Cooling Water System

[NPC5]

I have been performing a cooling water system calc in Excel. The system is being pressure balanced using butterfly valves. I have always heard that in real life it is incredibly difficult to actually balance a cooling water system this way. I am wondering if the calculations that I am performing would actually give me something that sort of resembles the actual system. Say I put in the pipe length, size, fittings, flows that I want to go through and use the butterfly valves to take the excess pressure drop I need to balance everything, would the flow rates that I have specified resemble what would happen out in the field? Thanks for any help!

 

[UmeshMathur]

If you want good flow regulation, a butterfly valve is generally a poor choice. The flow characteristics are just not suited to such a task. That's why one should specify a control valve if you want really close regulation.

A better choice than the butterfly is a globe valve for reasonably steady flow. I would estimate, as a rough rule of thumb, that the valve absorbs at least 30% of the total DP available.

It's very hard to provide advice when the details of the problem are not known.

 

[dcasto]

Butterfly valves are very good in your application. Water is a terrible thing to waste, but you waste nothing in a cooling water system. The use of control on a water exchanger is a matter of saving energy downstream of the water cooled exchanger, on the process side. For example, sub cooling reflux isn't the most energy efficient thing, so some controls are good. Getting the maximum cooling on a refrigeration condensor is good, so flow control isn't required other than lowering the flow so that other water users have some water.

 

[HEC]

OP is looking for pressure balancing, I have to agree wiht UmeshMathur, butterfly valves are not suitable for the application. A globe or angled seat globe valve is more suited to the application. Tor Anderson have a valve that is ideal and calibrated pressure drop vs flow too.

As far as using excel to calculate the network, you are brave! With correct modeling it can be done it is a case of iterative application of pressure loss calculations for each branch given desired flows. ensure there is continuity of mass (volume) for each distribution network and it can be done. If losses without the valve are found then the pressure loss diferences between parallel branches represent the required "trim" pressure drop across the balancing valve.

Mark Hutton

 

[katmar]

It seems we have a split vote here. I agree with dcasto that a butterfly valve is exactly the right type of valve to use for cooling water. Cooling water lines are usually large and globe valves are expensive, bulky and heavy. For fine control consider using a gearbox rather than just a lever (the bigger butterfly valves come standard with gearboxes).

Depending on the number of branches you have, it could be tedious doing the simulation with Excel. But by no means impossible. The valve manufacturers will supply charts of K value (or Cv) vs % opening, so once you have done your calcs you will be able to get starting points for the settings on each valve and then do the fine tuning on site using local pressure gauges.


Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[HEC]

The butterfly valve is a poor choice especially where large pressure drop (K value) is required to balance a branch. A butterfly valve has a very non linear characteristic vs its angular position. For a small angle from full open there is little increase in pressure loss. As the disc goes beyond around 45 degrees, the pressure drop (K value) increases dramatically. Hence the valve can be hard to set. An additional issue is that the flow is diverted against the side of the pipe and will remain that way. This can lead to local erosion of the pipe. Hence the reason why it is not recomended to use a butterfly valve as a balancing valve.

Mark Hutton

 

[dcasto]

If you need to pinch a butterfly valve my more than 75% in water system to balance, then you have a big problem, so shut down a pump on the cooling tower. A well designed and balanced system should only need so gross adjustments of a few psi drop, not drops so high that there is errosional issues and cavitation.

 

[quark]

There is a very good post by Jim at thread408-168102

I agree with dcasto. If the system sizing is properly done, butterfly valves can be OK. However, if the user points are more and very close flow control is required then reversed return piping will complement the balancing action of the valve.

 

[UmeshMathur]

Regarding the method of calculating the hydraulic network, I have actually programmed the Hardy-Cross method, which is a "beginner's" method, not too accurate because of the restrictive assumptions. This can be implemented in Excel (very tedious even for small problems; besides the spreadsheet would be a unique solution to a given network - therefore of no generic value). There are, needless to say, much better commercial programs available at modest cost.

Your real problem, of course, will be finding a proper way to characterize the hydraulic resistance of partially open butterfly valves. The pressure drop changes dramatically over just a few additional degrees of opening when you are in the effective regulation range for this type of valve. Unless you spend umpteen hours "fine-tuning" the parameters, I wouldn't be too optimistic about the results, as far as matching field results, even if you have programmed correctly.