Gasket seating width on pass-partition plates (heat exchanger)

[marty007]

Hello all,

Per ASME VIII-1 Appendix 2 Table 2-5.2, the effective seating width of a gasket is reduced by a factor depending on geometry (my understand is that is to account for flange rotation when torquing flange).

Now, for a heat exchanger that is also seating a gasket on the base of pass partition plates, TEMA paragraph RCB-11.7 provides guidance on how the ASME calcs should be modified to account for the additional force to seat the gasket under the pass partition plate(s). In the TEMA guidance, it refers to a variable: "b[sub]r[/sub] = Effective seating width of pass partition rib(s)".

How do most of you handle this effective seating width. Do you take it as the maximum width that can be compressed based on groove depth/pass partition plate geometry? Or do any of you reduce this effective width based on any guidance similar to ASME VIII-1 Appendix 2 Table 2-5.2 (if so, can you point me in the direction of appropriate guidance?)?

I am asking this question because I am designing a replacement channel for a heat exchanger, and my first-pass design (based on using maximum pass partition gasket width) shows insufficient bolt area. The exchanger pre-dates UHX, but not by much, and the flanges would have been designed to Appendix 2 regardless. I'm just trying to understand how the original design would have been calculated (original calcs are not available... sigh). The new design is complicated by this issue because I'm having to mate up to the existing bolt hole pattern on the shell, but my calculations show I need either large or bigger bolts to seat the gasket... ugh.

Cheers,
Marty

 

[SnTMan]

marty007, I am pretty sure different progams do it different ways, i.e. either full width of rib, or reduced rib width per Apx 2. I think in your situation, you would be justified to consider reduced rib width. Replacement work is hard enough

Regards,

Mike

 

[gr2vessels]

Marty, my two cents. Maybe, just maybe you forgot to use the tapered width of the partition rib as effective gasket width, as mentioned above!

 

[marty007]

I wish it was that easy, but the flange barely passes even if I completely disregard the pass partition plates... it seems as though they were completely ignored in the original design

 

[muld0020]

The effective width (br) is half the width of the Partition thickness up to 1/2" pp thickness. The actual equation is:

b0 = Gasket Width / 2 ; if b0 > 1/4", br = 0.5*(b0)^0.5 ;; if b0</= 1/4", br = b0

It is probably easiest to change the gasket material. Another option, depending on configuration, would be to not seal into the cover and have and independent bolting ring and cover. If you have a channel bolted to tubesheet that may be a problem.

 

[marty007]

Muld0020, do you have a reference for that? Unfortunately a gasket material change isn't in the cards.

We are going to be making a bolting change from SA-193-B7 to SA-354-BD. The existing shell flange is suitable, and with the slightly higher allowable stress we can achieve the required bolt load to seat the gasket with the existing bolt number and diameter. This approach just barely passes the code calcs.

 

[SnTMan]

marty007, muld0020 will no doubt tell you the referrence is Apx 2. Pass ribs were routinely ignored in the past. Mandatory now of course.

Regards,

Mike

 

[muld0020]

Sorry for late post, but figured in case anyone is also searching for this.

Yes App 2. You can also reference TEMA RCB 11.7

 

[marty007]

Thanks a lot everyone, it was useful to get a bit of a reality check. SnTMan, I've noticed it a few times now where I'm having trouble getting the calcs to work for a channel replacement, and I've often suspected that the original design ignored the pass plates. So far I've managed to find work-arounnds.

Cheers,
Marty

 

[dineshakabari]

I do agree with muld0020 that width shall be reduced to consider as effective width for gasket for partition plate gasket also.

 

[nick262b]

Not sure if this is closed but just a few general points.

The seating width calculations I believe were derived experimentally based on testing done on rubber gaskets a long time ago and are not necessarily the best way to design newer gaskets as kammprofiles for example won’t rotate or deform the same way under the load and the full area is more appropriate than an ‘effective width’.

The calculations in ASME VIII are based on a lot of general figures and data with safety margins put in and as you are aware give you a bolt load on the larger of the force needed to seat the gasket or the force to maintain operation of the gasket after seating and retain the pressure end load. The m and y factors I have found are not always representative of manufacturers more modern gaskets and the ASME VIII method can give quite different bolt loads to those calculated when using ASME PCC methods which also allows for things like relaxation, thermal effects and assembly spread etc. The bolt stresses that you are calculating with the design code (ASMEVIII) are not necessarily the ones that you will actually use in service. Refer to ASME PCC-1.

I have found it best to use the PCC methods described in appendix ‘o’ and use the manufacturers figures for min seating stress and min op stress as opposed to the code values. Bolt loads using this method are generally higher but will give better gasket compression in service under a range of conditions. As another point It’s always best to try and design your flanges to be stronger than the bolt yield to prevent damage from over tightening. I have dealt with many large joints that are poorly designed in this aspect and regularly leak due to low bolt loads. If you are finding your flanges are very large to accomplish this for the bolts you need then you can also refer to WRC 538 which proposes alterations to the ASMEVIII method and allows higher stresses based on the yield of the material.

Cheers Nick