Partition gasket m & y

[mechengineer]

Hello,
The pressure drop betweem the tube passes is very low (~1 bar)to compare with the internal design pressure (100 bar) of the heat exchanger, thus m & y valuse of the weld bar for the partition gasket should be much smaller than the m & y valuse of the gasket for the sealing pressure boundary so that avoid increase the bolting force unnecessary due to the higher valuses of m & y for the partition gasket.
Anyone has experince to request the m & y for the partition gasket aeparatly from the main gasket?

 

[SnTMan]

...so that avoid increase the bolting force unnecessary due to the higher valuses of m & y for the partition gasket.

I'd consider this a very poor practice. The bolting is not intelligent enough to properly distribute the seating force between the main gasket and the pass rib. Much better to have too much bolting rather than too little.

Anyone has experince to request the m & y for the partition gasket aeparatly from the main gasket?

Yes, one common example: A spiral wound main gasket with a jacketed pass rib. Each portion should use the appropriate m, y values.

Regards,

Mike


The problem with sloppy work is that the supply FAR EXCEEDS the demand

 

[r6155]

For DP 100 bar the gasket shall be metallic, one piece of uniform material. No leaks from pass partition gasket.

See ASME PCC-1 2019 -Where a gasket has additional gasket area, such as a pass partition gasket, which may not be as compressed as the main outer sealing element due to flange rotation, then a reduced portion of that area, such as half the additional area, should be added to Ag. (area of gasket)

See ASME VIII Div 1 -Tubesheet deflection, especially when the tubesheet thickness is less than the tube diameter, may contribute to tube-totubesheet joint leakage; likewise, deflection of a tubesheet or flat bolted cover may result in fluid leakage across a gasketed pass partition plate. Such leakages can be detrimental to the thermal performance of the heat exchanger, and deflection may need to be considered by the designer.

Regards

 

[SnTMan]

...deflection of a tubesheet or flat bolted cover...

Point of interest (or not): TEMA Standards does consider deflection in the design of bolted flat channel covers.

Regards,

Mike

The problem with sloppy work is that the supply FAR EXCEEDS the demand

 

[BJI]

Additional reference:

7.8.3 At least one-half of the gasket pass-partition rib area shall be added to the peripheral ring gasket area, Ag, unless otherwise agreed with the purchaser.

 

[mechengineer]

Disagree,
1.For example, a kammererite gasket, it limit by the maximum permissible gasket stress (350 MPa) and the minimum gasket seating stress (140 MPa) per API660 Table 3. In this case, an accurate, effective gasket area is very important.
2.The total rib area for the partition can reach 30~40% of the perimeter of the gasket for multiple passes. Even if we consider >50% of ribs area per API660, that will still have 15~20%. That means 15~20% of bolts are for the partition. However, actually the sealing pressure for the partition is only the pressure drop, mostly smaller than 1 bar, and has some drain holes on the partition plate. It should not be considered the partition rib to seal the design pressure or half of design pressure. In another word, theoretically speaking, it should not be considered the partition rib sealing as the design pressure sealing on the perimeter of the gasket. It just keep it without a gap in between the ribs and the gasket. Normally we use a full bolt load to design the flange, it has 30% margin if the allowable bolt stress is 70% as per API660. It seems not necessary to take the rib area to make the design ‘much better’again.

 

[BJI]

mechengineer, you are mixing min seating stress with min operating stress for the gasket. The required min operating gasket stress is much lower than 140 MPa for kammprofile gaskets. The problem is you don't know exactly how the load will distribute between the periphery and rib gaskets, or to what degree the flange will rotate and unload the pass partition, but like it or not it will take load, resulting in a reduced gasket stress on the periphery. You would be better off to design assuming a higher rib area than less, which is why it can't be reduced below 50%, otherwise you might end up with insufficient gasket stress in your periphery gasket. However, designing for 100% area is typically too conservative, particularly with a high number of passes.

You are correct in saying that with a low DP across the pass partition only really the min operating gasket stress is required, however, I don't see how you could accurately design and fabricate to the required accuracy when considering the complicated thermal and mechanical interaction at a TS joint.

 

[SnTMan]

mechengineer, go ahead and minimize on the bolting if you feel lucky

The problem with sloppy work is that the supply FAR EXCEEDS the demand

 

[r6155]

@ mechengineer
Do you have any experience in the fabrication and inspection of heat exchangers? Did you ever witness the bolt tensioning procedure?, and for bolts 3” or more in diameter?.
Do you have any experience during pressure test of heat exchangers?
Sorry, you are a little confused.

Regards

 

[mechengineer]

My point is to seek a possibility to use a different material (very low y and y) from the perimeter for the partition of gasket. If a gasket manufacturer can't do so and just use the same material for the whole gasket. In that case, there is no choice, we have to use a higher bolt load for the partition gasket.
@r6155, so may questions seem to be bragging that you know everything, but you did’t say anything valuable here. If you have no idea, no matter how smart you think you are, better to be quite.

 

[mechengineer]

See the attached sketch of the gasket, if this is fessible, the bolting load for the partition gasket may be negligible. The innovation gasket cost is another issue. If you saying that 'my gasket can reduce 20% bolting load', sure you will get a lot of orders.

 

[SnTMan]

mechengineer, for your proposed design you could certainly use a lower m, y for the rib than used for the spiral wound portion. Not much different than the example I cited earlier, except rib is not jacketed but uses soft facing instead. Proper m, y could likely be determined from the table in Apx 2, or in consult with a gasket manufacturer.

You need a means to insure the soft facings will remain in place in service. You cannot depend on the joint clamping force to do this. This is why corrugated or grooved metal cores (so called graphonic & kammprofile styles) are used with soft facings such as graphite. Trust me on this

Both graphonic (corrugated metal, graphite covered, CMGC) and kammprofile (grooved mnetal, graphite covered, GMGC) are routinely used for HX body flange gasktets. Either type rib could be supplied with a spiral wound ring portion. These ribs would have m, y like maybe 2, 2500.

Regards,

Mike

The problem with sloppy work is that the supply FAR EXCEEDS the demand

 

[Vikko]

The pressure drop betweem the tube passes is very low (~1 bar)to compare with the internal design pressure (100 bar) of the heat exchanger

According to the Taylor-Forge method both the constant m and y are independent of the pressure ! m and y are only a function of the type and material of the gasket and are tabled in tab. 2-5.1 of ASME BPVC VIII-1 or EN 13445-3 annex H.

y value is for the seating / assembly / bolting condition at ambient temperature, so when the pressure is zero (0). It is meant as a minimum stress value to prestress and compress the gasket up to a value considered safe in order to make the gasketed bolted joint work in the properly intended behaviour.

m value is just a safety multiplier applied to the design pressure which is considered enough to make the gasketed bolted joint work while in operating condition at design temperature and pressure, which have a direct effect on the bolt load and flange moment.

The Taylor Forge method is intuitive and has been demonstrated good in practice, despite it ends in overdesign of bolted joint. Most of the time in the past the partition area was not even considered at all in calculations, as far as I understood. If you think 1 bar of pressure drop is not a lot, why worrying in sealing it ? Just avoid the partition in the gasket and let a bit of space distance from the head internal partition to the tubesheet: how much flow can it pass from a little space ? I think that is not a lot. Many time just some silicone paste on the partition is more than enough to seal, but even without the silicone, the heat exchanger works more or less the same. So all this partition "sealing" is sometimes a waste of time and resource (of course it is not possible to really seal one tube pass from another in a heat exchanger, and in any case there is no way to test it).

* the gasket thickness and material shall be uniform for all the sealing contact surface. From a manufacturing and construction point of view I cannot imagine how to build a spiral wound gasket with a partition...

 

[SnTMan]

I cannot imagine how to build a spiral wound gasket with a partition...

Simple. A metallic pass rib(s) is welded to the inner ring of the SW gasket, much as in mechengineers' sketch.

And Vikko, I do not agree that sealing of the pass ribs is unimportant...

The problem with sloppy work is that the supply FAR EXCEEDS the demand

 

[mechengineer]

Thanks, SnTman. You got my point. I believe you. But I am going to discuss the possibility to make the partition gasket 'soft' with the gasket vendors and see how.

 

[Vikko]

metallic pass rib(s) is welded to the inner ring of the SW gasket, much as in mechengineers' sketch.

I have seen the sketch, and I do not think that the gasket metallic rib contact surface and the gasket spiral wound plates can be considered as a uniform material. Not to say about the thickness: even in the sketch the thickness of the partition rib of the gasket and that of the spiral wound plates are not the same... I do not think the Taylor Forge method is applicable in this case, and not even EN 13445-3 annex G and not also EN 1591-1. As far as I understand the contact sealing surface on a flange or on a tubesheet shall be at the same depth level with controlled roughness, while the gasket material shall be uniform and of same thickness along all the sealing contact surface, in order to apply any standard calculation method.

Here the reference of gasket for the application of EN 1591-1:

I do not agree that sealing of the pass ribs is unimportant

I did not say that. But can you test the sealing of one tube pass alone in a 2-tubeside-pass heat exchanger ? No. Never done. Not possible at all.

 

[SnTMan]

Vikko, no need for you to take my word on these things. I recommend you consult with the applications engineers of any reputable gasket manufacturer.

Correct, you cannot normally pressure test across tube passes. Does not mean by-pass flow can be disregarded. Run this by your rating engineer.

The problem with sloppy work is that the supply FAR EXCEEDS the demand