2 kN) for each employee attached; or
1910.140(c)(13)(ii)
Designed, installed, and used, under the supervision of qualified person, as part of a complete personal fall protection system that maintains a safety factor of at least two.

Personal fall arrest systems are rigged in such a manner that the employee cannot free fall more than 6 feet (1.8 m) or contact a lower level. A free fall may be more than 6 feet (1.8 m) provided the employer can demonstrate the manufacturer designed the system to allow a free fall of more than 6 feet and tested the system to ensure a maximum arresting force of 1,800 pounds (8 kN) is not exceeded.

So the only source for the 5,000 lb load is OSHA. And OSHA does not define whether this is meant to be an ultimate load or a service load. It only says "capable of supporting." To me, this sounds like an upper limit. It doesn't need to be "capable of supporting 5,000lbs with a safety factor of at least two"...it just says capable of 5,000lbs. So if you're using 5,000lbs, it should be an ultimate load. If it fails at 5,001lbs...that's okay. Because it was capable of supporting 5,000lbs. It's also worth noting that this is only the anchorage - you can conservatively argue that this load should be carried through the entire load path, but I don't think it needs to be. Just like impact vibrations aren't chased further than the girder supporting the impacted joist and blast load anchorage forces aren't typically chased through the whole load path, there's no need to do it here. Make sure your anchor is attached with a connection that meets the 5,000lb (ultimate) load and then follow a more reasonable load through the rest of the load path.

That more reasonable load comes from the IBC (and also meets the requirements of OSHA for design by a qualified person). If you look at the maximum allowable arresting force - 1,800lbs - that's the peak load experienced by a worker falling 6 feet or less. If there is tension greater than that, your system has already "failed" (1,800lbs is 5.8x the weight of a combined weight of worker and tools of 310lbs). So IBC says you have to design for at least 2x the maximum tensile force in the system: 3,600lbs. That's effectively a "service" load. Of course, this actually gives you a safety factor greater than two as it doesn't consider material safety factors used for development of allowable stresses or the combined effect of load and resistance factors in arriving at an effective safety factor. But it's quick and easy and conservative.

So, for a fall arrest anchor that a worker is tied to directly, I say design the connection for 5,000lbs (ultimate) to satisfy OSHA, and then design the structural response for at least 3,600lbs (service) to satisfy the IBC and OSHA. Increase if designed for multiple workers.

If you are looking at a horizontal lifeline, it's a totally different story - those loads can get extreme fast, and you really have to consider energy absorption devices and their ability to cut peak loads to the anchorage to have a change at making it work. I've considered energy absorption and deflection of the lifeline itself before to keep the numbers manageable.

 

[Geoff13]

For a Canadian location I would review CSA Z259.16 "Design of Active Fall Protection Systems".

In particular Section 6.2 and Annex A.4 discuss factored loads and how the philosophy for load factors and combinations for fall arrest differs from that of NBCC.

I would also reconsider the 5,000 lb load you reference. It is acceptable to use 2 times MAF as your anchor load, which is typically 2 * 8 kN = 16 kN (3,600 lbs). The obvious exception is an HLL, but that's a whole other design discussion.

 

[XR250]

Aren't the anchors typically installed on top of the roof which would tend to pull the beam sideways or up? Also, unless this is attached to a bar joist, is an i-beam really going to fail from this? It will deflect a long way before that happens which will tend to reduce the forces.

 

[phamENG]

XR - sideways at an offset, so there's a moment on it, too. Could be bending moment or torsion, depending on the position and orientation of the framing. And yes, I agree with you. That's why I generally use the 5,000lb as the ultimate capacity of the connection, and (now that the IBC has a defined 'equivalent static load' for it) use 3600lbs through the load path.

 

[Once20036]

Anchorages, except window cleaners' belt anchors covered by paragraph (e) of this section, must be:
1910.140(c)(13)(i)
Capable of supporting at least 5,000 pounds (22.2 kN) for each employee attached; or
1910.140(c)(13)(ii)
Designed, installed, and used, under the supervision of qualified person, as part of a complete personal fall protection system that maintains a safety factor of at least two.

This is the part of the conversation that I`m not a fan of. It used to be 5,000# (ultimte) OROROR part of a complete system designed by a qualified person. By increasing the stop distance you could substantially decrease the forces, and design for way less than 5,000#(ultimate). Now IBC requires 3,100#(service)/5,000#(ultimate) without consideration for any of the fall protection specifics.

I say design the connection for 5,000lbs (ultimate) to satisfy OSHA, and then design the structural response for at least 3,600lbs (service) to satisfy the IBC and OSHA

So would you apply these loads as extreme events? Or concurrently with other live/wind/dead considerations?

 

[CrabbyT]

ASCE 7-16 defines a fall arrest load as a live load.

4.6.5 Fall Arrest and Lifeline Anchorages. Fall arrest and lifeline anchorages and structural elements that support these anchorages shall be designed for a live load of 3,100 lb (13.8 kN) for each attached lifeline in every direction that a fall arrest load may be applied.

When you multiply 3100 by 1.6, you get 4960 (which rounds up to 5000 because of sig figs).

 

[phamENG]

So would you apply these loads as extreme events? Or concurrently with other live/wind/dead considerations?

Concurrently with other live loads:

Fall arrest and lifeline anchorages.
In addition to any other applicable live loads, fall arrest and lifeline anchorages and structural elements that support these anchorages shall be esigned for a live load of not less than 3,100 pounds (13.8 kN) for each attached lifeline, in every direction that a fall arrest load can be applied.