3 way valve 3

[flinana]

Can anyone tell me how the DP across the two ports of a 3 way valve affects its control.
This 3 way valve either lets glycerine flow through an exchanger or bypasses it. The dp designed in direction of the exchanger is 0,5bar and towards the byapss its 0,3bar.

We are trying to maintain a constant outlet temperature from the exchanger. The valve will divert feed around the exchanger to control this. Presumably there is a constant feed through the other side of the exchanger. The amount diverted will depend on the setpoint of the controller which will regulate within the possible limits of temperature.
I can only think that the DP is there for the actuator sizing. The DP across a valve will depend on the flow rate so I'm not sure what basis the DP figures have. These DP figures should not have anything to do with the control as the control is only based on maintaining the temperature.
Having said all this my concern is that the flow will always tend to go where there is less DP, so the valve will either fuly close or fully open and not regulate smoothly.

thanks,

 

[cliff6361]

Are you talking about a 3way ball valve? If so this type valve is NOT a control device. The valve is designed to change flow direction only. The valve port through the ball is the same a 90 degree elbow in a pipe system. This type valve should be stroked fully i.e. full 90 degree rotation to switch flow direction. This valve is not for modulating service.

 

[JimCasey]

Respectully disagree with cliff. 3-way ball valves are not universally wonderful, but no choice is. However they can be useful in temp-control bypass applications such as this.

Regardless of the type of valve you choose, you have the choice on a bypass temp control application to use the valve for diverting (installed upstream of the HX, or mixing (installed downstream of the HX.

The leg that goes thru the HX has more pressure drop due to the toruous path thru the fins, coils, fittings, etc.
It is frequently useful to install a balancing valve in the bypass leg, throttled down so that causes the same resistance as the heat exchanger. Usually the balancing valve is manual, tweaked until the system stabilizes, and then locked in position.

Another pitfall if temperature control is if the desired temperature is really close to either the outlet temperature of the Heat Exchanger or the bypass temperature. if that is the case you wind up controlling at one of the extremes of the valve's span and like any control loop that gives really unstable results.

So if, for example, the fluid coming down the pipe is at 50 degrees. and the fluid coming out of the HX is at 150 degrees, and you want to control at 100 degrees, a three-way valve will work pretty well.

If you try to control the same system so its outlet temp setpoint is at 148 degrees, then you just have a trickle coming thru the bypass and a tiny control signal change gives you a drastic change in outlet temp. In this case it would be beneficial to install a relatively large valve on the 150 degree leg, a much smaller valve on the bypass leg which operates in the reverse direction of the other valve, and operate them with the same control signal.

Back to the ball valves: A 90-degree 3-way ball valve has linear characteristic for the middle 60% of its stroke. With approprite selection of the other components in its system(valve size, actuator, positioner, zero-lash coupling) it can do a pretty good job. 3-way globe valves are generally just two unbalanced valve plugs in a common body, contoured to be linear throughout their stroke. So on the sucky control scale they are only marginally better than possible with a ball valve, harder to plumb, bigger for a given pipe size, and massively more expensive. Big advantage with a globe valve is that it is more likely to be able to tolerate higher temperatures, but unlikely to shut off either port completely if desired. (Metal seats) Ball valves (usually) have polymeric seats.

 

[flinana]

Thanks very much Jim, you sound very convincing, I was wondering if you could do me the favour of having a look at the valve spec which I attach, as well as the P&ID portion of the control and tell me what you think.
This has become a discussion with our basic engineering contractor.

Cheers,

http://files.engineering.com/getfil...47e-2870c6d848c2&file=3680-10-050-2_Rev.3.pdf

http://files.engineering.com/getfil...d9-7cbcd65ffdd0&file=3_way_control_valves.xls

 

[JLSeagull]

The valve described is a three-way globe valve. Control valve manufacturers sell three-way valves for diverting flow and for combining flow. The notes cover this.

Refer to ASME Class 150 and flanges per B16.5 and you can omit the finish requirement. The leakage specification B16.104 is obsolete; use FCI 70-2. If a seat tightness standard is used it needs to state the tightness class. Also, three-way globe valves lack tight shutoff and likely meet Class II.

I would provide more detail regarding the accessories such as the I/P - positioner especially if you want HART. I would provide more details about the air-pressure regulators for the accessories, too. However the vendor can specify the details with her proposal.

 

[JimCasey]

JLS, as always, expresses some good points.

I ran the sizing and you need a 6" valve if it's a globe valve.

I brought up shutoff: Shutoff is something that has questionable meaning in a 3-way valve because if you are asking for one port to shut off completely, the other port is WIDE OPEN. So you need to decide if you REALLY need to isolate one of the flows with the control valve. Otherwise your butterfly isolation valves 4192, 4194, 4193 perform that function. Also 4194 performs as the balancing valve I described.

I mentioned that the trim in a 3-way globe valve is unbalanced. With the 7 bar shutoff specified the actuator will have to supply over 2000 lbf (or about 1000 Kgf). Line 13 specifies the valve fails A-C, so the actuator needs a spring that supplies a ton of force, extended.

So the engineers have specified a 304L, 3-way valve. Everybody can GET 304L, but nobody I know offers it as a standard material. 316 (CF8M) is a standard material.
So you're looking at the price of a nice SUV for a single valve, and you will get to wait a casting cycle for delivery (18-20 weeks).

If it were MY plant, I would put a tee where TV004.23 is on the drawing, at the outlets of the tee I would use 2 eccentric rotary plug (ERP) valves (Valtek Maxflo3, Fisher CV500, Leslie K-Max) The valve in line with 4193 would be fail-open, and the valve in line with 4194 would be fail closed. Unless there is some really persuasive and not obvious reason, I would consider 316 for the body. 304L is really nice at resisting carbide precipitation in heat-affected zones around welds, but you are asking for a flanged valve.

ERP valves have great stability, unsurpassed turndown, and greater capacity per size. You could use 4" valves for this application. for TWO ERP valves(complete with actuators and electropneumatic positioners) , the cost would be about 1/3 as much as ONE 3-way globe and you can get the valves in 6 weeks or less. Run the same signal through both positioners and the assembly will act as a 3-way valve, with one opening as the other closes.

 

[flinana]

Absolutely fabulous. Thank u all.

Greetings from Spain.
For those interested this is for a pharmaceutical glycerine plant part of a Biodiesel plant in the south of Spain, next to Gibraltar.

 

[hydtools]

Use a thermostatic valve like one from Amot.

http://www.amot.com/pdfs/thermo.pdf

http://www.amot.com/us/products/index-temperature control valves.asp

Ted