Closed Loop vs. Open Loop in Flow Question

We're implementing a second pump in a piece of equipment that pumps primarily water, or fluids similar to water. This pump is located in a stand alone piece of equipment that will not always be used. This stand alone piece of equipment can be connected to our main piece, depending on which process the operator needs to run.

If we put the pump downstream of the flow meter, and use that flow meter as feedback, can we call our control "closed-loop"? We also have a second pump upstream of the flow-meter that uses the flow meter as feedback for closed-loop control.

System Diagram
First Controlled Pump -> Flow Meter -> Second Controlled Pump

I ask because I'm under the impression that we cannot call it "closed-loop", and must call it "open-loop". Here are the two arguments..

For Calling it Closed-Loop
1. Flow in "equals" flow out for this pump. See #1 for Open-Loop on why equals is in quotes.

For Calling it Open-Loop
1. Flow in "equals" flow out is a conditional statement. Although this is true for most scenarios, the flow meter will read a different flow rate going into the pump than what is coming out of the second pump. This is especially true since we use closed-loop control on the pump upstream of the flow meter. Other variables that cause flow in not to equal flow out include stuffing pressure to the second pump, and the pressure downstream of the second pump.
2. Most of my control experience is from classes taken during my BSEE, but I was taught it is best practice to measure your feedback directly after your control has been applied. Placing a flow meter before a pump and using it as "feedback" for closed-loop control seems to contradict this.
3. Our second pump's manual recommends using the stuffing pressure as the first pump's feedback source. My opinion is that you cannot claim that both pumps operate under "closed-loop" control if the first pump's adjustments cause the feedback for the second pump to vary, before the second pump's control has a chance to influence the feedback.

I should also say that I've advised adding a second flow meter downstream of the second pump, or having "jumpers" that redirect flow so (1) flow meter can be used downstream of either pump, depending on the process.

Let me know if I didn't explain something well enough, or if you have any questions. Thanks!

I forgot to mention... I've described having a flow meter upstream of the second pump as a way to measure disturbance, rather than feedback.

Is this an accurate statement?

Unless there is some way of accumulating water the flow in and out must be equal. Any mismatch will result in a pressure change between the two pumps. I would run both pumps at the same speed. In theory the flow through the two pumps should be the same.
HOWEVER, I would consider putting a pressure sensor between the two pumps. I would use the changes in pressure to "trim" the flow of the second pump.

See definition(s)

Closed loop may refer to: A feedback loop, often found in: Closed-loop transfer function, where a closed-loop controller may be used. Electronic feedback loops in electronic circuits. PID controller, a commonly used closed-loop controller.

Click to expand...

One type of control system in which the output has no influence or effect on the control action of the input signal is called an Open-loop system. An “open-loop system” is defined by the fact that the output signal or condition is neither measured nor “fed back” for comparison with the input signal or system set point.

Click to expand...

We do have two pressure transducers on the system, one after the first pump and one after the second pump. There is no way of accumulating water between the two pumps either.

The first pump is a progressive cavity and the second consists two (sometimes three) pistons that mixes the product homogeneously before sending it downstream. The pressure within these pistons can be controlled manually.

I'm aware of the definitions, and initially planned on classifying our control as closed loop, but thought about how I'd explain it to a client if they asked. If they asked why the flow meter that provided feedback for our second pump wasn't on the output of the second pump, I'd say because flow in equals flow out. If they said, that's not always true, how can you designate a signal as feedback if it doesn't tell you the flow rate coming out of your second pump.... That's where I started to consider the flow meter not being feedback for the second pump, but rather a way to measure disturbances going into the pump.

In regards to pressure, we also heat/cool water/product between the first and second pump between 280F and room temp.

Both loops are closed loop control of the flow through the flow meter.

Technically your system is open loop from the flow meter, through the second pump, to the downstream process, but as you said, that is a In = Out chain.

if you moved the flow meter downstream, they would still be closed loop control of the flow through the flow meter. You would still have open loop control of the flow through the pipework to your process, but the pipework is an In = Out chain.

For it to be completely open loop control, you would not take into account the flow meter reading.

So, even though a system is technically open loop, you can still classify it as closed loop? The in=out chain seems to only be true when you observe the system over some "x" amount of time. When I look at a block diagram I pulled from Peter's company website, I still have one thing that bugs me..



If I think about an instantaneous snapshot of the system where the operator adjusts the first pump's set-point, the flow meter before the second pump is going to measure a flow rate that is different than the flow rate coming out of the second pump. Also, the second pump needs at least 75 PSI of stuffing pressure, or else cavitation may occur, in which case I'm not sure, but I assume flow in would not equal flow out.

I can see now where a flow meter before the second pump has "technically" allowed the second pump's PID to affect the process and can be classified as feedback, since the effect of the second pump's PID can be measured before or after the pump.

Maybe it just seems like bad practice to do it this way since the goal of the second pump is to provide a specific flow rate at the end of the process. Also considering the amount of variables (adequate stuffing pressure to pump 2 dependent on product viscosity and temperature within the second pump, set-point changes to either pump, external disturbances) that can affect what a single flow meter before the second pump will read, and this reading not matching the flow rate out of the second pump, which is the desired flow rate for the entire process.

Confuse process

Just because you have the equipment to operate in a closed loop does not mean you have a closed loop system. If you ride your bicycle at a leisurely pace, that is open loop. If you are following someone, and control how fast you ride based on your distance from that person, you have closed the loop. If you put a speedometer on your bike and try run 10 mph, then that is closed loop. If don't care how fast you ride, but take note of the speedometer reading, and don't change your speed because of the reading, you are back to open loop.

proof said:

Just because you have the equipment to operate in a closed loop does not mean you have a closed loop system. If you ride your bicycle at a leisurely pace, that is open loop. If you are following someone, and control how fast you ride based on your distance from that person, you have closed the loop. If you put a speedometer on your bike and try run 10 mph, then that is closed loop. If don't care how fast you ride, but take note of the speedometer reading, and don't change your speed because of the reading, you are back to open loop.

Click to expand...

If you're controlling how fast you're moving based on the rotation of your wheels and there is ice on the road causing them to spin, is that still closing the loop if the goal is to control your speed? Especially if another feedback signal is available, like distance from an object.

Yes, measuring the rotation of your wheels will tell you your speed for most instances, but there are some where it will not. That's kind of how I'm currently looking at using a flow meter upstream of the 2nd pump. The end goal is to control how fast flow is coming out of the equipment, downstream of the second pump, yet that metric isn't what is being used as feedback. Yes flow in = flow out for most instances and when observed over time, but based on Mickey's definition he posted... Is the output signal "measured and fed back" to the system with a flow meter upstream of the 2nd pump when it reads a different value than a flow meter downstream of the 2nd pump?

I'm a little confused as you what questions you have. Are you wanting to get the definitions of closed and open loop control? Are you asking about how you should control pumps, or are you trying to understand some basics of physics? Or are you getting confused by semantics?

jethridge said:

The end goal is to control how fast flow is coming out of the equipment, downstream of the second pump, yet that metric isn't what is being used as feedback. Yes flow in = flow out for most instances and when observed over time, but based on Mickey's definition he posted... Is the output signal "measured and fed back" to the system with a flow meter upstream of the 2nd pump when it reads a different value than a flow meter downstream of the 2nd pump?

Click to expand...

It does not matter whether the flow meter is before the pump or after the pump. The flow is the same, so the reading is the same. "Measured and fed back" is not referring to a physical location of the flow meter being after the pump so the reading goes 'back' to the pump. Instead of saying back, say into. The reading from the flow meter goes into (back to) the controller, the signal from the controller goes out to the pumping system (VFD, pump stroke, variable speed drive) adjusting the pumping rate. The flow meter reads the flow. That is the loop.

proof said:

I'm a little confused as you what questions you have. Are you wanting to get the definitions of closed and open loop control? Are you asking about how you should control pumps, or are you trying to understand some basics of physics? Or are you getting confused by semantics?

It does not matter whether the flow meter is before the pump or after the pump. The flow is the same, so the reading is the same. "Measured and fed back" is not referring to a physical location of the flow meter being after the pump so the reading goes 'back' to the pump. Instead of saying back, say into. The reading from the flow meter goes into (back to) the controller, the signal from the controller goes out to the pumping system (VFD, pump stroke, variable speed drive) adjusting the pumping rate. The flow meter reads the flow. That is the loop.

Click to expand...

I wanted to keep it theoretical, but I should've include our controls/physical implementation. The flow meter's signal is fed back to the PLC, the PLC controls the pump's VFD via a PID block w/ PID parameters calculated via Cohen-Coon method. If we use the flow meter as feedback for this PID, it does not always represent the flow out of the pump. We will account for proper stuffing pressures and the like in our control, but as stated in my 3rd post (5th of thread), I wouldn't know what to say to a customer in the rare chance they challenge our implementation with a flow meter upstream of the 2nd pump since flow in = flow out is a conditional statement and only true when the system is operating at predefined conditions and observed over time. This prompted me to create the thread.

In summary, here are the questions...
1. With proper control implementation, can a flow meter upstream of a pump be classified as "closed-loop", despite flow in not always equaling flow out?

E.g. instantaneous snapshot of the system after a set-point change, 2nd pump having inadequate stuffing pressure, etc.

2. Would a better way to classify the control w/ a flow meter upstream of the 2nd pump be: closed-loop only when the system is operating at predefined conditions?

In summary, here are the questions...
1. With proper control implementation, can a flow meter upstream of a pump be classified as "closed-loop", despite flow in not always equaling flow out?

E.g. instantaneous snapshot of the system after a set-point change, 2nd pump having inadequate stuffing pressure, etc.

2. Would a better way to classify the control w/ a flow meter upstream of the 2nd pump be: closed-loop only when the system is operating at predefined conditions?[/QUOTE]

1) Yes. The loop is closed. Even with improper implementation, the loop is closed. You seem to think that having the flow meter ahead of the pump is an improper implementation. I don't see that it is. I would be more concerned with appropriate piping lengths before and after the flow meter than I would about the flow meter location in relation to the pump. You are probably not going to be getting much error in the reading from having pump leaks and etc.

2) No. It is closed loop. Whether you have accurate readings, fast or slow response with the pumping, good or back tuning of the PID, the end result of great control or horrible control, it is still closed loop control. Closed loop control is not the same, and does not imply that you have good control. That is why you see the reports that XX% of the control loops perform better in manual control than in automatic control.

Thank you for the response proof. That was my disconnect.. I assumed the source of feedback would need to be a direct and accurate representation of the flow coming out of the pump to be classified as closed-loop. Chalk that up to lack of real-world experience.

My last question... We are also planning on adding this second pump in another piece of equipment, but this equipment will have a different physical configuration. This will include a holding tank. The system configuration would look like:

Pump 1 -> Flow Meter -> Pressure Transducer -> Tank -> Pressure Transducer -> Pump 2 -> Pressure Transducer

I've proposed that this implementation would require a flow meter after pump 2 to be classified as closed-loop. Is this correct?

For my sanity, as someone who has had a long time since ME classes in college, and little practical experience with pumps and pumping processes:

How is flow in = flow out a conditional statement? Under what conditions could more/less water leave the pump than enter it? Am I misunderstanding the whole conversation?

I guess leakage is one instance, but I would hope that pumps don't leak enough water outside the system for that to be a measurable variable one needs to consider.