Pump Control Valve vs. VFD

[hibbonite]

I am working with a local municipality to replace their aging water tower booster pumps. The big question is whether to replace the pumps and Cla-Val booster pump control valves, or just put all the pumps on VFD. We're worried about water hammer and I don't see how the VFD wouldn't work just fine. Any experience or thoughts on this would be appreciated.

thanks.

 

[BigInch]

When the electricity goes down and your pump shuts off and you've got a big water hammer on the way, what's your VFD going to do?

**********************
"Being GREEN isn't easy"
.... Kermit

http://www.youtube.com/watch?v=hpiIWMWWVco

http://virtualpipeline.spaces.live.com/

 

[whammett]

BigInch beat me to the punch. Unless you have a UPS system tied into the VFD, your much better off going with a FCV. There is a lot to be said for being able to set a valve to fail open, closed, or stationary.

"Scientists dream about doing great things. Engineers do them." -James Michener

 

[BigInch]

I take issue with that VFD controlling water hammer strategy anyway. True waterhammer is too fast of a phenomenom for a VFD and pump to control. The pressure wave hits the pump before the pressure sensor even figures out it isn't working anymore. In water, transient pressures travel at near 3000 fps. If your PT is 1/2 mile away, you've got roughly 1/2 second to adjust the pump and its going to take another 1/2 second or so for the new discharge pressure to affect flow in the line between the pump and the PT, and that's if your pump has a low rotating inertia and can change speed that fast. By that time, the transient wave has already traveled all the way to the pump. You might be able to do something about the second wave, but that's only about 1/2 the amplitude of the initial wave's pressure. Any damage that was going to be done, has already been done.

**********************
"Being GREEN isn't easy"
.... Kermit

http://www.youtube.com/watch?v=hpiIWMWWVco

http://virtualpipeline.spaces.live.com/

 

[SteveWag]

In Municipal Water System pump design/selection, very little attention is placed on analyzing transients. The common design is to ramp up/down the flow slowly so as to keep from causing pressure disturbances in the system. The “Cla-Val” and other hydraulic valves allow a Municipal Engineer to do this “safely”. Now that VFD’s are common-place, some salesmen are suggesting that they be used to do the same thing. As “BigInch” has pointed out, this doesn’t hold-up when the power fails. The big problem with the control valve is that, typically, the operator does not understand the function/need for the valve, nor do they understand how the valve works. (They don’t understand how the VFD works either, buts that’s the electrical guy’s problem.) When VFD’s are used, a non-slam check is typically used. The VFD’s are used as soft-start devices only. I tend to use the control valves, but am bucking the current trend.
Steve

 

[hibbonite]

I forgot to mention that we are also considering installing a Cla-Val surge anticipator valve in the main header to deal with the water hammer issue.

One argument I've heard is that the Cla-Val pump control valve allows the pump to start against zero pressure, since the valve is still closed when the pump starts. This is true, as I've seen the valve and pump do exactly that.

Shouldn't the VFD be able to start the pump against the pressure of the main header?

Even a soft start can be ramped up and down (up to 4 minutes). I would think this would be gradual enough.

A soft start with a non-slam check valve seems like it should do the job just fine and, be much cheaper than a Cla-Val.

Thoughts?

 

[BigInch]

You can start with a VFD, but you will have to ramp it up pretty high to get enough head to finally overcome static head and begin flowing into the tower, so I doubt it will really do much good there. Buy a soft start box.

A non-slam check won't protect the pump against water hammer if closes too slowly, or if it breaks. If it works, properly, it can protect the pump, but then it wouldn't do anything to lower the transient pressure that it itself creates in the pipe when it closes. A fast-acting check can cause a water hammer more easily than a slowly operated ball valve. Flow through a Cla is inversely modulated by its differential pressure, hence can control transient pressure changes when it is properly adjusted.

Pumping to a tower, I would think you'd not have so much of a problem starting (unless you're overpowered) and you'd be more likely to get a water hammer when the pump trips off and the flow in the line is subjected to a quick reverse by the action of gravity.

**********************
"Being GREEN isn't easy"
.... Kermit

http://www.youtube.com/watch?v=hpiIWMWWVco

http://virtualpipeline.spaces.live.com/

 

[hibbonite]

After studying my pump curves, here are my latest thoughts.

When the pump starts, even if it is slowly ramped up by a vfd or soft sart, there will be a point where the pressure will exceed the main header pressure and the check valve will open. At that time, the pump will be trying to suddenly push a given amount of water into the main. According to my pump curve, it will try to push 2500 gpm into the main at this pressure. I'm thinking this would be a water hammer issue.

Likewise, 2500 gpm will be suddenly not allowed into the main when the pressure drops below the main header pressure and the check valve closes. Another water hammer issue.

Hence the need for the Cla-Val, which gradually increases the flow into the header. The VFD is not the answer in the situation.

Thanks all for your comments, they really helped me think this through.

 

[BigInch]

That's exactly as I see it. Whether the 2500 gpm is actually enough to cause waterhammer in your pipeline or not depends on its velocity going into the initial linepack. It could be troublesome, as that linepack may be rather "solid" under its static pressure, thus any fluid entering anew would be able to rapidly increase pressure with the addition of relatively little volume. Got a transient analysis program available?

**********************
"Being GREEN isn't easy" ..Kermit the

http://www.youtube.com/watch?v=hpiIWMWWVco

http://virtualpipeline.spaces.live.com/

 

[Yellowtetra]

First, you really need to perfrom the required transient anaylsis or get someone qualified to do it.

Second, VFDs will generally work fine. They don't have to be fast, they are cheaper than flow control valves, easier to install, and easier to maintain. You also get the advantage of running the pump at it's most efficient point unlike the valves. Soft starters are generally even a better deal but they can't vary the flow like a VFD.

Third, VFDs, soft starters, and flow control valves cannot help reduce tranisents during power outage. All depend on the pump to keep the water moving so the device can safely slow down the water.

You need to know how bad the pressure spikes will be if the pump sees an unprotected power failure. The resulting pressures might not damage anything or justify a surge anticipator.

Fourth, watch out for surge anticipator valves. They can cause low pressure spikes to get worse, which is often the major consequnce in a net worked distribution system. Customers will complain about no or low pressure water. The high pressure spike often never happens in a gridded system. Use a standard pressure relief valve instead if anything is necessary.

My recommendation is softstarter with quick closing check valve, but it is your responsbility to determine what's best in your situation.

Also think about how often power failures will happen.

Dan Barr, PE
Burgess and Niple