16 replies to this topic

[tdeb]

Dear all,

I would like to know if anyone knows of any book or source that addresses the topic of the title. I'm investigating in which cases it is necessary to provide symmetrical arrangement in the inlet/outlet of an air-cooled heat exchanger, besides the obvious case of two phase inlet or outlet. I've heard there's a criteria which determines if it's necessary or by calculating the max and min pressure drop (the one for the bundle furthest away and closest respectively to the distributor inlet/collector outlet), but I've not been able to find any reference to cite.

Thanks in advance

[Bobby Strain]

Best to use symmetrical piping. Then you never get it wrong.

 

Bobby

[shvet1]

Asymmetrical manifolds are related to AC margin whether it is called design margin, fouling margin, safety factor. If your AC has 100-200% margin it is able to tolerate most improper designs of manifolds. If 5-10% - AC will not work soon. What is margin in your case?

 

Note that AC's flaws will appear only during several hours during a couple of days during the warmest month. What is the propability personnel will operate unit at max capacity during these days? Will personnel not be able to decrease unit capacity during these 2 days? What is the probability personnel will operate AC at max design fouling conditions during these days? What is probability factors above occur simulteneously during a year when worst combination of air temperature and humidity occur?

 

The most important - will personnel associate this upset with improper design? In any case you are able to blame personnel in AC's overfouling as noone is able to measure that.

 

Impunity is the reason many designers use a random design of manifolds.

Edited by shvet1, 09 November 2023 - 08:42 AM.

[tdeb]

Hey shvet1, thanks for your reply. The AC margin is 20%. The analysis I made was a comparison of the pressure loss calculated in the piping vs in the AC. The pressure loss in the inlet/Outlet piping is about 1% of the one calculated with HTRI in the AC, which leads me to conclude that there will not be a tendency of the fluid to distribute asymmetrically, as it would mean a much bigger change of DP in the AC than in the piping. What I'm trying to investigate is wether there's something else I should be evaluating to determine if it's acceptable to avoid providing symmetrical arrangement. The inclusion of symmetrical piping would mean a very complex piping arrangement as it is a quite large air cooler, so if it's possible to avoid it would be pretty beneficial

[Bobby Strain]

What are you cooling? Given the pressure drop I assume it is a liquid.

 

Bobby

[tdeb]

It's actually a condenser, overheated vapour comes in the inlet, and the outlet line is designed for self venting

[Bobby Strain]

What is the fluid? Is it a refrigerant condenser?

 

Bobby

[Vegeta]

Dear tdeb,

How many bays (or bundles) does this AC have?

[tdeb]

Yes, it's a refrigerant. The AC has 18 bundles

[Bobby Strain]

You can arrange the inlet header from end-to-end. The outlet header should run the same way. Thus the highest pressure in the inlet and outlet are at the same position. You can look at the arrangement in one of the LNG plants. Your engineering contractor should have this knowledge. Also, the coolers should be elevated such that turning off fans will not blow vapor to the outlet. And my experience is to also use a low pressure drop for the cooler.

 

Bobby

Edited by Bobby Strain, 13 November 2023 - 12:13 PM.

[Pilesar]

Also, the coolers should be elevated such that turning off fans will not blow vapor to the outlet. 

 

 

Bobby, I have practiced the end-to-end header arrangement and spec'ing low pressure drop. Would you explain further the effect of elevation on vapor blow? Is this describing putting the process outlet a bit lower than the inlet to keep a liquid seal or something else? Is this something to put in specifications or is it a standard feature?

  My air cooler designs were usually 'two deep' so one half had variable speed fans for turndown or louvers and sometimes steam coils for winter conditions. I left the detail design to the specialty vendors who should know better than me how to build these things.

[Bobby Strain]

I once had a YORK manual with an illustration. But it's long gone. If you still have one around, check it out. If not, I'll do some work. But the idea is that there is a vapor line from the condenser inlet to the refrigerant receiver. Thus, a liquid head is required on the condenser outlet. When a fan is turned off, there is an increase in the outlet pressure from that bay. The higher pressure depresses the liquid in the outlet line. There should be sufficient elevation to prevent vapor flow at this condition. You have also seen such operation with a vapor condenser on a column where the pressure is controlled with a valve on the liquid outlet to the drum to provide variable exchanger surface area.

 

Hope this helps.

 

Bobby

[Pilesar]

Thank you, Bobby. Okay, here is what I think I learned. The vapor line from the condenser inlet to the refrigerant receiver maintains the pressure in the receiver. If there were no equalization line, the pressure in the receiver would drop due to all the vapor condensing. If the pressure dropped too far, then the condensing temperature would also drop below the temperature of the cooling medium and condensing would stop until pressure builds again. The condenser receiver system would operate sporadically. So it is better to keep constant pressure on the receiver by use of the vapor bypass line in order to run smoothly.

  With the vapor bypass equalizing the pressure, then the condensed liquid must flow by gravity into the receiver. The piping from the condenser to the receiver should have low sections or 'P Traps' to serve as constant liquid barriers to vapor blow. These traps should be sized to handle all expected flow conditions. There is pressure drop through a coil while condensing. When one part of the condenser is 'fan off' then there is less pressure drop through the coil in that section. Since there is a pressure equalization line, the pressure drop in the active coil must be compensated by increased liquid level in the downleg of that section. So the downleg height must be long enough to offset the coil pressure drop.

  There is an illustration and explanation of the elevation requirements on page 12 of this link: https://www.evapco.c...e 131B 0121.pdf . I think the brochure explanation is the same as you gave. I will have to think more on it so it will sink in. Epiphany events come slower to me than they used to!

[shvet1]

@Pilesar

The idea is that condensers are gravity-driven and negligible pressure difference has a drastic effect causing overflooding and reducing capacity. Every point of system where pressure build-up is possible shall be vented:

- a liquid receiver

- condenser bonnets

- liquid lines from a condenser to a receiver

 

Also note that liquid contains entrained bubbles which accumulate and reduce effective bore of piping.

 

See an example attached.

Attached Files

•  1.png   182.92KB   1 downloads

Edited by shvet1, 14 November 2023 - 01:40 AM.

[shvet1]

@tdeb

 

20% might be a lot or a little depending on number of bundles and fouling factor applied.

 

If someone has the article marked bold please share.

 

 

Foster Wheeler's std. 302

 

4.8 Distribution Piping

It is usually impractical to have a completely symmetrical piping arrangement at the air fin inlet or outlet. To minimize the effect on the flow distribution of non-symmetry, the pressure drop in the distribution piping should be small in comparison with the total overall system pressure loss.

In general, the preferred piping system is as shown in type A&B with inlet and outlet headers running the full length of the air fin header box with direct connections from the header to each pass.

 

 

Type A (Co-Current)

The characteristics of this type are:

*Simple arrangement

*Low Pressure Drop

(1) Poor flow distribution and hence unbalanced heat transfer.

 

Type B (Counter Current)

The characteristics of this type are:

*Simple arrangement

*Low pressure drop

(1) Poor flow distribution but better than Type A.

This is the preferred type for liquid flow. ...

 

Type C (Multipass)

The characteristics of this type are:

*More complex structure

*High pressure drop

*Good flow distribution

In multipass two phase service the numerous bends and T's promote fluid separation.

For additional information on flow distribution see Chemical Engineering June 17, 1968, pages 210-213.

One known refinery has solved it's two phase mal-distribution problems by adding notched stand pipes to the individual inlets in the inlet header. Apparently the notched stand pipes distribute the condensed liquid to all inlets allowing the vapor to readily flow through each coil pass.

(1). The design of piping distribution systems should be reviewed very carefully, particularly in large units operating at low pressure and where piping frictional losses are to be minimized for vapor services.

 

 

UOP's std. DM-EQUIP-2073-402

 

3.4 Bundle Arrangement

The arrangement of bundles is listed as a means of specifying the header layout for the air cooler.  ... There are three options:

a. None

Use for a D type header.  There is no limit on the number of bundles.  The flow enters one end of the distribution header and branches to each bundle as it flows by.  The collection header reverses the process with the flow out the end of the header at the same end as it entered.  This arrangement is acceptable for single-phase fluids at the inlet and low pressure drop designs.  This is the most commonly required layout and the default value.

b. Limit Bundles to 2, 4, ... 2n

Use this statement for a C type or F type header.  In these arrangements, the inlet pipe is split only into two lines.  If necessary, each of these is symmetrically split into two more lines, etc.  Therefore, only 2, 4, 8, 16, ... 2n bundles are possible.  The flow is mirrored in the collection manifold in a type F manifold.  In a type C arrangement, the collection manifold is like a type E header (even number of bundles).  This arrangement provides the best possible distribution of two-phase flow at the inlet.  However, the manifold has a high pressure drop, requires a lot of room, and is quite expensive.  For very small units when only one bundle is required for a service that usually requires this note, “None” may be specified instead.  See Unit Specific Guidelines for more guidance on when to use this option.

c. Even Number of Bundles

Use for an E or Y type headers.  This arrangement is similar to a D type except that the inlet to the distribution manifold and the outlet of the collection manifold are in the center of the manifolds.  Therefore, the number of bundles to either side of the inlet/outlet pipe is equal.  This will provide a better distribution of fluid for two-phase flow at the inlet.  It also provides symmetry in the headers.  The pressure drop is higher due to the use of more elbows in the piping.  The Y type header is a variation in this concept in that there are two nozzles but only a single bundle.  See Unit Specific Guidelines for more guidance on when to use this option.

 

Edited by shvet1, 15 November 2023 - 03:52 AM.

[shvet1]

 

BECHTEL's std. 3DG-P22-00004,

 

8.4.2 Air Cooler Piping Arrangements

The piping arrangements called for at air coolers can also greatly affect their placement and elevation. In cases where true symmetrical (cascaded) piping is required to stabilize two-phase flow, the required elevations of supports for the piping can become prohibitive. In such cases the use of header box inserts may reduce the need for symmetry to the point that a rake style arrangement may be used (Fig. 30 and 31). The Responsible Engineer shall be consulted about such possibilities.

Edited by shvet1, 20 November 2023 - 12:29 AM.

[Bobby Strain]

Here is a propane condenser on a large LNG plant. The outlet is not shown but each is not sealed with a P-trap.

 

Bobby

Attached Files

•  Screenshot 2023-11-23 at 17-18-02 (PDF) Angola LNG Project Progresse Photos May 2009 to May 2010 - DOKUMEN.TIPS.png   717.19KB   0 downloads