12 replies to this topic

[cea]

Dears,

 

We have two cooling water circuits for common cooling tower (one at 5.5 kg/cm2a to cater requirement of user at higher elevation with dedicated pump and one at 4.5 kg/cm2a to cater requirement of user at lower elevation with its respective pump) . Please refer attached schematic for better understanding.

 

We intend to connect both the headers into single header before it goes to cooling tower deck. This is view of having single point connection & thus reduce load on cooling tower structure.

 

Following are my queries.

1) Due to elevation difference from where HP CW & LP CW returns to common header, is it expected that HP CW return header may offer some back pressure on LP side & thus reduced flow through LP circuit?

 

2) If so, what could be the remedy. Does intoduction of RO in HP circuit will overcome above problem?

 

Kindly note that line sizes are adequate with respect to standard velocity criteria & to supply required flow to users. The sizes are not indicated in diagram & assumed is not the matter of concern.

 

Kindly help.

 

Thanks & Regards,

Attached Files

•  Scheme.xls   91.5KB   240 downloads

[shan]

You do not need do anything. The hydraulic system will be balanced itself. For example, if HP circuit pressure is higher at the return point, more CW will flow through HP side, which leads increase of line pressure drop and pump operating point shift right (lower discharge pressure).

[katmar]

Your diagram shows vents at the highest point of each circuit.  This is the correct way to do it.  The vents make the pressure at the top of the return lines equal for the two circuits - both are at zero gauge pressure.  It is only the static pressure in the return lines that drives the water back to the tower.  An important point here is that the vertical legs of the return circuits must be sized for self venting flow to ensure that air is not sucked down the vents.  Drawing air in this way will cause maldistribution in the towers.  A search through this site will give many discussions on self venting flow.

 

The only way that you could get interference between the HP and LP circuits would be if the return lines backed up all the way to the highest point of the LP system.  This would be extremely unlikely, especially if you have put some effort into sizing the return lines properly.

[Atttyub194]

Dear All

 

Good day!

 

Please note the following:

1. The decision  can not be done without detailed calculations which of course require dia of each pipe, expected flow and more detailed hydraulic evaluation
2. The vents provided are only used for start up of line and shall be closed once the loop is pressurized and in service
3. We have similar problem and then we have addressed by  adjusting the inlet and outlet size of the headers. Till resolution of the issue, we used to pinch inlet valve of exchanger at lower elevation as per process requirement and we operated as such for couple of years as we have margin available in pumps.The return header back pressure increased from 1.1 to 1.8 kg/cm2 which after modification came back to original
4. Please do check suitability of your cooling water distributor and fills, if the modification is new with increased flow
5. Lastly, once you install RO it will be difficult for you to gain operating flexibility  till next turnaround and you may face frequent chocking & MIC issues due to low velocity in exchanger at top . Therefore, please give due consideration

Best regards

 

Ahmed Attyub

[S.AHMAD]

1. The pressure drop in the return line to the cooling tower is expected to increase and this gives higher back pressure (system characteristic will be difference)

2. Higher back pressure will reduce the flowrate of each pump and using the new system curve and the pump curve you can determine the new flowrate.

3. If the new flowrate is acceptable, then do nothing. Otherwise you need to increase the size of the single return line to cooling tower.

4. Adding orifice will make the situation more critical. Don't do it!

Edited by S.AHMAD, 24 March 2013 - 07:16 PM.

[katmar]

Cooling water circuits can be operated as described in post #4 by adjusting valves and changing header sizes, but it is much better practice to keep the vents open and to allow the water to flow back to the cooling tower under gravity.  This is the most flexible and safe way to do it. If your highest exchangers are more than 10m above the towers then you run the risk of vibration in the return piping if you do not keep the vents open.

 

Keeping the vents open also solves the problem discussed in post #5.  If the vents are open and the return is by gravity then there is no effect on the back pressure of the pumps. There will be no interaction between the pumps if the vents set the return pressure to atmospheric.  This is the universal way to do it and whoever designed your original circuit knew what they were doing. Leave it that way.  Just check that the return piping is big enough for the combined flow under gravity, and that the vertical sections are self venting. Keep it simple.

[cea]

Cooling water circuits can be operated as described in post #4 by adjusting valves and changing header sizes, but it is much better practice to keep the vents open and to allow the water to flow back to the cooling tower under gravity.  This is the most flexible and safe way to do it. If your highest exchangers are more than 10m above the towers then you run the risk of vibration in the return piping if you do not keep the vents open.

 

Keeping the vents open also solves the problem discussed in post #5.  If the vents are open and the return is by gravity then there is no effect on the back pressure of the pumps. There will be no interaction between the pumps if the vents set the return pressure to atmospheric.  This is the universal way to do it and whoever designed your original circuit knew what they were doing. Leave it that way.  Just check that the return piping is big enough for the combined flow under gravity, and that the vertical sections are self venting. Keep it simple.

Thanks to all for your valuable input. 

 

However, still a small doubt in my mind.

Let's assume that

1) all lines are  adequately sized, fullfilling requirement of self venting.

2) CT. height is 10 m

 

In that case the back pressure at common header would be about (1 kg/cm2a (atm press) + 1 kg/cm2a (equivalent to 10 m height) + 0.015kg/cm2 (15m total length x 0.001 kg/cm2/m frictional drop) = 2.015 kg/cm2a

 

Now from LP header, pressure available at return point due to elevation would be 1 kg/cm2a (atm press) + 1.4 kg/cm2a (equivalent to 14 m height) - 0.015kg/cm2 (15m total length x 0.001 kg/cm2/m frictional drop)= 2.385 kg/cm2a

 

For HP header, pressure available at return point due to elevation would be 1 kg/cm2a (atm press) + 2.7 kg/cm2a (equivalent to 27 m height) - 0.03kg/cm2 (30m total length x 0.001 kg/cm2/m frictional drop) = 3.67 kg/cm2a

 

If I am right in my understanding, LP header return pressure is almost same / just above the requirement to overcome system pressure drop after exchanger outlet to common header.

 

However, for HP header the available pressure is much more than this requirement. At some point it shall be killed, so as to have one pressure level at common point (at header) or it will create back pressure in another system.

 

Correct me, if I am wrong.

 

I would like to highlight that the flow requirement at highest elevation user with allowable pressure drop in exchanger as back pressure, have been considered as basis, while carrying out pump hydraulics. In short, pump calculation loop doesnot include downstream system of exchanger. However, common header size have been checked for pressure drop, in order to ensure that pressure gradient available is sufficient to overcome pressure drop in return header.

 

Kindly help.

[katmar]

cea, this sounds like a good design that you have got.  The pump hydraulics have been correctly calculated if the return piping after the highest exchanger has been excluded.  As stated earlier, the driving force for this return flow is gravity.

 

On your LP circuit, the 10m of head you gain going down from the CT level to the header is offset by the 10m of head you lose going back up to the CT deck (as you have shown) so the only thing to check is if the head available above the CT is enough to overcome the friction losses.  You have 0.015 kg/cm² friction loss in the header and another 0.015 kg/cm² friction loss in the LP return line making 0.03 kg/cm² total friction loss.  The available head you have is 4 m of water column, or 0.4 kg/cm² which is way more than you need. As you have shown, the HP circuit is even safer. 

 

You have said that we can assume that the vertical sections are self venting.  The pressure drop per meter indicates they are not.  You should check this.

 

 

[cea]

cea, this sounds like a good design that you have got.  The pump hydraulics have been correctly calculated if the return piping after the highest exchanger has been excluded.  As stated earlier, the driving force for this return flow is gravity.

 

On your LP circuit, the 10m of head you gain going down from the CT level to the header is offset by the 10m of head you lose going back up to the CT deck (as you have shown) so the only thing to check is if the head available above the CT is enough to overcome the friction losses.  You have 0.015 kg/cm² friction loss in the header and another 0.015 kg/cm² friction loss in the LP return line making 0.03 kg/cm² total friction loss.  The available head you have is 4 m of water column, or 0.4 kg/cm² which is way more than you need. As you have shown, the HP circuit is even safer. 

 

You have said that we can assume that the vertical sections are self venting.  The pressure drop per meter indicates they are not.  You should check this.

Dear Katmar-

 

Thanks for validating our design.

 

Yes, as you rightly pointed out that there will be additional 4 m head available in LP circuit & about 14 m in HP circuit after overcoming all the losses. My query is, where this additional pressure will get killed at downstream of exchangers? And secondly, whether the additional head in HP circuit will create any back pressure on LP circuit, that may cause reduction in flow in LP circuit?

 

Kindly address these above two queries.

[katmar]

The vents set the back pressure to atmospheric on both circuits. The back up in the return lines will set itself automatically to overcome the friction in the return line to the cooling tower. The 14m and 4m are the maximum the backups can be, but as you calculated yourself, they will actually be 0.45m and 0.3m in operation.

[cea]

The vents set the back pressure to atmospheric on both circuits. The back up in the return lines will set itself automatically to overcome the friction in the return line to the cooling tower. The 14m and 4m are the maximum the backups can be, but as you calculated yourself, they will actually be 0.45m and 0.3m in operation.

Thanks Katmar for your brief explaination.

 

I have tried to picturised your explaination in attached sheet. Hope, this is in line with the concept.

Attached Files

•  Scheme.xls   127KB   118 downloads

[S.AHMAD]

Refer to post#5 for flowrate reduction. However this problem can be solved by proper sizing of pipeline.

[katmar]

No, your diagram is wrong.  I misread the height of the HP circuit.  Here it is corrected to 17 m.

 

The 17m and 4m (above the CT deck) are the maximum the backups can be, but as you calculated yourself, they will actually be 0.45m and 0.3m (above the CT deck) in operation.

 

Because the HP and LP levels in the vertical sections will be below the LP return they will not affect each other.  You do not need an RO or other device.  Your return piping (if you have calculated the pressure drops correctly) is already adequately sized to ensure the two circuits do not interact. The only thing you have not convinced me of is the sizing for self venting flow.