Pressure drop over valves

[Drexl]

Hi,

I had two questions regarding pressure drop in valves, hopefully quite easy for you guys.

1. I'm performing a pressure drop calculation on various pipes, still in early design stage. Valve types are specified but no exact models, so I can't ask the manufacturer for Kv-values. I have used some "typical" Kv-values for the valves from some specific manufacturers, but the customer does not like this. Is there some norm that would supply Kv/zeta values per valve type/size that i could refer to instead? My old school books did not impress either.

2. I have a problem with possibly too high pressure drops in a system and asked the valve supplier what the pressure drop in his valve would be at the wanted flow. I got the answer that:

dP = rho * (Qs / Kv)^2

dP = Pressure drop [bar]
rho = Density [kg/dm3]
Qs = Flow [m3/h]
Kv = Supplied by manufacturer [..]

Now my question is, how can the pressure drop be the same with syrup and water? This equation must clearly be for water only, then how will i calculate pressure drop for other fluids with various viscosities?

 

[iken]

Hi Drexl,

1. You need to start somewhere, and the Client needs to appreciate that until the valves (make) have been confirmed you need to make some form of guess. You could always pick a valve make, stipulate this is the type to be used, or state all calculations based on make xxx. If alternative valves are used, re-calculations may/will be required.

2. Doesn;t the density take care of different fluids? Syrup and water wouldn't have the same density. Also, does the Kvs value relate to "fluid" or "water" passing through valve?

 

[KoachCSR]

1. In addition to iken's response, try taking a look at a Crane 410. Assuming fully turbulent flow, you can find K-factors for a large variety of valve and fitting types. A good starting point without any hard data.

2. I also agree with iken here; the density will resolve fluid differences.

 

[gerhardl]

Agree with both above answers.

Actually the answer is somewhat self evident if you go to the description and certifications for a certain valve's flow data, either zeta values, Kv or Cv:

The flow has to be physically measured within a given description, and valid only for the specific make and construction (and including normally one size up or down in same pressuere class, or some other similar limitation on the validity).

The reason is, as known, that the same type of valve, same size and pressure class, can have a fairly large spread in flow and pressure loss form one construction/manufacturer to another. The generalization has obviously becaause of this limitations. Better to take a certion well renowned valve brand, and state (with kv values - not necessarily brand name)) that valves bought should have this Kv (Zeta, Cv) or better.

For a complex pipeing system consisting of a number of different valves, pipes, bends, materials, walltypes and intruding instrument etc, the total pressure loss will anyway have a degree of uncertainity.

You will have to agree with the customer how you should approach the problem, and what the overall technical/commercial targets for the pipeline construction should be, including cost and quality, and upper/lower limits for accuracy for pressure loss.

It is not even certain that the most critical issue is the loss over the valves, what about bends, pipesize, pumps and material etc. etc.?

(Opposite sample:for a hydroelectrical plant the water turbine inlet valve loss will be highly significant!)

Anyway, as said by others: you will have to start somewhere.

Note: If you are held responsible for the total loss over the system, remember to agree exactly how this should be measuered (in detail) and by whoom, and be sure you have stated the accuracy limits wide enough, and that each supplier is sub-responsible for their calculations/values for each of their components.

 

[PaoloPemi]

the ISA equation (liquids) keeps in account several factors including the viscosity of fluids (the procedure calculatea a "pseudo" number of Reynolds based on the port diameter and a correction factor), once you have calculated the CV the difficult is to convert it in a proper valve size as different types of valves have different CV values, for example a rotary valve as butterfly or ball offers a maximum CV larger than a globe but the characteristic curve is very different, selecting the proper valve is not easy as you have to consider many different factors.

 

[JimCasey]

If the valves are control valves, there is another dimension in the consideration. a 4" (100mm) valve may have a Cv of 200(Kv 168), but that's only in the fully open case. Control valves are not ever supposed to be fully open, and typically operate at around half-stroke or 20% of capacity,(cv~40, Kv=~33) making a LOT more pressure drop than the unrestrictive isolation valves installed next to them.

Viscosity. It's a funny thing. There is very little effect from viscosity as long as the flow is turbulent. Then the standard Cv(Kv) equation works well. As pointed out earlier, syrup and water have different densities which are reflected in the Cv (Kv) calculation and account for the variation in results. At low Re (<40,000) the viscous effects become significant. In the laminar flow region there is a different equation to describe the relationship between the variables. Checking and correcting for viscous flow is tedious. Computerized sizing programs automatically evaluate and correct for viscous effects.

 

[stanier]

Suggest you download the Applied FLow Technology software :-

http://www.pumpsystemsmatter.org/content_detail.aspx?id=110

Then you can get access to the valve loss criteria from Crane, Miller & Idelchik when you select a valve node.

In support of AFT you may then choose to invest in the full blwn packages such as Fathom or Impulse for steady and unsteady state modelling software.