Heat Exchanger Pass Partition Plates

[marty007]

I'm working on a heat exchanger design and am having a disagreement with another engineer regarding tapering of a pass partition plate at the gasket surface.

Exact geometry aside, let's consider a simple 2-pass AEU heat exchanger with a single pass partition plate in the channel. Our customer has specified that we should follow TEMA R requirements, so for sizing the thickness of the pass partition plate, we refer to section RCB-9.13 and equation RCB-9.132.

Now let's assume that the pressure drop across the plate is quite high (30psi), and we calculate a required thickness of 3/4".

Questions -
[ol 1]
[li]Does this thickness have to be maintained for the entire pass partition plate (not allowed to taper at gasket seating edges to reduce gasket seating loads), or can the partition plate be tapered to reduce the gasket seating loads?[/li]
[li]If the exchanger is being designed per ASME VIII-1, are there any additional complications due to the pressure drop exceeding 15psi?[/li]
[/ol]

Thanks all.

 

[dlmeteng]

Marty,

As to your first question, TEMA RCB-9.131 states that "pass plates may be tapered to gasket width at the contact surface." Good design practice needs to be used to determine the extent of the taper. I tend to see no more than 1/4" difference between the pass plate thickness and the end of the taper.

In regards to the high pressure drop, I am not aware of any place in VIII-1 that addresses this. RCB-9.134 mentions that "special consideration" is required when dealing with "extreme differential pressures". However, while 30psi is slightly higher than the pressure drop in the units I generally see, I would not classify it as "extreme".

DL

 

[GoodCoffee]

Marty,

I do believe it is a common practice to taper the ends of a pass partition plate. Whether this affects the ability of the pass partition plate to hold pressure, I would say no.

We have to step back and see what the TEMA differential pressure formula for a 2-pass partition plate is calculating. This formula comes from the classic Timoshenko theory for flat plates, presented again in "Roark's Formulas for Stress and Strain". Roark's states that for a flat plate with fixed-simply supported boundary conditions (fixed on two opposite sides, simply supported on the other two opposite sides), the maximum stress occurs in the middle of the fixed edge when uniform normal pressure is applied onto the plate face. While the Roark's version solves for stress, the TEMA version rearranges the formula to solve for thickness, setting the stress to 1.5xAllowable for the material. Keep in mind, that this is the stress on the plate, in the middle of the fixed edge (the longitudinal center of your channel). Changing the end geometry of the partition plate would not change the maximum stress in the middle, so long as you are not changing the meaningful length of the edges.

What TEMA does not do, is determine what the impact of high DP flat plate bending is on the channel cylinder itself. As the flat plate absorbs load and displaces, it does induce some meaningful stress in the channel pressure boundary that usually goes unaccounted for. It is especially significant when the channel cylinder thickness is not very large (i.e. low tubeside pressure, high DP). There is a Mandatory Appendix in ASME Section VIII Division 1 where this is explored. Appendix 13-13 discusses "Vessels of Circular Cross-Section Having a Single Diametral Staying Member", and provides formulas to solve for stresses in the plate and in the cylinder. I suggest you take a look at these when you get into the high DPs. The industry has generally not paid much attention to this interaction, but it is very real and can affect the pressure boundary. Usually the plate fails before the shell can yield, at the welds or by jumping the groove, but there may be instances where the pressure boundary can yield.

Good luck!
-Andrey

 

[CuMo]

Our company practice is to taper pass partition plates to a minimum of 6 mm thick edge if needed.
The taper ratio is usually 1:4.
For the pressure drop - as advised above you can check stresses by Roark & Young formulas for flat square plates.

 

[muld0020]

1. Taper is ok. Tapers are allowed in code construction and are common in industry.

2. No complications over 15psi; except below some comments:

High pressure drop requirements (over 15psi) can significantly increase thickness of these plates. I once had to design a partition plate for 100% design pressure acting on one side in case the heat exchanger became blocked during operation. In general these welds are highly stressed. I have performed FEA on many of these plates the taper is fine. You really need thickness or stiffeners in the middle of the plate to reduce the deflection. If you are welding a thin partition to a thick channel you should consider preheat (ie 1/2" plate to 1" or thicker channel in carbon steel). If you have a very thick plate you can also look at partial penetration welds on the attachment to the channel, but these plates benefit from full penetration when possible. I would avoid double fillet welds as much as practical. I agree with good coffee that these attachments can significantly impact your pressure vessel, as always, when you are beyond the scope of Sec VIII D1 you can use U-2(g).

I agree with the other posters you can also use roark's