I am designing a quarter turn ball control valve. I have system inputs for pressure ranging from 350-750 psi and a flowrate of 52 gpm. It has also been given that due to another valve downstream, the backpressure at my valve's outlet will be 150 psi. With these values it seems I automatically have a pressure drop for the valve. What I am not understanding is, how can the outlet pressure of my valve be forced to be 150 psi when based on the valves internal geometry, is likely to be something else? In other words how can I just assign a pressure drop to my valve given inlet and backpressure? Thanks for any and all input!
Eng-Tips is the largest engineering community on the Internet
Intelligent Work Forums for Engineering Professionals
-
Congratulations waross on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!
Valve Back Pressure
- Thread starter nctexan22
- Start date
If the downstream valve modulates to hold a constant pressure it would correspond to your specifications.
If the downstream valve is in a constant position, then 150 psi would occur at one set of service conditions. Other properties of the fluid staying unchanged, your pressure would vary with the square of the flowrate.
ManyMany valve data sheets are filled out by lazy or unthinking engineers who specify, for example, 100 psi upstream and 50 psi downstream at three widely varied flowrates.
If the downstream valve is in a constant position, then 150 psi would occur at one set of service conditions. Other properties of the fluid staying unchanged, your pressure would vary with the square of the flowrate.
ManyMany valve data sheets are filled out by lazy or unthinking engineers who specify, for example, 100 psi upstream and 50 psi downstream at three widely varied flowrates.
not sure to understand the question (since you are designing a valve I would suppose your are familiar with cv / flows etc.), however in 1950 Masoneilan introduced the cv defined as
Q = cv * Square Root(dp/sg)
Q gallons per minute (gpm),
dp in psi
sg specific gravity
so at costant cv (which corresponds to a specified position as said by JimCasey) there is a relation between pressure drop and flow and your problem would be that to estimate (or measure) the curve cv/aperture for your valve.
Q = cv * Square Root(dp/sg)
Q gallons per minute (gpm),
dp in psi
sg specific gravity
so at costant cv (which corresponds to a specified position as said by JimCasey) there is a relation between pressure drop and flow and your problem would be that to estimate (or measure) the curve cv/aperture for your valve.
- Thread starter
- #4
Jim,
It seems that I have interpreted the data incorectly. My valve must be capable of inducing a maximum pressure drop of 600 psi (750 in vs. 150 out), and a minimum delta p of 200 psi. So my challenge is to determine at what positions I can maintain a 150 psi pressure at the outlet, all while passing at least 19 gpm. Did you intend to state that pressure drop varies as the square of the velocity? I guess they are essentially the same thing, only the flow being with respect to area.
Paolo,
I am familiar with the flow coeficient, however the methods of determining a Cv based on the resistance coefficient K (as in Crane 410), do not apply to my valve. The seat trims' internal geometry consists of multiple orifices that are only exposed to full flow at certain angles of opening. I can determine what my Cv NEEDS to be based on my given flow rate and required pressure drop, but my pressure drop is my unknown. This is why I've turned to CFD software to hopefully determine my delta p and in turn develop my Cv for differing angles of opening.
It seems that I have interpreted the data incorectly. My valve must be capable of inducing a maximum pressure drop of 600 psi (750 in vs. 150 out), and a minimum delta p of 200 psi. So my challenge is to determine at what positions I can maintain a 150 psi pressure at the outlet, all while passing at least 19 gpm. Did you intend to state that pressure drop varies as the square of the velocity? I guess they are essentially the same thing, only the flow being with respect to area.
Paolo,
I am familiar with the flow coeficient, however the methods of determining a Cv based on the resistance coefficient K (as in Crane 410), do not apply to my valve. The seat trims' internal geometry consists of multiple orifices that are only exposed to full flow at certain angles of opening. I can determine what my Cv NEEDS to be based on my given flow rate and required pressure drop, but my pressure drop is my unknown. This is why I've turned to CFD software to hopefully determine my delta p and in turn develop my Cv for differing angles of opening.
When I perform a valve sizing (assuming that the medium is water or std. oil) I see that at dp=600 psi the valve is choking/cavitating. During choked flow the downstream pressure is no longer determined by the valve. This might explain why your downstream valve can control the intermediate pressure.
Regards,
Terje
Regards,
Terje
With 750 PSI in and 150 PSI out, (Please specify Absolute unit or gage units) This valve will have some form of critical pressure drop. Probably caviation, possibly flashing. Choked flow almost guaranteed.
Since the flow is a liquid, you'd better know the critical pressure, vapor pressure, temperature, viscosity, density, and the Fl of your trim should be very very close to 1.0. 1.0 would occur if the valve trim exhibits pure frictional adiabatic flow and would correspond to the flow through something like a cigarette filter, where energy is lost through shear and there is no acceleration of the fluid.
More than likely you will show cavitation at this kind of pressure drop. It is also likely that you are designing the tortuous trim to avoid cavitation.
The Masoneilan equation listed above will not apply as it does not address critical drops. Use ISA equations as appropriate.(ISA Handbook of Control Valves). Les Driskell wrote many books with regard to valve analytics and theory. Search Amazon and Stock up your library.
You did not specify which liquid you are designing for. Chemically pure fluids cavitate with damage likely. Water is probably worst since it is dense, has a lot of surface tension, and releases a lot of energy as it condenses from a vapor back to a liquid at very precisely defined conditions. Hydrocarbons are usually a mixture of chemicals with similar but not identical vapor pressures, so when the cavitation bubbles form and collapse they tend to cushion each other. Damage is thus minimized.
One last thing: It is not impossible but it is difficult to design this sort of trim into a ball valve. Globe anticaviation valves are a pretty mature concept and available from almost every major control valve company. Only you or your superiors know your situation and whether is is justifiable to re-invent the wheel with this product.
Since the flow is a liquid, you'd better know the critical pressure, vapor pressure, temperature, viscosity, density, and the Fl of your trim should be very very close to 1.0. 1.0 would occur if the valve trim exhibits pure frictional adiabatic flow and would correspond to the flow through something like a cigarette filter, where energy is lost through shear and there is no acceleration of the fluid.
More than likely you will show cavitation at this kind of pressure drop. It is also likely that you are designing the tortuous trim to avoid cavitation.
The Masoneilan equation listed above will not apply as it does not address critical drops. Use ISA equations as appropriate.(ISA Handbook of Control Valves). Les Driskell wrote many books with regard to valve analytics and theory. Search Amazon and Stock up your library.
You did not specify which liquid you are designing for. Chemically pure fluids cavitate with damage likely. Water is probably worst since it is dense, has a lot of surface tension, and releases a lot of energy as it condenses from a vapor back to a liquid at very precisely defined conditions. Hydrocarbons are usually a mixture of chemicals with similar but not identical vapor pressures, so when the cavitation bubbles form and collapse they tend to cushion each other. Damage is thus minimized.
One last thing: It is not impossible but it is difficult to design this sort of trim into a ball valve. Globe anticaviation valves are a pretty mature concept and available from almost every major control valve company. Only you or your superiors know your situation and whether is is justifiable to re-invent the wheel with this product.
- Thread starter
- #7
Jim,
These pressures are specified as gauge, and the fluid is water.
I have realized the difficulty of this design. My superior has patented the trim in the past for noise control but it has never been produced to provide specific and "made to order" Cv's. Now that the design has been "sold" I'm responsible for solving the Cv puzzle.
I'm also not too sure how the ISA equations could relate to my situation. This is mostly because I don't believe I could qualify my valve into either of the two ball categories (full bore/segmented) to obtain factors required to obtain a flow coefficient, again due to it's trim geometry. Please correct me if I am off and thank you for your responses.
These pressures are specified as gauge, and the fluid is water.
I have realized the difficulty of this design. My superior has patented the trim in the past for noise control but it has never been produced to provide specific and "made to order" Cv's. Now that the design has been "sold" I'm responsible for solving the Cv puzzle.
I'm also not too sure how the ISA equations could relate to my situation. This is mostly because I don't believe I could qualify my valve into either of the two ball categories (full bore/segmented) to obtain factors required to obtain a flow coefficient, again due to it's trim geometry. Please correct me if I am off and thank you for your responses.
nctexan22: You need to test your valve to determine the flow coefficient, Cv and pressure recovery factor, FL at different travel positions.
These cannot be calculated or even approximated for a multistage 'low-noise' trim ball valve.
Check the Neles Q ball or Fisher V-ball with insert to compare similar type valves.
You are wasting your time speculating on inlet and outlet pressures etc without knowing what your valve Cv is for different openings.
terje61: please note that downstream pressure is never determined by a valve but solely by the system hydraulics. The valve is simply a variable restriction in the overall system.
These cannot be calculated or even approximated for a multistage 'low-noise' trim ball valve.
Check the Neles Q ball or Fisher V-ball with insert to compare similar type valves.
You are wasting your time speculating on inlet and outlet pressures etc without knowing what your valve Cv is for different openings.
terje61: please note that downstream pressure is never determined by a valve but solely by the system hydraulics. The valve is simply a variable restriction in the overall system.
- Thread starter
- #9
Scotsinst,
Thanks for your input, this is what I've always speculated and realized after pouring over formulas and methods in Crane 410 and ISA. It’s ironic you mention the Q-ball because the valve was sold by assuming a Cv curve similar to that obtained by a Neles Q-ball of the same NPS.
However, after deeper scrutiny I believe a better relation would be achieved in assuming my valve as having two of their "A-plates", one situated in both upstream and downstream positions.
This is because my valve actually has no ball trim, as the Q ball valve does.
I still have one main area of confusion with my design requirements: pressure. You mentioned that "downstream pressure is never determined by a valve but solely by the system hydraulics. The valve is simply a variable restriction in the overall system."
The way I see it, for a given flow rate, or velocity, at x angle of opening, there will be y pressure drop across the valve. As the ball is closed and a smaller opening is available for the fluid to flow through, to maintain a particular flow rate, the pressure must be increased on the upstream side.
Would the pressure drop not always be the same, regardless of upstream pressure, at that angle of opening?
In other words, and relating back to my initial post, how would system hydraulics affect (and thus determine) the downstream pressure if the valve’s angle of opening will only produce a certain pressure drop for a certain flow rate? How is this accomplished by the system?
My valve must pass a q= 19 gpm (not the 52 gpm as previously given), with input pressure available from 350 psig to 750 psig, and maintain an outlet pressure of 150 psig.
I have modeled the fluid into CFD software, using the given flow rate to determine a velocity for the inlet based on its diameter. The software has given me a magnitude of pressure drop across the valve of 163.5 psi, at my smallest angle of opening. If downstream pressure must be maintained at 150 psig, then the CFD would imply that the upstream pressure could be no more than 313.5 psig at 19 gpm. This would tell me that I must modify the valve geometry to produce a larger pressure drop because minimum inlet pressure is 350 psig. Would you agree with my thinking/methodology?
Thanks again for any input!
Thanks for your input, this is what I've always speculated and realized after pouring over formulas and methods in Crane 410 and ISA. It’s ironic you mention the Q-ball because the valve was sold by assuming a Cv curve similar to that obtained by a Neles Q-ball of the same NPS.
However, after deeper scrutiny I believe a better relation would be achieved in assuming my valve as having two of their "A-plates", one situated in both upstream and downstream positions.
This is because my valve actually has no ball trim, as the Q ball valve does.
I still have one main area of confusion with my design requirements: pressure. You mentioned that "downstream pressure is never determined by a valve but solely by the system hydraulics. The valve is simply a variable restriction in the overall system."
The way I see it, for a given flow rate, or velocity, at x angle of opening, there will be y pressure drop across the valve. As the ball is closed and a smaller opening is available for the fluid to flow through, to maintain a particular flow rate, the pressure must be increased on the upstream side.
Would the pressure drop not always be the same, regardless of upstream pressure, at that angle of opening?
In other words, and relating back to my initial post, how would system hydraulics affect (and thus determine) the downstream pressure if the valve’s angle of opening will only produce a certain pressure drop for a certain flow rate? How is this accomplished by the system?
My valve must pass a q= 19 gpm (not the 52 gpm as previously given), with input pressure available from 350 psig to 750 psig, and maintain an outlet pressure of 150 psig.
I have modeled the fluid into CFD software, using the given flow rate to determine a velocity for the inlet based on its diameter. The software has given me a magnitude of pressure drop across the valve of 163.5 psi, at my smallest angle of opening. If downstream pressure must be maintained at 150 psig, then the CFD would imply that the upstream pressure could be no more than 313.5 psig at 19 gpm. This would tell me that I must modify the valve geometry to produce a larger pressure drop because minimum inlet pressure is 350 psig. Would you agree with my thinking/methodology?
Thanks again for any input!
nctexan22,
in normal conditions (not limitated as in critical/cavitating/flashing cases) there is a relation between flow and pressure drop and providing you get a curve CV/Aperture it shouldn't be difficult to calculate these values. However when there is a change of phase, flow becomes critical , there is cavitation etc. etc. things are difficult to evaluate and really you should program a series of specific tests for the different working cases, then you can attempt to correlate the results with some form of generic correlation (for example ISA formulation) by finding the paramters (FL etc.) which minimize the errors. As far as I know also manufacturers with large experience follow this way.
in normal conditions (not limitated as in critical/cavitating/flashing cases) there is a relation between flow and pressure drop and providing you get a curve CV/Aperture it shouldn't be difficult to calculate these values. However when there is a change of phase, flow becomes critical , there is cavitation etc. etc. things are difficult to evaluate and really you should program a series of specific tests for the different working cases, then you can attempt to correlate the results with some form of generic correlation (for example ISA formulation) by finding the paramters (FL etc.) which minimize the errors. As far as I know also manufacturers with large experience follow this way.
- Status
