Ball Valve Operation 7

[bccwwes]

Recently I had an operator tell me that a ball valve will be forced open/closed by the force of the liquid flowing through it. (i.e. when the valve is nearly closed the force of the flow will "snap" it closed, and similar for opening.)

I have not heard of this. Any truth?

 

[zdas04]

I have seen throttled ball valves rattle open (not closed). It is not a snapping action, but more of an incremental thing. It has usually taken a couple of hours in strongly vibrating service for a ball valve to go from 5% open to 15% open (where the throttling stops) and then it will tend to stay there forever. This is one of the many reasons that I will not spec a ball valve for throttled service.

I have done the free-body diagram on the forces on a ball and while it is a bit indeterminate (positive feedback is hard to get into equilibrium), the forces to slam a ball in either direction don't seem to be available in a flow stream, even at high flow potential and very high dP.

David

 

[gerhardl]

As zdas04 I believe this to be very uncommon, (never heard of it) and that it would require several unlikely conditions to occur at the same time. This could be for instance unfortunate or unsolid construction, high wear or low natural torque, possible combination of outer forces as mentioned above, high pressure, high flow, special sizes, special fluids etc..

Like zdas04 I only exceptionally and for special 'ball valves' use this construction for throttling. V-port and double eccentric 'ball valves' are acceptable under certain and controlled conditions.

To avoid problems:

Change the ball valve to better constructed or valve of other type.

Combine with gear, if necessary with lockable gear.

 

[bccwwes]

Thanks for your input.

This is not actually a throttling application, but is a situation where rapid closure could cause a pressure surge problem.

The discussion was that the valve should be closed slowly and the operators thought that the flow would force the valve closed out of their control.

Based on your feedback I believe that their concern is likely unfounded. Thanks again.

 

[JLSeagull]

The ball valves that I use are pneumatically actuated. Properly sized actuators can open and close ball valves with no snapping type forces observed nearing the seat. On large valves I often observed jerky motion in mid travel on a shop floor with no process fluid applied inside the line.

 

[bcd]

The dynmaic forces from flow around the edge of the ball port can create a positive or high pressure upstream, and a lower pressure on the opposite side in the ball port as the valve rotates. So flow can and does assist or resists ball rotation depending upon the direction of rotation and flow velocity. Key point is that it assists a very small amount, consider it negligible, and it will never fully open or fully close the valve. Ball valves are inherently capable of rapid opening and closing. Good way to slow them down is to use a multi-turn gear operator. Electric actuators also tend to be slow operating. If using a handle or pneumatic or hydraulic actuator, they can be operated quickly and initiate serious water-hammer.

 

[gerhardl]

The best possible control against rapid closure is a mecanically secured and controlled closure.

Choice one: gear and/or electrical actuator with handwheel and/or UPS if critical.

Choice two: oil hydralic with throttling valve on return line from actuator (with or without, depending on size and industry/application mechanical fallweight for closing, or oil pressurized reservoir vessel)

Choice three: pneumatic opening and air or spring return wih throtteling valve on return air and/or backup pressurized reservoir vessel)

Ranging of alternatives may vary a bit depending on industry and application, backup or redundant solutions on how critical the application is.

 

[btrueblood]

We have a 4" full port ball valve in service (as an isolation valve, not a throttling valve) on a water flow test stand, and it has a tendency to rattle closed rather disconcertingly at flow rates of 50 to 250 gpm. We added a mechanical restraint to lock it in open or closed positions.

 

[zdas04]

btrueblood,
That is interesting. A fully opened ball valve looks to the flow like a chunk of pipe. The available forces are tiny even at those flow rates. There has to be something very specific going on for it to consistently rattle towards shut (the farther shut, the higher the dP, and the higher the forces that will normally tend to open the valve). Could there be something in the flow that is genertating von Karmann Streets? I'm not sure where I'm going with that question, but I've seen an amazing amount of energy included in those shed vortex pairs.

Is the topworks of the valve on the top or the side? I have seen the weight of a side-mounted operating lever try to cycle a ball valve in really low dP applications, but I've never seen one succeed past about 85% shut (at that point the dP increases rapidly and the ball shifts to the downstream seat and gets harder to shut).

David

 

[btrueblood]

Dave,

Agree with you regarding vorticity (see below), but it was disconcerting that the seal friction did not do a better job holding the valve in the open position. This was a first for me, and a heck of a surprise.

This is a 4" NPT, female threaded, brass full port valve; if memory serves McMaster-Carr # 47865K73. The valve is installed vertically, about 18" above a horizontal run and buttweld elbow. Within a day or so of commisioning the pump skid, we noticed the handle seemed "loose" in the open position, a week or two after that, it started to exhibit the behavior described earlier. A bit disconcerting, as the back pressure on the item under test would suddenly depart from test conditions.

The stem is perpendicular to the plane of the piping bend, thus a 2-vortex swirl as develops downstream of bends, would (I think) explain a non-uniform force or couple on the ball. The force is not huge, a fingertip on the handle is enough to restrain the ball - but there is zero friction on the ball from the seals until fully closed. The handle is oriented vertically up, so the weight of the handle is enough to keep the ball moving towards closed...

Our fix was (laugh if you want) an O-ring tying back the handle to force it to snap over-center to either the closed or open position. Quicker, easier, cheaper and more convenient (fit for purpose?) than a locking handle.

 

[JimCasey]

I agree with those who put this into the "unusual" category.
1. McMaster-Carr is a great source for many things, but a 4" brass threaded valve is not going to be of industrial quality. Unusual to see NPT connections larger than 2" just because it's hard to find room to swing a cheater bar long enough to tighten a connection that requires that much torque.
2. "Floating" ball valves load the downstream seal to shut-off IN PRINCIPLE but the manufacturer should assemble the valve with the seats in preload. The preload friction would mitigate any possibility of the valve jiggling to a new position. Dynamic effects are ignored because they are so small.

3. A ball valve typically requires more torque to close because the trailing edge of the waterway has to clear the lip of the seat, and differential pressure is exerted on the ball when it closes, also raising torque. Like pointed out earlier, when the valve is in the wide-open position, the flow only knows it went through a short piece of pipe with a better-than-normal surface finish.

If the ball does insist on rattling to a new position, using a crescent wrench to twist a little snug on the packing adjustment should add enough drag to keep it from wandering.

Butterfly valves do exhibit dynamic torque effects, and the dynamic effect peaks about 15 degrees from full-open.
The torque moves the disk toward the closed position. This is why it is unusual to see butterfly valves larger than 6" with levers. Gears become mandatory for safety reasons.

 

[btrueblood]

Certainly not a usual or expected mode of operation for a valve. Disparage the quality if you want, but to say it can't happen is a bit misleading. I've seen other, heavily-worn, commercial-quality valves have similar low torques, but not to this extent and in this size of valve.

"...using a crescent wrench to twist a little snug on the packing adjustment should add enough drag ..."

This is an o-ring stem, so there is no packing adjustment, just a nut to hold the lever in place. I actually prefer the method mentioned - it provides a fairly positive locking action to the two positions the valve needs to be in, and reminds operators that the damn thing is faulty and should be watched. A packing nut (itws), or a washer under the handle to add friction, would just be a point of maintenance easily ignored.

At a guess, I think the o-ring used for preload on the upstream seal may have been missing during assembly. But it's never going to be known, as I have no intention of hacksawing the darn thing open.

 

[JLSeagull]

If an operator told me that a ball valve is forced open/closed by the force of the liquid, I would believe that something like that occurred. Check that specific valve; perhaps have it repaired - or replaced if small.

 

[bccwwes]

The operator had not used the valve yet. He just felt that this would happen. I'm now pretty sure that it will not.

Thanks to all for your input.

 

[FMJalink]

bccwwes, your operator was probably speaking about floating ball valves. With these valves the differential pressure over the valve presses the ball on the downstream seat, which is different from trunnion ball valves.