PRV Directly After Booster Pump..why? 2

[ziptron]

Hi Everyone,

So I have a system where a building is boosting the domestic cold water from the incoming city pressure up to the 30th level. For this we have some booster pumps (standby and lead) which boost the pressure so that at level 30 the pressure is around 60 PSI. To get 60 PSI on level 30 I need approximately 200PSI at the booster pump level. 200 PSI at the booster pump level is what I have however this 200 PSI is not created by the booster pumps. Directly after the pumps (1 ft away) I have a PRV valve. So, my booster pumps pressurize the system up to some pressure (no gauge so I do not know what) and the PRV valve lowers this back down to 200 PSI.

I'm curious of why they bothered installing a pump that is too big, and then just reducing its pressure. The reason I ask is that this valve keeps failing and I keep on having to replace it because with it failing I am getting pressures that are too high throughout the building. Would it not just make more sense to install a pump with the same flow capacity but lower head capacity? I would then not need to worry about the PRV valve failing so often.

Any input is appreciated, thanks!

 

[btrueblood]

Yes.

 

[btrueblood]

Alternatively, run the booster pump on a VFD, and have the VFD control the pump speed and thus output pressure, via feedback from a pressure sensor on the 30th floor.

 

[JJPellin]

Ask yourself what it is you require of this pump. And then ask yourself what the pump is capable of providing. I am not familiar with this sort of domestic water system, but I have other systems that may be comparable. Please correct my interpretation where it is wrong.

You need a widely varied flow rate to account for water usage at different times. In the middle of the night, the water usage may drop to zero. At half time of the Superbowl, the water usage may spike incredibly high. (You didn't say if this was residential or business) In order to provide consistent water pressures for the end users, you need a constant (or nearly constant) water supply pressure over the full range of flow rates.

The pump is (I assume) centrifugal. The curve for a centrifugal pump provides a range of pressure depending on the flow rate. As the flow rate increases, the pressure drops. As the flow rate decreases, the pressure rises.

In order to get this pump to meet that need (variable flows at constant pressure), you have to do something. You can design the pump to produce a slightly higher pressure than you want at the highest flow rate that you ever expect and then control that pressure using a PRV. That is what we normally do. A VFD may not work well. As you change the speed of a centrifugal pump, the head changes even at a constant flow. And you want a constant pressure. If the problem is as I have described it, I don't think a VFD would work very well. We usually use VFD's for applications where we want a variable pressure, not for a constant pressure application.

You could address the problem with a controlled spill-back system to spill back the excess flow so you can run the pump at constant flow even when you have variable demand. But, this wastes a lot of energy. You could use a fixed head tank (like a city water tower) and cycle the pump to maintain level in the head tank. But, you may not want to build a tank 140 feet above the highest water user. You could use a positive displacement pump with a variable speed drive. This would be the most energy efficient option that occurs to me.

Without telling us more about your pump and your system, you are asking us to guess and speculate. So, that is what I have done. With more information, I could be more helpful.


Johnny Pellin

 

[zeusfaber]

This post is more about getting into the mind of the original designer than about offering a solution (Johnny does that better than I could)

This sounds like the sort of thing we see in the boost/servo circuits of some hydraulic powerpacks, where a requirement for a constant pressure but substantial variation in required flow is satisfied by a small positive displacement pump and a downstream PRV.

The thinking in the hydraulic example is that the PD pump is a constant flow device, not a constant pressure one. When deadheaded, it will produce silly pressures, so something (and the PRV is about as simple as it gets) has to be provided to control the pressure. The downside is that once you start to scale this simple system up, you lose a lot of energy in the PRV.

A.

 

[ziptron]

Thanks everyone for your replies. This is a residential system and yes, I do realize that over sizing the pump and putting in a PRV allows you maintain very specific pressure, I suppose I would have expected that to be done with VFD's though perhaps you do get better pressure control with PRV.

 

[JJPellin]

A VFD on a centrifugal pump set to hold constant pressure tends to be very touchy, especially at low flow where the curve may be rather flat. I would offer an example. In the attached curve, I show the curve for my pump running at full speed (3600 rpm) and a second curve with the same pump running at 3220 rpm. If the constant head demand in my system was 250 feet, the pump would run from 1040 gpm down to zero flow over this speed range. The VFD might control pretty well between about 1000 gpm and 600 gpm. Below 600 gpm, a change of just 60 rpm will drop the flow by 200 gpm. Since 60 rpm is only about 1.6% of full speed, this gets very difficult to control. It could be done. But, I would want to consult with a VFD expert to understand the control characteristics over these speed ranges.

Johnny Pellin

 

[77JQX]

Something else to consider - repairing a failed PRV is pretty easy, and we can get one of our apprentice mechanics to rebuild one in an afternoon, and we have the parts in our warehouse. A failed VFD generally takes greater skill and time to repair.

 

[MintJulep]

As noted above, a PRV may be the correct choice for this application.

However that does not mean that you have the correct PRV installed.

Rather than constantly repairing or replacing the same valve, take some time to understand why this valve is failing and look for a valve that would not be subject to the same failure mode.

 

[BrianPetersen]

The head-tank need not be 140 feet above the highest level if it is a pressure tank. If you use a pressure tank, it can be placed anywhere, although obviously the pressure required to be inside it will vary depending on where it is.

I grew up in a rural area, with water coming from a well. Water for the house came from a pressure tank (compressed air plus internal bladder) in the basement. The pressure tank was fed from the pump deep down in the well, which contained a check valve. The pump cycled on and off based on a pressure switch. Nothing exotic, nothing hard to understand, nothing that is wasting energy when there is low/no demand.

 

[LittleInch]

If the PRV (I assume this is Pressure Regulating Valve?, but can also be Pressure Relief valve) is failing a lot it could be that your pump is simply too big to cope with the variable demand. You often see multiple pumps which can come on line in parallel as pressure falls to cope with the huge variance in flowrates from a domestic building water supply. Installing a single pump with a standby may easily have been cheaper than say 3 or 4 smaller units, but you are now seeing the downside whereby it looks like your valve is constantly throttling at low flows and becoming damaged in the process.

I think some sort of buffer pressurised tank would be a good idea or if there is one a bigger one or retrofit some smaller pumps to reduce the requirement to throttle the outlet so much.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[zeusfaber]

Interesting point. I had assumed the PRV was a Relief (which is certainly the way it works in the hydraulic example I mentioned) rather than a Reducer.

A.

 

[LittleInch]

The phrase"lowers this back down to 200psi" makes me think this is regulating valve... That's why you should never use abbreviations without spelling it out first.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[gruntguru]

Yes, pressure regulators come in two categories:
- throttling type. A variable restriction in the supply line.
- relieving type. Requires a return line to relieve excess pressure to the pump inlet. Energy wasteful.

Simplest solution IMO is a more appropriate pump ie one with the desired pressure and a "flat" characteristic.

Link



je suis charlie

 

[gruntguru]

JJPellin you say "Below 600 gpm, a change of just 60 rpm will drop the flow by 200 gpm."

The control system setpoint is "pressure", what happens to "flow" will be a function of the system characteristic.

je suis charlie

 

[ForeverStudent]

Two questions.
1. How do you know that the PRV (I assume pressure-reducing valve) lowers the pressure back to 200 psi, if you do not have pressure gauge
2. Why do you need 200 psi on the first floor residential? Looks as way too high. 60 psi downstream of the valve seems to be more than adequate

 

[JJPellin]

Gruntguru,

Agreed. So, if the flow demand drops by 200 gpm the VFD will have to change the rpm by only 60 rpm to to maintain the target pressure. That is a very small change in speed and may be difficult to control well enough to maintain the required pressure within an acceptable range.

ForeverStudent

He has a pressure gauge after the PRV and knows that the pressure there is 200 psi, which is what he needs. The pressure upstream of the PRV is not known since there is no gauge there. They do not need 200 psi for the first floor residents. They need 200 psi at ground level in order to have 60 psi for the residents on the top floor. If they had 60 psi downstream of the valve, no water would reach the top floor.

Johnny Pellin

 

[ForeverStudent]

JJPellin,

Thank you for the clarification, I thought that riser to upper floors was upstream of PRV. This is let's say ... unusual setup. Pressure reducing valves (PRVs) should be on each floor. Pump needs to be changed in my opinion.

 

[gruntguru]

Agreed. So, if the flow demand drops by 200 gpm the VFD will have to change the rpm by only 60 rpm to to maintain the target pressure. That is a very small change in speed and may be difficult to control well enough to maintain the required pressure within an acceptable range.

Yes but if the flow demand drops by 200 GPM (say from 600 to 400) the pressure increases by about 7%. Not very far from target pressure and the VFD hasn't done anything yet.

The VFD is controlling pressure and the relationship between output pressure and speed has a sensitivity which is conducive to stable control.

je suis charlie

 

[JJPellin]

I have generally considered a VFD contolled centrifugal pump to be a good option for a system with a steep system curve and a poor option for a system with a relatively flat system curve. The system described above has a system curve which is intended to be very flat. Could it work? Absolutely. Would it be my first choice. No.

Johnny Pellin