I would like to know what people think about using PID controls on a liquid transfer

bezure

Member
Join Date
Dec 2004
Location
McKinney TX
Posts
13
I would like to know what people think about using PID controls on a liquid transfer
to understand what I am asking let me give you a little more information.
I work in a plastic manufacturing plant.

I would like to know what people think about using PID controls on a liquid transfer system.

I have as many as 10 different plasticizers. any one or combination of plasticizers can be used in a recipe.

that vary from a few ounces to 100 pounds of liquid in a variety of combinations.

I need an accurate way of measurement for each material added in a specific amount to maintain a consistency

of my material .

I have an engineering firm recommending a PID control with an elaborate programming scheme.

I am uncertain if these PID controls will stand the test of time. they will be operating in a very harsh dirty environment.

and be service by individuals with minimal service skills or understanding of how they function.

personally I would believe a positive displacement metering device would be more accurate and less prone to damage .

and all I would require to add to my PLC is a high-speed counter and a few simple math calculations.

I would appreciate any feedback or advice on the subject

thank you

 
Well...
Since you're the customer and the customer is always right.......

Why is the engineering firm not pursuing your idea?

Let me guess, the complex system will take a lot of engineering, which equals lots of billable hours.

I'm a little biased as you may have guessed. In my opinion, engineering firms, if not closely watched, will always choose the most complex method for any application. I think this is due to the fact that they want to try something different and exciting (to them) and it always takes lots of time, and makes lots of money.
 
Are you sure the recommendation is for PID? Maybe it's a question of terminology. I'm curious as to what the engineering firm means by PID.

There is a typical batching technique where the flow has two stages, initial fill stage at a high flow rate, then a secondary reduced flow rate, a trickle rate, when the total reaches some preset value, like 95% of the total.

I've seen this done several ways,

- with VFDs driving a pump at two different rates,

- with modulating control valves driven from 100% to 5% open,

- with solenoids on two feed lines, where the secondary feed line was a small diameter tube, compared to the main fill line pipe.

But even if done by driving a pump at one of two speeds, or operating a valve at one of two conditions, or switching between high flow and low flow lines, I wouldn't consider that PID. It's an on-off logic decision based on whether the total is greater or less than a totalizer switch point.

Your idea of using a PD meter and totalizing the pulses sounds fine to me.
 
If I see that one more time I think I will....

I would use some sort of positive displacement pump. The pumps may need to be synchronized to maintain the proper ratios as in a reaction injection machine. The question is how do you synchronize the positive displacement pumps, with a controller of some type? It is not clear if this is a batch or continous process.

Well...
Since you're the customer and the customer is always right.......
There are few statements out side of, math and physics, with word always in it that are true. I wish my customers were always right. It would reduce tech and application support significantly.
 
I would consider weight cells (load cell) we have many and they work very well, as far as the PID, you could use whatever loop you prefer, math or a PID. Then a 4-20 proportional valve would work, I have several PID loops that are very reliable, it’s the transducer that you need to be concerned with

 
something about the original post leads me to believe that someone thinks that a PID controller is a physical piece of equipment rather than a control method .

I have done a few metering and mixing applications , though not with liquid plastics , and have used Endress & Hauser promass mass flow meters with very good results - I believe they will also handle some nasty chemicals as well .
 
apparently I was not clear on how my process works.

I am using load cells to weigh the material being discharged from a manifold

into a hopper. each liquid is weighed individually and the scale tears prior to additional way up.

my concern is not only the proportional integral and differential (PID).controls.

I have a total of 750 pounds capacity of load cells. (3) 250 pound load cells.

my problem is I use in my recipe sometimes as little as two pounds of liquid

and as much as 150 to 200 pounds of another liquid..

some of my recipes call for two pounds of a liquid. five pounds of another.

twelve pounds of something else. thirty-five pounds of something. and possibly 60 pounds of something else.

my problem is the load sells accuracy at lower weights are less than at the higher weights.

when I am calling for 2 lbs i may be getting actually 2 and one-half possibly three or more lbs of material.

and with all the other equipment running and the vibration caused by my mixers.

I am uncertain that their method is any more accurate than what I have already.

it all boils down to the accuracy of the scale and how fast of a response time they can achieve..

to me a positive displacement metering device would suffice.


the system they want to add is basically the same thing i already have.

the only difference is they want to incorporate the PID programming.


I just can't see how a bit the tricky programming is going to solve this problem
 
Have you looked at mass flow meters? Might be very good for your application needing different weights, with good accuracy over varying ranges.
 
i use pid controllers with vfd's often for liquid movenments, its accurate, i've also in the past on more critical processes used turbine flow meter for the highest accuracy.
 
The mass flow should be a great solution for your application. These would be much more accurate then load cells, and immune to vibration.
 
bezure-

As you said in your last post, any basic feedback control system can only be as good as the feedback device. If you are questioning your feedback device then a feedback control system (PID or otherwise) will have trouble.

The way I see it you have two options.
1)...Find out the source of your load cell system error and either eliminate it or map it so you can account for it.
2)...Use a different feedback device.

You say your accuracy is worse at low weights that at higher weights. Is this statement based on absolute numbers or percentages. That 1/2 pound error at two pounds that you refer to is a 25% error. But that same 1/2 pound at 100 pounds is 0.5%. In reality it is the same error.
If vibration is your problem there are load cell amplifiers on the market that can filter out most vibrations. If it's a linearity issue, you can map that and compensate for it. If it's a hysteresis issue, that's tougher to deal with.

All that being said, I also believe that a positive displacement metering system is your best bet. I would say you need to push your engineering firm on what they perceive the base problem is and why they think PID will help.

Keith
 
I might be missing something here but :- PID controllers are continuous controllers - they are just fine for ensuring that when you specify that you require 2 kg/minute flow - that that is what you get , but not in the least bit suitable in this application -

I think someone has been trying to "sell" you something that they don't really understand .

If compatible , mass flow meters would really be the ticket , if not , what is the possibility of using a small prebatching pot where you could use high resolution Load Cells for the smaller quantities.
 
I do not see where PID loop controllers will help you with this application. You may wish to control a feed rate for a period of time (batching) but you are using scales / load cells as a consistent method of measuring each material. It sounds like your issue is measuring the smaller total weight of some ingredients. Load cells are accurate and provide consistent measurment (accuracy or inaccuracy). You should not see differences in accuracy. You are potentially seeing response time problems due to trying to measure 1 pound and it is too late when you are at 1.25... pounds. One method is to use a second scale/load cell measuring system for the smaller material additions. I would suggest using two feed rates for each material (bulk feed and dribble feed) as indicated in some of the posts above. Set up the bulk feed to run until you hit 90% of the desired weight then run at the dribble feed rate. For the smaller weight materials you dribble feed may start sooner. I have been successful in doing both for dry and wet material batching.
 
Last edited:
The most common method for adding small precise quantites of different ingredients is probably the old reliable metering pump. Give it a setpoint, turn it on, wait for it to shut off.

PID control is probably not needed here. What is really needed is an efficient recipe-handlling method. There are many HMI programs that have this built-in, including RSView.
 
I think bezure that your engineering firm may have the wrong idea of why you have incorrect mixture..Reading your post and looking at it from all angles,i think somepeople would think you are overshooting or undershooting the target..Perfect for PID control..However what i think you are saying is that the Loadcells are not that acurate under low weights..That being said i think everyone else here is on the right track..go with a flowmeter, A PID would not be needed just on off control...as Ken stated some engineers will try to make a mountain out of a mole hill
 

Similar Topics

I have a burner that uses two fuel sources. Currently we only have one PID tag that runs the fuel controls for the active burner, but if we...
Replies
6
Views
2,110
Hi everyone, I have a problem with EN2TR during I using the EN2TR web page. I was trying to use EN2TR web via connecting thru E-TAP and go thru...
Replies
1
Views
1,223
Is there a way to create proportional logic for a control valve without using a PID function? I want to create a simple P controller without...
Replies
8
Views
2,584
Hi Everyone! My name is Vidy and I am new to PLC AND PLC PROGRAMMING As Part of my final year project I have been tasked to program a PID...
Replies
5
Views
2,981
I had a problem on a hydraulic servo controlled cylinder, which I believe was caused by the PID loop and the cylinder reaching a hard stop at one...
Replies
10
Views
2,423
Back
Top Bottom