VFD controlling a pump by GPM instead of RPM

[Robb B]

Hello, I was wondering if anyone has some insights into a VFD controls scenario. We have a VFD controlling a pump which is currently setup so that the operators can manually input a specific RPM (600 - 1800 I believe) which the drive will obviously maintain very effectively. The operators have requested the ability to input a flow rate in Gallons Per Minute instead of RPM, in order to try to achieve a flow through the pump. This seems reasonable, but I wonder if there is some caveats that I should know of before implementing such controls. I do have a minimum speed approx. 20% as a hard lower limit on the drive, but I wonder if the drive will end up "hunting" for a speed which satisfies an inputted GPM.

 

[GaryS]

First a VFD doesn’t hunt for anything, it will run at the speed or torque commanded of it.
As long as it is in its operating range.
I assume you have a flow meter connect to show the operator the actual flow.

To do what you want a simple PID control is what to need. Feedback from the flow meter is the process variable for the PID control. The PID will adjust the speed reference to the VFD s the actual flow = the set point. As for the hunting proper tuning of the PID control will minimize or eliminate any hunting.
Some VFD’s have a PID control as part of their control software.
I that case no external control is needed.

 

[Robb B]

There is no flow meter, but the VFD is using a pumping routine which has an internally calculated flow. Would that be a reliable enough feedback?

 

[TheWaterboy]

If the VFD has created that data, it might have the means to use it for feedback. Its docs will tell you. I can't imagine why it would provide that info if it was not capable of using it as a PV.

 

[Bit_Bucket_07]

There is no flow meter, but the VFD is using a pumping routine which has an internally calculated flow. Would that be a reliable enough feedback?

If there is no flow meter, then the VFD must simply be inferring flow rate from pump motor speed. While the calculation might employ some linearization, the flow rate is still an inferred value.

 

[RedSoxFan3]

You cannot control what you do not measure!

 

[cardosocea]

You cannot control what you do not measure!

Or to put it more eloquently:

When you can measure what you are speaking about, and express it in numbers, you know something about it, when you cannot express it in numbers, your knowledge is of a meager and unsatisfactory kind; it may be the beginning of knowledge, but you have scarely, in your thoughts advanced to the stage of science.

Not my quote, it's from a bloke called William Thomson.

 

[ndzied1]

A lot will depend on the type of pump and the media you are pumping? Is the pump positive displacement or a centrifugal pump or something else?

Is the media stable in its temperature and viscosity?
Does the pressure remain constant or fluctuate?

How accurate does the actual flow have to match the entered number?

 

[Tom Jenkins]

1) What nOrM said is 100% on point.

2) I assume this is a positive displacement pump. Since the operators have been getting by as is I also assume accuracy requirements aren't extreme. If so the flow can be reasonably calculated by an equation in the form y = mx + b where y is flow, b is slip or no flow rpm and x is rpm.

I have controlled flow successfully with positive displacement pumps and blowers many times. Plug in a flow, and calculate the rpm using x = (y-b)/m.

3) If the pump is centrifugal then you should use a flow meter.

 

[TheWaterboy]

Grundfos is very proud of their ability to determine flow just by speed and pressure of their packaged systems... But yea, use a flowmeter.

 

[Tom Jenkins]

Grundfos is very proud of their ability to determine flow just by speed and pressure of their packaged systems... But yea, use a flowmeter.

The problem with this is that it assumes the pump is in as-new condition. Impeller and casing wear will degrade accuracy. Also, it assumes the pump is running on the curve - most of which are ± 3% or so. The method is OK for most applications, but end users assume an accuracy that may not exist.

 

[TheWaterboy]

Of course - I think its silly to assert such a claim but such is marketing these days.

 

[TheWaterboy]

Hey Tom,
I have a pump performance issue I have been hunting for a long time (years) - is there a good online forum for discussing it with folks like yourself ?

 

[BryanG]

I will give a different view. If the operator entering a speed works, and gives you consistent flow results, then the operator entering a value in some 'units' (Gallons) can also work. I have seen it with a very bodged chart on a wall that the operator used to know what speed to set.

ABB drives can be programmed to work in 'units' rather than RPM or Speed %. So it depends on the drive you have on whether it can be changed from RPM to something else, and of course you will need to know how much flow a range of RPMs give so that you can convert RPM to Gallons. Sometimes a simple solution is good enough, sometimes you need flowmeter feedback and automatic speed control. I like simple solutions, they match my personality

 

[TheWaterboy]

Just a lookup table... X RPM = X GPM. It does require swallowing a lot of "IF" though. But yea, maybe that's OK here.
Such a table could be printed an just posted near the device.