PID Loop for Glue System

Hi everybody,
I am looking for a way to use pump output pressure and tank level in a PID loop to control motor speed. I am pumping glue to a tank on our line. My inputs that I have are a pressure transmitter after the pump and tank level. Both have setpoints, and I am looking to find a good way to combine the two and create a floating variable for the PID loop to compare with set point. Currently my plan is convert both inputs to percentages then take the average as the process variable. With this plan the set point would be 1 if both were at perfect levels. Am I on the right track with this?

Welcome to the forum.


Typically a relationship between related process variables (tank level and pump output pressure in this case) is dependent on the process, so controlling both to fixed, independent setpoints with one control variable (motor speed) will be problematic.



We need more information about the process. Can you sketch it out e.g. is the pump pumping glue into the tank or out of the tank? If pumping into the tank, what is setting the flow out of the tank, or if pumping out of the tank, what is setting the flow into the tank? What kind of pump is it (e.g. positive displacement)? Is there a partial recycle of glue from the output of the pump back upstream of the pump, and if so, what controls that recycle flowrate?

Typically a level is controlled to provide capacity for transient process variation between upstream and downstream flows without overflowing or emptying the tank, so the average variation over time must be zero. Is that the rationale for the tank level setpoint i.e. why is the tank level being controlled to a fixed independent setpoint?

What is the rationale for the pressure setpoint i.e. why is the pressure being controlled to a fixed independent setpoint?

Don't re-invent the wheel.

Buy a machine e.g. https://www.nordson.com/en/divisions/adhesive-dispensing-systems

Glue machine maintenance (on an OEM designed machine) isn't fun. High temperatures, and naturally, sticky.

Send the machine a digital signal when the head should be extruding.

So the pump is approximately 120 ft from the tank it is pumping to. The positive displacement pump is sending glue to a 100 gallon tank on the line. The glue is then pumped from there to a glue curtain where product runs underneath it. What glue is not falling on the product is recirculated by gravity back to the tank. There is no flow control on the line to the “day tank” as we call it. There is flow control out of the tank, which varies dependent on how much glue we need on the product. The issue is that this glue is really tough to deal with in cold temperatures and if it is not moving it thickens. That is why we have a setpoint, or sweet spot I should say we are desiring for line pressure. If the pressure is too high, The tank level is set to 60% because they want to keep it from sitting for too long. I am trying to come up with a way to keep it in our sweet spot without having an operator constantly checking and changing speed manually. This is a new process and I am trying to work through some of their ongoing issues with this glue system.

If the glue is not at the correct temperature to flow properly, then I would suggest looking at some sort of wrap for the tank to maintain the tank contents within a certain temperature range. This would help with the flow rates and is not really that expensive. I have used these on different tanks in the past to help keep the contents flowing.

EHinson4 said:

... What glue is not falling on the product is recirculated by gravity back to the tank. ...

Click to expand...

So the 100-gallon tank is at (or near) atmospheric pressure (i.e. 0 PSIG). And the target pressure (setpoint), which is the outlet pressure of the pump, is effectively the pressure difference to get a certain flowrate through 120ft of pipe.

Also, I don't understand this:

If the pressure is too high, [???.] The tank level is set to 60% because they want to keep it from sitting for too long.

Click to expand...

Is there something missing at the blue question marks?


Does the sweet spot have something to do with the non-Netownian nature of the glue (fluid)?

Drbitboy,
Yes I apologize that part must’ve gotten deleted. From what we’ve experienced, high pressure means high viscosity or a blockage which means the pump is essentially deadheading. This is causing the glue to lock up the pumps and ruin the seals, bearings, etc.
The glue is Newtonian, but it starts curing when it is not moving. Also when it’s cold the viscosity is much higher. When it is running through the glue curtain head it is about 85°F.

Ron,
After the tank it passes through a heat exchanger, however a small percentage of the outgoing glue does recirculate so wrapping the tank would be a good idea to pitch.

I don't think what is desired in the OP is possible.


To maintain a level in the tank, the pump flow needs to equal the curtain flow. And whatever that flow is, that sets the pump outlet pressure (at the upstream end of the 120ft of pipe).


The only way I see to get a particular outlet pressure would be during the "on" half bang-bang level control: run the pump at the required speed to maintain that pressure when tank level is low; stop the pump when tank level is high. But then the glue would be static in the pipe some of the time, at which time it would cure, starting at the walls and working its way in, and eventually plug the pipe. However, long before that the unblocked flow area would be small enough that the setpoint output pressure would not be enough to raise the low level in the tank.


Also,

• For a system like this there should be a rupture disk at the outlet of the pump to protect the it.
• Would heat-wrapping the 120ft of pipe make sense, or would that accelerate the curing process? Maybe only apply heat on cold days, e.g. to control the temperature of the glue entering the tank?

Drbitboy,
I appreciate your information and help. I will look for other ways to help this process. Heat wrapping the lines is on our to do list, along with adding a way to keep lines flushed when not in use.

the_msp said:

Don't re-invent the wheel.

Buy a machine e.g. https://www.nordson.com/en/divisions/adhesive-dispensing-systems

Glue machine maintenance (on an OEM designed machine) isn't fun. High temperatures, and naturally, sticky.

Send the machine a digital signal when the head should be extruding.

Click to expand...

You beat me to it.
My one word answer would have been NORDSON as well.
This is all they do....why reinvent the wheel just because you think you CAN, or because your boss is directing you to????