Oil and water accurate dosage

[shubhamw]

Hello!!!!
For an emulsification process we are feeding oil and water into a nozzle using two different gear pumps .
We are settng the dosage using frequency drive by calculating flow and output equivalent frequency. But there seems to be some error.
Is there any method of achieving more accuracy and less error maybe using a PID controller or something like that? So that the error could be made nil everytime .
Thanks

 

[MikeHalloran]

How do you know there's an error?

How big is the error?

How big do you want it to be?




Mike Halloran
Pembroke Pines, FL, USA

 

[Latexman]

How do you calculate the flow? Sounds like you adjust based on the theoretical displacement of the pump/RPM. Is this done on an absolute basis or a relative basis? There is always some slippage/internal recycle inside a pump. It varies with speed, viscosity, time since new/rebuild, etc. If this is done manually, there is human error. I'm sure the error can be improved. A P&ID or PFD would help us understand the current process and progress the conversation to something that meets the need. Btw, what error is acceptable? "Nil" may get expensive.

Good luck,
Latexman

Need help writing a question or understanding a reply? forum1529

 

[shubhamw]

I know there's an error because we calculate there volume that has been pumped after say an hour an there is always a major difference between theoretical calculation.

Some error is acceptable about 2% .
Another reason might be the wearing of gears of the gear pumps we are using but they are fine for the application.

We calculate the flow like this -

currently we are using a 100 LPM gear pump connected to a frequency drive .

The motor is 1440rpm and frequency is 50hertz.

So, I move the drive to 25 hertz to get an output of 50 LPM and so on.
But, this is not highly accurate.

Maybe the picture is clear now.

Please suggest a way not too expensive of bringing down the error and also minimising manual interference.

 

[MikeHalloran]

Did you measure the motor RPM, or are you reading 1440 rpm off a nameplate or a datasheet?

I.e., if it's not going ~1500 rpm at 50Hz, maybe you could use a motor with less slip. ... or a tach or encoder on the motor. What you can do in that area depends on the specifics of the motor and the controller. Since you have revealed squat about either device, it will be hard to offer more help.



Mike Halloran
Pembroke Pines, FL, USA

 

[shubhamw]

We read the rpm through the rpm meter.

What I want to do is place a flowmeter ahead of the pump with a transmitter .

The flowmeter should signal the flow to a controller which should make the error nil maybe by controlling the frequency drive .

Is there any other possible way wherein we could digitally set the flowrate and be able to achieve it.

I have less idea about automation being a mixer designer.please help

 

[LittleInch]

For that level of accuracy, you stated the answer in your post. Fit some sort of flow measurement device in the piping, turbine meter or ultrasonic meter on a straight length and the feed the output into a PID controller which feeds back into the vfd to increase speed or decrease speed to maintain your flowrate set point. You may find the vfd controller already has this function built in as it is a common requirement to control the speed of a pump to maintain a fixed flow or pressure.

My motto: Learn something new every day

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

 

[shubhamw]

Through the built in function of the vfd, I can control the rpm but not the flow setpoint.

Anyways, I think this should work and nil my headache.

Thanks!

 

[engineeringcalc]

I agree that the first approach would be to have two flowmeters one for water and one for oil. With the signal from the flowmeters it will be easy to adjust the speed of the pumps. Probably your VFDs already have a PID controller built-in.
For a better accuracy you would need a oil in water analyzer. If you are mixing in a pipe you can measure the content of oil in water and close the loop. If you are mixing in a tank this probably will be very diffucult or impossible.

I know a company that is specialist in this type of blenders (Jiskoot). They are really good (and not cheap) and only competitive for big systems. Anyway they approach the problem like I explained above.


REgards,