pumps in series how to regulate??

I have an application were I must regulate the flow from two pumps in series.. The pv comes from a flowtransmitter and the sp should come from an operator that sets desired flow.. Should i make it easy and only have one signal to both pumps or is it a better way??

Two pumps in series?
Discharge from pump 1 goes directly to inlet of pump2?

Are pumps same size?

If so, I would have pump 1 run slightly faster than pump 2, use pump 2 for control.

PID output goes to pump 2, pump 1 speed is (PID output * 1.02) or 1.03

First thanks for the answere.. But can someone explain more deeply why you should drive the first pump a little bit faster..??

you want to make sure the pipeline is absolutely full with enough pressure
to prevent the second pump from cavitating

I would recommend you consider a few problems I foresee with running two pumps in series. One you need to consider what type of pumps you are using reason is due to hydrostatic head for example its more than possible that the first pump is going to fight the second pump if they are say diaghram pumps the second pump may cause to much head pressure for the first pump to operate. Also if the first pump should break down then you can damage the second pump due to loss of volume creating cavitation.

I'm not going to pretend I'm expert on this, but I've seen it done. First, a few thoughts:

I am assuming these are centrifugal pumps. If not, you are probably in trouble.

Cavitation of the second pump is not a flow related function, it is pressure and temperature related. Cavitation occurs when the pressure at a pump impeller falls below the vapor pressure of the water. If the inlet to the second pump is pressurized by the first I wouldn't expect cavitation to occurr.

This is not a PLC or control problem. It is a mechanical engineering problem. Once you get the system requirements right the programming is easy. I strongly suggest you contact your pump supplier and get the system requirements from them.

Thats essentially what I was trying to describe although your explanation is the more accurate. The one time I've seen this attempted was with two 30 hp centifical pumps from Gormon Rupp due to poor monitoring or fault detection the system regularly broke down although I definetely agree that this example was never properly engineered.

Ditto all the cavitation comments.

First look for prior examples. I am sure this has been done before.
A pressurized accumulator is necessary between the two pumps. I think two pumps without the accumulator would be very tricky and the second pump will cavitate when things go wrong. It may not survive commisioning. The accumulator liquid volume can be determined by the pressure in the accumulator and that would need to be maintained. The second pump would regulate the flow, the first pump would regulate the pressure/liquid level in the accumulator. Obviously the second pump can not operate unless there is enough pressure in the accumulator. I would have a set point override so when the pressure started to drop too far I would subtract from the SP which would slow down the second pump and keep the pressure from dropping too low.

Today the operators control the system with two potentimeters wich gives a SP to the VSI and it isnt working soo good.. And my firm (Im a consultant) is there to do some other job (a new operator panel and such) and they asked us for a automatic soulution for these pumps aswell.. So it is a "working" process but they want it automatic beacause now they have to finetune it several times a day to get the right levels..

Hope you understand what I´m saying my English is a little weak..

But I will try calling the pump supplier and talk with them thanks for the help anyway..

Assuming these are centrifugal pumps, I would run them at the same speed.

With pumps in parallel, you just add the pump Q-h curves horizontally, in series add them vertically. As long as you haven't got any strange characteristic there should be no problem.

If the pumps are the same, if the NPSH requirement of the first pump is satisfied, then the second one must be satisfied, for pretty well any speed difference.

If they are positive displacement pumps, then you may have a problem.

As everyone has said net positive suction head (NPSH) must be positive for a centrifulal pump. An example of centrifugal pumps in series is condensate (50 HP 5 psi) and feed water (150 HP 600 psi) on a steam plant. The condensate suction is from condensor hotwell (vacuum at 28 or so inches) which discharges to the feed pump suction. Between condensate and feed pumps was surge tank approx 10 feet above both so there is always 10 head at feed pump suction. Surge tank overflowed to reserve feed tank.

Purpose of the surge tank was to always ensure NPSH to feed suction and to ensure NO air to steam generators.

Condensate pumps were variable speed 3 phase controlled by 1965 era VFD control setpoint was hotwell level.

The first pump in these series setups MUST pump a little more than the second to ensure 2nd NPSH otherwise 2nd will cavitate or worse yet flash and "air bind" which will wipe out the seals quickly.

IF my setup I would run the 1st pump at set speed and have some kind of overflow to take care of surplus and yet maintain NPSH at 2nd. 2nd pump would be variable speed controlled by suction pressure.

ALTERNATIVE
1st pump is VFD controlled by discharge pressure high enough to ensure NPSH to 2nd. 2nd runs at steady speed. 1st starts first of course. This may be better setup. Pressure switch on 2nd suction that will not let it run with low NPSH.

Dan Bentler

A lot of really good comments on this one.........

My input would be as follows: The OP stated that the process to be regulated is "FLOW". The main reason to add centrifugal pumps in series is to increase "PRESSURE". For example, multi-stage pumps are (largely) just stages in "series" for the purpose of building a pump that is more efficient in a higher pressure range.

If the pumps are located together, I would consider running them at the same speed (conceivably with same analog input to two drives OR if the horsepower is low, consider running both pumps with the same drive.) EITHER WAY, it will be VERY important to monitor the performance of both pumps, so that if either one goes off the curve, BOTH will go out on safety.

In the scenario as described, "suction pressure" would be monitored only for the sake of determining the proper performance of each pump within its engineered performance curve.

The process variable is "flow".

The PID concept will work just as well for "flow" as for "pressure". The key is to keep the pumps operating within their engineered performance range as determined by measuring "pressure", and to limit the "max Hertz" to keep them from going "off the curve". This would be a simple drive parameter setting at startup, backed up by some carefully set pressure switches at the intakes of EACH pump.

While you're developing the automatic system of safeties, you should consider adding (1) "overtemp" sensors and logic, and (2) careful setting of the "overload" feature of the VFD(s).

Happy pumping,

Stationmaster