2 Positive Pumps

[Greg Dake]

Leadfoot, appreciate your feedback, as well as others. To answer your last question, I'm not sure...I don't know what control model to use yet.....I need to ponder this for a few days. I'm kinda trying to decide weather or not to leave the control up the the VFD's, I'm using ABB ACS550's or run the control out of the PLC were're integrating....A PLC really wouldn't be required, but we're integrating an HMI with it also, so they can have visualization as to what's happening downstream...


Greg

 

[jamesau]

Sounds like fun!


It’s always a good idea to speak with a knowledgeable technical representative from the homogenizer and pump manufacturers; explain what you’re trying to do. If they don’t offer direct insight into mastering the application, they may at least advise you to the absolute constraints of the machines and what the implications are if you violate them.

Maybe consider the following for a process modification / control strategy.


1) Use a recycle line around the homogenizer (from discharge to suction) that contains a sanitary control valve; this will act as your pressure relief. Nominally, you’ll want to maintain some suction pressure, Ps, on the booster during normal operation. Control the sanitary valve with PI controller to some higher pressure, Ph, which is meant as a conservative high limit to protect the pumps. During normal operation, actual pressure will be below Ph and the recycle valve will be closed.



2) Hopefully (I can be optimistic from my armchair), flowrate from each of the pumps will be nearly linear with pump speed and under normal conditions the ratio of the pump speeds will be nearly constant. Through experimentation, you can determine what this ratio is; call it R. Now set up a controller where the speed of one pump is ratioed to the other. So S1 = R * S2, where S1 is the speed of one pump and S2 the other. Another PI controller attempts to maintain Ps by having as output a trim or limited change to the ratio, R.



3) Next have a flow controller for overall rate that controls S1. This controller will be detuned quite a bit for an overdamped response (because you’re more concerned from a control perspective with maintaining pressure).



So, how does all this work? On startup, put all controllers in manual. Open sanitary valve to some preset. Turn on pumps at low speed at nominal R, wait x seconds, and put pressure controllers in automatic. Wait y seconds and put flow controller in automatic. The sanitary valve controller will act as a constraint control to safeguard equipment yet allow the other controllers to recover gracefully from startup transients and other unforeseen hiccups.



That’s enough for now. Good luck and have fun.



Jim

 

[fluidpower1]

Could you use an Eddy Current Drive on Pump 1 that is controlled by a feedback from a pressure sensor in the line between the pumps?

If this is completely crazy just consider the source.

 

[Leadfoot]

Could you use an Eddy Current Drive on Pump 1 that is controlled by a feedback from a pressure sensor in the line between the pumps?

If this is completely crazy just consider the source.

That don't sound crazy, sounds like you are trying to use the KISS principle.

If he can use a VFD, he certainly could use the eddy current control. I would think the eddy current drive would be a bit softer response depending on the vintage of the contro. The VFD would be more power efficient. With a bit of tuning, the VFD would be able to respond as needed. Most VFDs today have excellent P and I adjustment capabilities on their own.

 

[darrenj]

Correct me if i am wrong..Wont you have 3 pumps?..A so called "stuffing" pump for the homo?

As for controlling pump 1 it is not usualy an option. These are basicaly timing pumps..They put X amount of liter/gallons per min. This is used to calculate Pasturisation times etc. so they run at 100%.

Speaking of KISS The two Homos we have are cabable of pumping a 2" line to 100'?..what brand are yours?

 

[Mark_H]

I have no experience in what is being discussed here, but I am curious about something.

If for the first pump was replaced by 2 or 3 smaller pumps (plumbed in parallel) driven by the same shaft but having their outputs offset by 180 or 120 degrees, you would have smaller "pulsations" going down the pipe to the 2nd pump. Would the smother flow help improve the life of the piping and make the control of the pumps simpler?

 

[darrenj]

I have no experience in what is being discussed here, but I am curious about something.

If for the first pump was replaced by 2 or 3 smaller pumps (plumbed in parallel) driven by the same shaft but having their outputs offset by 180 or 120 degrees, you would have smaller "pulsations" going down the pipe to the 2nd pump. Would the smother flow help improve the life of the piping and make the control of the pumps simpler?

Not a bad idea but the big factor here is that it is a Homo pump..
it takes 2 things..Like cream and milk and mashes them together so they dont seperate..kinda more complex than that but that the basic idea.

D

 

[Lancie1]

D,
Yes, but Greg said earlier that the Homo Pump was not really being used to mix two things...

Greg, why not go even farther with the KISS principal? Replace the Homo timing Pump with one larger booster pump and eliminate the second pump. This may be cheaper in the long run. The controls for two pumps in series may not work, but even if they work, they may cause cavitation in the pipes, and a second pump introduces more connections, valves, and places for contaminant entry into a sterile system.

If the goal is to keep the pudding clean and still transfer it to another place, then eliminate one of the pumps.

 

[drewcrew6]

Keep in mind I'm clueless of about half of the terms as far as the food.


Sounds like youre mixing 2 ingredients then pumping a long way .
Is it possible to pipe each one a long way then mix?
Seems like it would be easier to control and maintain then the other way other the the more pipe and fittings. But my thought is if you have pump "surge" enough to hammer pipes then the system will be short lived and I'd hate to have to clean pudding off the ceiling!!


Drewcrew6

 

[darrenj]

D,
Yes, but Greg said earlier that the Homo Pump was not really being used to mix two things...

Greg, why not go even farther with the KISS principal? Replace the Homo timing Pump with one larger booster pump and eliminate the second pump. This may be cheaper in the long run. The controls for two pumps in series may not work, but even if they work, they may cause cavitation in the pipes, and a second pump introduces more connections, valves, and places for contaminant entry into a sterile system.

If the goal is to keep the pudding clean and still transfer it to another place, then eliminate one of the pumps.

True but he also stated they were using it as a timing pump..This is obviosly critical..(I would assume the pudding still has to be pasturised)But i still cant figure why the Homo would have such a hard time pumping 100'.....

D

 

[darrenj]

Drew..dont think they he is using the mixing part..however..just to put things into perspective..The 2 hom "pumps" we have are 50 hp motors, driving 3 or 4 pistons about 3" in diameter, The floor plate is about 8' x 8'.. They are not things you can move easily!! Obviosly the homo is used in more than one product so usualy the homo stays in one place and you desighn/ build around it..It is only moved as a last resort..

Greg..Can you describe in detail what you are tying to do? (PM me if need be) i can then run it by one of my plant Quality control guys..(He is a HTST/Homo Guru..He actually should be an engineer but dosnt want to leave the floor.)

Sounds interesting and challenging..far above my abilities..

d

 

[PhilipW]

Sorry I did not read this thread sooner...too busy watching a really tight election here in NZ.

I've done exactly this kind of application on a multi-stage vacuum system. The intermediate and final boosters pumps in this system are positive displacement and all of them where controlled by VSD's.

The trick to getting it all to work well was to run the intermediate pumps VSD's in TORQUE mode. In this case we had to control the intermediate vaccum to a specific level in order to share the load effeciently between the final and intermediate stages and to optimise condensation in a solvent recovery system that is placed between them.

From a control point of view all I did was run a PID loop that controlled the intermdiate suction side to a setpoint, and send the controlled variable to the torque setpoint of the VSD controlling the intermediate pump motor. In this case I was using AB Powerflex 700's all networked so setting this up was easy.

The final booster stage PID loop sends a setpoint to the SPEED of that motor, so in effect this pump determines the FLOW through all the stages. The intermediate pump only now needs to be commanded to generate enough TORQUE to maintain the required vacuum between the two stages.

It turned out to be remarkable easy and stable to set up. It would also totally eliminate any need for an accumulator.

 

[drewcrew6]

thank you Darren


What is meant by the term "Timing Pump" ?

Okay so if the Homo pump can't be moved and the booster pump is 100' and also cannot be moved . What were the original intentions of the 2 pumps?

Drewcrew6

 

[DickDV]

To Tom Jenkins. I see your reasoning behind the difficult tuning of the follower pump in this situation due to the incompressibility of liquids.

Wouldn't this be an application that would benefit from having both p.d. pumps follow the same speed signal---volume for volume with a PID pressure loop set up as a trim on the first pump.

Seems to me you could stabilize the tuning problems somewhat by keeping the trim down at 5% or some other low value.

The original poster indicated that he was using ABB ACS550's on this job which I believe have this capability. If not, the ACS800's certainly do.

Be interested in your thoughts on this.

 

[Jimmy_Oz]

i have a very similiar system here making aseptic dairy products; pump 1 is the master and pump 2 is the slave, pump 1 has a PT on the outlet that sets the speed of pump 2. Pulsation dampener on the inlet of pump 2 (shear plate due to hygine reasons not a pressure releif valve)and away you go, oh and of course a PLC which does the speed calculation for pump 2 (using pumps 1's PT read out) and just to add a little spice to the system we also have a temp/time correction factor so it looks at our process temprature set point against current measured value and will trim up to +/- 5% of the speed of pump 1 (which in turn alters the pressure and therefore the speed of pump 2) to either increase or decrease the time at temprature for our process