Variable Speed Pump Control

[smersh]

I have two Variable Speed pumps in a Duty/Standby configuration. Only
one pump runs at a time, controlling to a pressure setpoint. A new stream is being added which will produce large surges in demand. I need to change the code so that the second pump will ASSIST the first
during spikes in demand. What is the best way of coding the assist pump.

1) Just enable the second pump to control to pressure as it normally would. Therefore, you will have two pumps controlling to a pressure.

2) Set the assist pump output to a fixed speed, say 10%. Increase this figure if the pressure remains too low for a time. Decrease when output of Fist pump falls below a preset value (say 25%).

Any ideas?

 

[Fritz_frederix]

Hello,

I used to keep the pressure of a watersupply constant. We had 4 pumps which ofcors were in parallel. One of the pumps was controlled with a frequency converter.Two pumps were normally giving 100% and the frequency converter driven pump varies according the needed pressure. When the frequency converter driven pump reached about 80% of it's normal flow/presure then the third pump would come in and the frequency converter driven pump should run at 25 % of the
normal flow/presure. When the frequency goes below 15% then the third pump stops and the frequency converter driven pump goes back to the necessary freq.

We used a Danfoss frequency convertor en VLT for it.

Rudi

 

[DickDV]

Usually it is desireable to alternate which pump leads and which one lags periodically to even out the wear on the pump/motor packages.

If you want to do this, then you need a drive with a Pump/Fan Macro in it. Some manufacturers may call it a Lead/Lag Macro. ABB's ACS550 model series has this macro as well as other manufacturers.

If alternation is not important to you, then a system similar to what fritz frederick mentions will do the job. You need to be careful with that arrangement tho because it has a definite tendency to oscillate the second pump on and off repeatedly at certain pressures and flows. Most of the time you can tweak the on/off setpoints to get away from this.

You could, as you mention, put a second inverter in and add it to the pressure transmitter loop. Set it up as a setpoint controller and both will work in parallel. The benefit of this system is that it will have the fastest response to sudden increases in flow. The disadvantage is cost, both initial in having to buy two drives and operating, in having both motors running so slow they are pumping almost no water during periods of low demand.

Which system to use is your call!

 

[smersh]

Both pumps already have inverters.
DickDV, are you saying that I should apply the output from the control
loop to both pumps? If i bring in the second pump at when the output is 80% then there will be a sudden surge in pressure.

Would it work if both pumps had their own PID control loops?

 

[Rube]

smersh

We used to do the exact thing you are describing. We did not worry about a lead pump, however. Both pumps were on VFDs and the pressure transducer was logically used to control both pumps. If pressure dropped too low from the primary pump, we ramped the second pump up to compensate. Conversely, when the demand dropped, we spooled the second pump down. Worked very well.

And, yes, both of our pumps were on PID loops.

 

[DickDV]

Since you have two inverters already and don't seem interested in alternating the lead pump, run the 4-20ma transmitter loop through both drive inputs.

Set the lead pump to regulate to the desired pressure and set the lag pump drive to regulate to a slightly lower pressure than the lead drive. Don't set the lead PID controller too tight. This will allow the pressure to fall on increasing flow just enough to get the lag pump to start and regulate to the slightly lower pressure.

You could, of course, run the instrument loop just to the lead drive and then use the analog speed output of the lead drive for the speed input of the lag drive. Set the analog speed input of the lag drive with a lot of low end offset so it will not start until, say, 14ma. That way the lag pump will only begin operating when the lead pump is running above the speed represented by 14ma. Both pumps would be at full speed when the lead drive is at 20ma speed output. You would have a better system if you set the lag start signal to come on at 13.5ma as well. If the drive relays have level detection, they could be used for that. That way, the lag drive would not run at all until the lead drive was above 13.5ma on its speed output.

This seems like a nice way to do this job but it has a major weakness. If the lead drive fails or faults for any reason, the whole system goes down. If you use my first arrangement, either drive could be faulted and the remaining drive would continue to maintain pressure up to its maximum capacity.

 

[harryg]

PID is the way to go. In a two pump system I use a smart relay, if lead pump cannot achieve or maintain setpoint for 15 seconds then the lag is called to "help". I agree with DickDV about the Lead/Lag scenario.

 

[Tom Jenkins]

One word of caution. You have to examine the system curve, the pump curve, and the operating characteristics carefully at minimum and maximum operating conditions. In most cases, especially with a high static head system, you will not be able to run a pump below a minimum speed. Doing so will cause recirculation and heating in the pump because it cannot create enough pressure at low speed. The result will be eventual pump failure. This minimum speed will usually be much higher than 10% speed.

If you are uncertain about your pump system performance limits, contact the pump supplier.

In general I use Dick's system, running both pumps in parallel at the same flow rate. Running one pump at 100% and the other at minimum speed will usually result in a big step in flow rate, since you will go from 100% to, say 130% flow, due to the minimum capacity limits mentioned above.

 

[DickDV]

Tom's concern about the lag pump minimum speed is certainly valid. Essentially, when the lag pump starts, it must be at a high-enough speed to cause a small amount of forward flow. This is usually easy to determine by seeing that the check valve opens at minimum lag pump speed.

If you don't take this precaution, the lag pump could possibly be turning the pump with no flow. The liquid heats up and soon the pump seals are gone. Something to avoid!

If the lag pump output relays have level detection, simply set one of them to operate at the desired control loop level. Connect the relay contacts to the drive run/stop input and set the drive minimum speed to a level that causes slight forward waterflow at the system pressure that exists at the operating level of the relay.

If the drive doesn't have level detection relays, then you need a smart relay or equivalent to do the same thing.

 

[Alan Case]

I have gone the other way with this problem. The first pump runs off the pressure transducer (with a PID loop)and when it is at full speed for x seconds I start to ramp the second pump up. Just use a timer in the PLC to slowly increment the set speed. Pump 1 will PID down in speed. When the speeds match I tell pump 2 to follow the speed of pump 1.
If the speed drops below Y rpms I drop off pump 2.

Regards Alan

 

[Junsheng Liu]

Pump control

Hi, Dick:

I prefer to use PID loop for both pump control combining with some kind of interlock in case one pump failed or couldn't run.

 

[DickDV]

Well, there! It just goes to show that there are several good ways to do this.

One potential problem with running both drives in PID mode on the same process transmitter is that most 4-20ma drive inputs are 500 ohm to convert 20ma into 10vdc. Two of these in one loop will likely exceed the max voltage capability of the transmitter loop.

An easy solution if your drive has input offsetting and gain adjustments is to shunt the ma input you are using. A 1000 ohm shunt would equal 333 ohnm or 6.7vdc max and a 500 ohm shunt would equal 250 ohms or 5vdc max.

Adjust the offset and gain on the input accordingly and the loop voltage problem will go away. The rest of the drive setups will be the same as described above.

To improve reliability in case one drive fails or faults, you may want to use a separate loop power supply. If you use the drive power supplies, at least use steering diodes to combine the two supplies. That way, if one goes down, the other will cover for it instantly.

I don't know if I'm allowed to say this but, if your brand of drive doesn't have all these nice features (offsets, input gains, loop power supplies, etc) maybe you should consider another brand! (hint, hint!!)

 

[harryg]

One potential problem with running both drives in PID mode on the same process transmitter is that most 4-20ma drive inputs are 500 ohm to convert 20ma into 10vdc. Two of these in one loop will likely exceed the max voltage capability of the transmitter loop.

I use a signal isolation module for this and have ran up to 4 drives on a single pressure transmitter with no problems