Unusual Diff. Pressure switch sought

A DP transmitter with a programmable device give you flexibility that you could well need with this sort of thing. You're likely to find once you try it that there are things like time delays or hysteresis settings that need to be there for operation to be reliable.

I've been around a lot of pumping installations over the past few decades and can't recall seeing anything like you describe above (it actually sounds a bit dangerous to me).

Pardon my ignorance, but why does the check valve need to be motor operated when a simple swing-check would do??

United Electric

make a huge range of diaphragm-style DP switches and we've specificied many of their J120K-series over the years on pumping applications.

I hope this helps.

Cameron:-)

Originally, I meant to have the valve open on high differential pressure (switch contact closes) and then close on low D.P. as the contact opened. The problem was, at low flows on these pumps, the D.P. across the valve would be so low the contact would open. Since this was conceived, I've learned that this valve, made by Golden-Anderson here in the U.S., will close like a conventional check valve if pressure on its outlet exceeds pressure on its inlet. As its closing, a (separate) high pressure switch between the pump's discharge and the check valve will cause the pump to trip.

The client does not trust a conventional mechanical check valve in this application (we'll be controlling pressure in the header downstream of four pumps, each with a variable speed drive.) The maintenance supervisor believes in this case, the pump will start, and do nothing for a bit, and then flow will come gushing out in a burst, making pressure regulation hard. These motorized check valves will open (and close) gradually.

Thanks for the feedback.

Iconoclast

Using pressure switches as you describe is a major hassle. Using the check valves themselves to put the pumps on line is by far the cleanest way to do this.

The standard way to address this combination of events is to design a "shutdown" mode into the pump sequence.

As the pump accelerates, it pushes the check open and follows the command from the pressure controller. When the pump is to stop, the VFD decelerates the pump to a speed where there is about 10% backflow through the pump. This backflow permits the check to close, and a limit switch detects the closure of the check valve and only then completely stops the pump. If the check valve doesn't close, the pump doesn't stop, and the 10% reverse flow prevents the pump from overheating. If the pump is again needed, it simply accelerates, the flow reverses and the pump picks up its load. Even though the check is stuck the system stays in control. If the valve were to stick and the pump were simply shut off, it would backspin, drawing large quantities of fluid out of the header and very possibly damaging the pump. Too, a backspinning pump is very difficult to restart. The shutdown scheme avoids all this grief.

If you have never done it before, don't underestimate the issue of sequencing and staging the pumps. Depending on the system curve, you may use almost none of the speed range of the VFDs. Also consider that, when pumping into a common header, pump speeds must be matched; running one pump at 100% and the next one at any speed less than that will result in no flow from the second pump! This is perhaps the origin of the maintenance supervisor's concern, some doubtful schemes try to run one pump at full speed and an additional one a bit slower to provide incremental additional flow. This makes the second pump go from zero to full flow with a very small change in speed, and is nearly impossible to control satisfactorily.

It's not necessarily that bad, although I agree that the hydraulics need to be checked. If there's enough hydraulic resistance in the individual discharge lines *before* they common then you can get away with a degree of speed difference, maybe quite a lot, and still retain flow sharing. There's an additional proviso there ... provided that your speed control is solid. This is usually the case with VVVF and similar modern drives, but some devices such as fluid couplings have nasty nonlinearities with load that can make load sharing very difficult to achieve.

...

I had the same questions when I read Iconoclast's post. It seems to me that the plant designers want to make motor-controlled (butterfly?) valves behave like check valves by adding instrumentation. As you say, that seems to be -- expense aside -- a possibly dangerous and certainly cumbersome approach.

...

Jerry

...

Then bring the pump up to speed gradually. There's no need to close a check valve slowly. It closes by itself when flow stops; that's completely shockless. A motorized valve that closes slowly must brake a backflow.

Jerry

Well, that's not true. With fixed speed pumps, the water-hammer from check-valve slam can be quite severe. It's a function of how fast the pump costs down and how far the valve disk has to travel from full-open to fully-shut. If the valve travel is very large (like a reverse-angled seat), the reverse flow can build up considerably before the valve reaches the seat. And depending on the piping configuration you can get hydraulic shocks of several hundred psi.

That is why VFD systems have the 'shutdown mode' that BFoelsch discussed. By lowering the speed only partially, the reverse flow doesn't build up to nearly the same flow rate and then when the valve seats there is much less water-hammer.

daestrom

daestrom

So what you have is commonly referred to as a 'stop-check' valve. When the stem is withdrawn, it acts like an ordinary check valve. However, the stem can push/hold the disk in the shut position when run in.

Using D.P would be difficult in this sort of thing because of the VFD. Take the case of one pump running at some speed supplying the fluid flow. The D.P. across the full open check valve would be rather low. Now, you want to start the second pump and place it in service. If you run it up to the same speed as the running unit, the D.P across the shut stop-check will depend on how much flow the running pump is supplying. If is supplying very little, then the discharge pressure of the two pumps will be very nearly equal and the DP across the starting unit will also be very low. If the flow of the running pump is high, then there will be a difference in pressure between the already running and starting pump. But exactly how much difference in pressure depends on the pump curves. If the pump curves are fairly 'flat', then there won't be much difference, if they are 'steep', then there will be a difference.

So your use of DP will have reliability issues, it may open the valve sometimes (enough difference in discharge pressure caused by flow on a 'steep' pump curve), or it may not.

You don't go into a lot of details about these pumps. I *assume* they are the centrifugal type. If the pump motor starts, how likely is a mechanical failure going to prevent the pump from developing head? Most designs don't consider this sort of failure to be very likely. Many set ups that I've seen will simply tie the valve operator to the pump start/stop logic. For starting, first start the pump and after a short time delay open the valve. For stopping, first shut the valve and when it reaches fully-shut, stop the pump. This sort of logic is very easy compared to what you're proposing.

daestrom

About 10 years ago, I did the pump controls for a central sewage pump station that used a scheme more like the G-A check valves. The pumps were 5 x 200 HP, rated for 16MGD with VFD control and hydraulically actuated rotary check valves on the 30" discharge lines. The pumps started on deadhead, and once the pump speed was high enough the discharge pressure closed a _gauge_ pressure switch which triggered the check valve to open. On shutdown, one relay dropped out and triggered the valve to close,while a second relay kept the pump running until it was unlatched by the valve closed limit switch.

That worked, until the day the wetwell level reached an extreme high and caused at least one check valve to open with its pump still stopped. (No interlock between my pump PLC and the valve supplier's PLC.) The engineering firm wrote us a change order to replace the switches with transmitters wired to my PLC and add logic to open the check valve only when the pump was running and the pump discharge pressure to wetwell level was above a minimum.

I agree about the sequencing and staging. The sequence changed only when just one pump was running, and multiple pumps ran at the same speed.

Mike