NOZZLE LOADs - EFFECT ON ANCHOR BOLTS 2

[roca]

Good Day
Has anyone any experience with adding the effect of piping loads on vertical vessel / columns to the anchor bolt design?
I think it depends on the vessel / column design details and the way the piping is run.
On a short thick wall column (therefore stiff) with an overhead which does not run all the way down the column and goes off to the piperack at a certain elevation, then this could give you a problem due to expansion and the shear force produced which the column cannot absorb due to its stiff shell (no flexibility).
Therefore a bending moment is produced which has to be added to the wind / seismic moments at base.

I have personally never done this before and I am being told that some anchors bolts have been pulled or failed due to these additional loads / moments
Have anyone heard of this before?

These days it seems to me that equipment are used as anchors

Thanks

 

[JohnGP]

Vessel design, including anchor bolts, necessarily should include all possible loadings. Whether the vessel is stiff or not, any forces induced by attached piping will result in forces and moments at the supports. Vessel flexibility though will assist in limiting the size of local forces and moments experienced in the region of the nozzle, and in the attached piping.

So, yes, attached piping forces should be included in support design, usually seismic and wind are not fully additive together (unless you see that as a definite possibility).

John

 

[roca]

Hi John
Thanks for the reply
Wind and seismic are not usually applied together.
Do you have a method for determining which nozzle loads to take into account given that nozzles can be in various orientations and could either be acting together or against each other.
Regards

 

[prex]

You must adopt the worst case combination.
If your design nozzle loads are specified as plus or minus, you need to apply them according to the nozzle orientation with signs chosen in order to get them in the worst combination.
If this gives an unrealistic result, your only option is to revert to the piping designer for a closer analysis and specification of piping loads.

prex

http://www.xcalcs.com

Online tools for structural design

 

[JohnGP]

roca,
I agree with prex - if you have results of piping flexibility analysis, then use the reported forces and moments at vessel nozzles. Otherwise you will need to make some judgement on worst possible (realistic) combination of loadings from attached piping. (Adding up design nozzle loadings could make things kind of scary though).

Careful attention to arrangement of pipe supports can assist in limiting the size of loadings experienced by the vessel.

Hope that helps, it's a bit hard to be more definitive.

John

 

[bvi]

One other point, you want the anchor bolts to be pretensioned to greater than the expected loads during service, including wind, earthquake and nozzle loads. Otherwise, they could be subject to fatigue failure due to load cycling. As long as the preload exceeds the operating loads, the stress in the bolt will remain essentially constant.

 

[jte]

roca-

As JohnGP and Prex have stated, you have to consider all of the loads. Generally you'll find that piping or other external loads don't affect anchor bolt design. Unusually high loads, however, can. I've had to increase anchor bolt diameters at least once to accomodate the loads from an overhead line.

cb4- Thanks for the hot ductwork analysis y'all did for me recently. Maybe its because I'm out in earthquake country and not in hurricane territory that I've only seen one set of anchor bolts fail in fatigue. That was a case where I could literally lift the bolt out of the foundation - but it was due to too short of a skirt and thermal cycling. I was asked about bolt pretensioning when we recently installed a new vacuum column, and was told by the field guys that we can ask for whatever torque we want. They'll put an impact wrench on the nuts and stop when they decide that they're tight enough. I think that it would take quite a bit of engineering effort to change that attitude.

jt

 

[roca]

Thanks for all your replies.
I have spoken to some other people I know and I now feel that only realistic unusually high loads from a proper stress run calculation should be taken into account.
From a mechanical point of view bolts should not fail as we use quite low allowable stresses - (Shell typically use 80 N/mm2 in their DEP's). Pull out from the concrete is a sperate issue and this could happen.
If you are designing a column from scratch with 20+ process nozzles this could become a nightmare to evaluate depending on nozzle orientations. There must be a rule of thumb method which dictates which nozzles to assess dependent on their size or nozzle to shell ratio.

 

[prex]

roca,
don't agree with you. If you have any piping loads specified, you need to include them in the calculations, unless they are manifestly negligible, and irrespective of how they are specified or calculated.
Think of what would happen in case of problems with the anchor bolts: if you didn't include those loads, you'll be faulty irrespective of the cause for the failure, whilst if you did, someone else will take the responsibility.
And if the calculation with many odd oriented nozzles bothers you, then take them aligned along the same vertical line! Of course your design will be excessively safe and the bolts will be huge, but a design that is not on safe side is not what you are required to do.

prex

http://www.xcalcs.com

Online tools for structural design

 

[juancito]

When "transferring" your nozzle loads to anchor bolts, you need to consider the difference between permanent loads like weight and loads that normally produces secondary stresses like thermal loads. For example, a pure thermal radial load, once the flexibility of the column allows it to move in the same direction of the load, makes the load to diminish. It is very difficult to consider this factor in a hand done calculation.
Besides, if the nozzle loads you were input are all +/-, there is a way to combine it , and if I`m not wrong, accepted by ASME III NB, that is the SRSS (square root square sum).
Hope this helps

 

[bvi]

Anchor bolts, just like nozzles, can't tell the difference between piping loads due to thermal expansion of the piping, and those due to weight and pressure. The loads, with respect to anchor bolts and nozzle stress, should be treated the same. Note, for example, Section VIII, Div 2, Appendix 4 treats stresses in nozzles due to piping loads from thermal expansion and other loads the same.

 

[JohnGP]

cb4 is right juancito, anchor bolts need to resist whatever loads that the column sees, whenever those loads might arise and from whatever cause. Also, I'm not convinced that to determine loading on anchor bolts you would combine nozzle loads by means of square root sum square. Surely you would consider components of each nozzle load, multiply by corresponding moment arm with respect to column base, and add/subtract moments according to sign.

As you say, for pure thermal radial load there will be some equilibrium loading taking account of column and piping flexibilities, but simplified (conservative) approach would be to take the full load estimate.

John

 

[roca]

Hi All
Thanks again for the responses.
Can any of you quote any actual examples where anchor bolts have failed due to nozzle loads?
Thanks

 

[JohnGP]

Hi roca,
I suspect you won't hear of many instances. Personally I haven't had significant nozzle loadings to add into the mix, except for one occasion when lack of piping flexibility caused some initial concern. With factors of safety applied in normal design, it will have to be a severe loading situation to cause actual failure of the anchor bolts.

I think we have come full circle - how to decide whether to take account of nozzle loadings in foundation design. As jte said, piping loads will generally not have significant effect, but should not be disregarded without firstly reviewing the design, especially for larger nozzles with significant thermal effect (for example).

John

 

[ganvis]

If the nozzle loads are to be considered for anchor bolts design, the impact will not be limited only to the anchor bolts but will include anchor chair design too (base ring, compression ring etc.). What about skirt thickness? In some cases(tall columns), bottom shell course(s) thickness may have to be increased too (longitudnal stresses may govern due to bending moment due to nozzle loads)!
I think consideration of nozzles loads on vessel design shall be reviewed case by case. Otherwise, we will be generating hundreds of pages of unnecessary calculations.
Nozzle loads create local stresses and this has to be checked based on WRC 107, 297, BS5500 or FEA. They die down with distance for most of the cases and will not transmit loads all the way to the bottom. For short vessels, the whole vessel shell may act as rigid body and this will impose loading to the anchor area. Here again, as pointed out by Roca, some judgement to be made with regard to direction and simutaneous occurance of loading. Imagine having a nozzle at top of 120 ft tall column and with nozzle loads considered, increasing the skirt, bottom shell course(s) thickness goverened by nozzle loads.
Further, if anchor bolt has failed, it cannot be concluded that the failure is soley due to nozzle loads! It may be under designed for wind or earthquake or simply used commercial structural quality bolts.
It is not practical to consider all nozzle loads for design of vessels. It can also be argued that even if we considered all nozzles and loads acting at the same time, it may not reflect the true operating conditions and loads given by pipe stress analysis, as some of the lines may not be operating with conditions assumed in pipe stress analysis. This may create more number os cases to be analysed and will go for ever!

 

[codeeng]

The key here is flexibility. Treating a vessel as a rigid anchor is pure ignorance. Even the shortest, stubbiest, thickest vessel will have some flexibility, you just have to find it. Modern tools like FEA will provide spring rates for a given nozzle/shell geometry. These spring rates are given back to the pipe stress group and a revised stress run generated. Loads are reduced and re-analyzed. I recall a 42" flare line attached to the bottom of a flare stack. The first stress run had 100's of tons thermal load. Spring rates reduced this load by more than 90% without changing anything. If you use Compress, this has a FEA (FE Pipe) module built-in and will run using WRC input. You'll also need a FE Pipe security key.