Control of pumps in a closed potable water distribution network 1

[waseem19]

Hi all,

I'm currently working in a closed system of potable water distribution system, pumping directly to consumers, there are no storage tanks that I can use to shut off the pumps when it gets filled, so what is the most common way of controlling the pumps in this case ?

1- Pressure sensors at the furthest hydraulic point in the system ( or even several pressure sensors in several locations ) these will have several set point keep tracking the pressure from a minimum of 2bar to a maximum of 4.0bar.

2- A pressurized tank at the pumping station, is this the same as the hydropnumatic tank ? what is the difference between this and the surge vessels ? will this hydropnumatic tank work for surge protection ? what should it be sized for if it will work for both ?

The pumps will be VFDs. 2 duty and one standby. Pumping 180l/s at peak. 90l/s on average.

Your help is very appreciated.

 

[kchayfie]

When the treatment works I design deliver straight into the distribution network without storage, the final water pumps are controlled by a pressure transmitter at the outlet of the treatment works. We rely on the network designers/modellers to tell us what the pressure setpoints need to be at the outlet of the treatment works in order to maintain distribution to the whole of the network.

 

[Joesteam]

This sounds like a typical pressure booster used in many buildings. Normally there is a tank rated for the pressure and water for domestic use. It is not really for excess demand (the pumps must be sized for maximum demand) but rather to supply a comparitively small amount pressurized water so that the pumps do not cycle as much. For example during very low demand periods should someone use the a small amount of water (ie: flush a toilet), the water will come from the tank and the pumps will not have to turn on and off.
As for the pressure, although you do need to control the pressure at a MINIMUM at at the furthest point, the pump control is usually done locally. This is primarily due to safety - so if a valve somewhere is shut off the pumps will not overpressurize the entire system.

 

[bimr]

kchayfie has outlined the typical design.

I would add that you should do a surge analysis to check for possible water hammer. You should also design the system with a surge relief valve to prevent over pressurization.

 

[Pump2005]

The water system I operate at has two water towers which I can run the pumps off of either tank level. I can also run the pumps by system pressure which would work for you since you don't have storage tanks.

 

[thewellguy]

A pressurized tank can be the same as a hydronumatic. A Hydro tank has a mixture of air and water in it. Some tanks have a rubber bladder to seperate the two. (Well-x-trol by amtrol is a big name brand) the larger tanks 200 gallons plus require that air be added to maintain the correct working ratio. This can be done manually or with a tank mounted compressor. This types of tanks are used to store water. A surge tank is not designed to store water. Its purpose is to act like a shock absorber in a car. It purpose is to reduce water.

I would use vfd's to control the pumps from a transducer located in the pumphouse.

 

[cvg]

your pumping rate is quite large (about 2,800 gallons per minute peak) and it seems that this is a fairly large service area. a hydropneumatic tank sized for your application would be very large. A better approach if you can afford the capital investment would be to construct a gravity tank to help deliver water to your system at a more uniform pressure and to reduce the wear and tear on your pumps.

 

[waseem19]

*Pump2005,

you said system pressure : how do you measure that and where do you measure it and how do you send the signal to your pumps ?


*thewellguy,

"I would use vfd's to control the pumps from a transducer located in the pumphouse"

Can you please explain more, do mean pressure transducer in the discharge line ? I cannot use level transducer cause I have no gravity tanks, even If I do they wouldn't be located in the pumphouse.

the idea of measuring the pressure of the discharge line directly after the pumps seems interesting , I thought about it. would this work ? it is definitely cheaper and easier.


*Cvg,

thanks for your input, I'm now confident that the hydropnumatic tank is not an option since it has to be too big to serve this community.

unfortunately building a gravity tank is not an option.

Guys,

I just need someone to confirm that using pressure sensors in several hydraulic critical points "these could be 2km away from the pumping station" would be a reasonable solution to control the VFD pumps. Please note that these pressure transmitter will be placed on the distribution mains.

Regards.

 

[thewellguy]

I work on the city water systems around the Chicago area. Most of the system are set up this way. They have pressure transducers close to the pumps that send a 4-20 ma signal to the vfd to control pump speed. I do not understand why you want to measure the pressure 2km away. If is is a closed system that has correctly sized mains, the loss of pressure should minimal. There may be a differential due to elevation changes, it this is the case I would try and location the pumps at a low point.

To answer you other question, yes it is very common in this area to control systems like this.

 

[bimr]

I agree with the thewellguy's comments.

 

[cvg]

if you are going to install a gravity tank, you will need to measure the level in the tank so that a) you don't overfill b) you don't allow it to drain completely. By installing the tank, you could pump from your well at night to fill the tank and then during the day, the tanks provide the flow to your community. During peak hour or during fire flow demand, the pumps would come back on. This might save money on electricity, because it is often cheaper during the night.

 

[Pump2005]

Having a storage tank producing the pressure by elevation is the way we can run our pumps if we want to rather than the water level in the tank. Our effluent finished water pipe is also our incoming water feed from the system and we have a pressure readout just as it comes into our facility. You could also put a pressure sensor with a radio transmitter at a point in the system that is most reflective of system conditions. We use one tank as our control tank and have our on/off set points where the pumps will come on when there is a drop of 10 feet and off when it rises back up. This way we are always maintaining a near full tank.

 

[waseem19]

Thanks a lot everybody. That was very helpful.

I’m new in the water distribution network design. There is a senior guy in the office who is convinced that controlling the pumps near consumers “2km away” would give the best indication of system pressure.

I checked the model for severe changes in demand to see how would that affect pressures in the system as overall , I noticed that when pressure was reduced by 50% at 2km away from pumps the discharge pressure just after the pumps was reduced by only 17%. The main reason for this big difference is because there is a relatively larger main feeding the smaller branches where the demand change was assumed to occur. I’ve assumed fixed speed pumps in this scenario to see the direct impact of a huge demand change.

The 50% reduction in pressure was due to a modeled fire incident.

It seems that there is some sense in what my senior guy is saying. Even though the 17% still give enough flexibility to detect the change but smaller demand changes could have much smaller impact at the pump discharge.