API 650 frangible roofs/Railings at roof/tank size

[kalluna]

Hi,

Looking for some input or suggestions as to how to handle fall restraint/railing at vertical tanks.

As far as I understand from API 650, one of the conditions that must be met before having a frangile roof is that diameter of roof must be 15.25 m or more in diameter.

Most of our vertical tanks are smaller than that. However we have traditionally built (all) our tanks with a railing only on a portion of the roof edge,as someone way back when thought that would enable the roof/shell joint to be frangible.

Since we are now developing a proper fall restraint policy, I am proposing that we put railing all the way around the top of tanks (smaller than the 15 m dia), with a lower rung placed so one cannot slide off even if you fall down onto tank roof.

Using fall restraint harnesses while on tank (tied off to central painters post) is fine if there are 2 people, but if, (despite policy), one person goes up by themselves and they fall they could be hanging in harness for along time and possibly die (these task are in isolated locations).

For larger tanks which can have frangible roof we'd probably need to use and enforce the fall restraint/harness/2 person rule, and I guess there we would not need railing round full perimeter..

Any comments would be welcome.

Thanks!

Heather

 

[JStephen]

I would not assume that a standard handrail kept the roof from being frangible.

 

[kalluna]

Well neither would I -but am no welding expert.
I suspect that the decision not to put railing all round tanks edge was actually made
a) to save money on railing, and
b) due to a (misguided) notion that "if we don't put railing there then people won't go to that part of tank roof and therefore won't fall off"

 

[vesselguy]

Welding hand rails around the edge of the shell at the edge of the cone roof does not help make it frangible. That's a new one on me.

Read API 650 for all the requirements that must be qualified for a roof to be frangible. I have to assume you know what is a frangible roof and how it works. If not, state so and others will explain it to you here.

As for handrails around the top of the roof, do you want to walk up to the edge of a sloping cone roof tank and walk around there without hand rails, even if you have fall restraint? If your answer is no then you might want to put some handrails on.

 

[vesselguy]

Oh yes, and you're right to only want to put rails in the segment of the roof where you typically want access to. That's common practice and it saves money.

 

[JStephen]

Although that gets into potential OSHA problems & may require fall restraint to access the roof, even if you stay near the handrail.

 

[kalluna]

Sorry if post was not clear.
I understand what a frangible roof is and what the API conditions are for it-but at our organization there were those who believed that if we had a handrail all around a tank roof edge it would stop the shell-roof weld from being frangible. I don't think that is true but please correct me if so-I'm not a welder.
We are concerned about fall restraint but seeing the requirements needed to use harnesses etc on tanks if no handrail, it would probably be simpler and easier to just run the handrail all along the edge of tank. (Most of our tansk are below the size for frangibility at roof weld anyway.)

 

[Motorspirit]

This is not the first time Ive heard this suggestion.

Firstly, as you rightly point out, for smaller diameter tanks it will be difficult to practically get your tank to meet the code requirements for frangiblity. However I always try to meet the principles where I can as ultimately, regardless of tank size, you want the roof to fail before the floor to shell joint.

For larger tanks that meet frangibility requirements, it could be argued that a handrail could stiffen the kerb angle area and affect the buckling capacity of the roof to shell joint. Therefore frangibility could be affected. However I doubt this affect would be significant, assuming you are using conventional handrails that are not overly rigid.

It would be difficult to prove other than in a physical test as the mechnanism is reasonably complex so good engineering judgment must apply to this situation.