PID controlling 2 three-positional valves, no feedback

[ylle]

Hi,

 

I have to make logic were one PID loop controls two series-connected valves (look File 2). In logic, PID SP is temperature what comes from dynamical graph (depend on the external air temp), PV is outgoing temperature (after valves), PID CV is used to activate PLC relay outputs- OPEN, CLOSE (look File 1)  There is no feedback from valves, from technical data is known that OPEN position to CLOSE position takes 150 sec (both valves are same type).

 

Actions:

1. If its needed to raise water temperature, then its activated OPEN output, which is linked to valve "Direct" (look File 2, 11GM1). When valve "Direct" is full-open, but temp. isnt still high then end- switch Open will be electrically close valve "Bypass".

2. If its needed to lower water temperature, then its activated CLOSE output, which is linked to valve "Bypass" (look File 2, 11GM2). When valve "Bypass" is full-open, but temp. isnt still low then end- switch Open will be electrically close valve "Direct".

 

First positions (physically) are , before PID Auto -tuning 40/60: "Direct" valve start position is 60% opened and "Bypass" valve start position is 40% opened (maybe 50/50 would be better?).

 

I have looked U90 Ladder example "PID with motorized valve". It seemed very helpful, in example primary loop PID is looking SP-PV error and secondary loop compares PID CV and Virtual Valve Position.

In example there is only one valve, how to do it with two valves what are series-connected? OK, in logic OPEN output is increasing "Direct" valve position and CLOSE output is increasing "Bypass" valve position,  but what to do in logic when movement physically goes over to series-connected valve (look algorithm and File 2)?

 

Ylle

 

[Walkerok]

You did not say what model of controller you are intending to use. I tried looking up the model of valves you are using online and could find nothing on them. I have many questions:

1. Is this a closed loop hydronic heating system in which the temperature and pressure changes of the returning water are slower but heat load can be radically different?
2. Is this a consumption loop where you are sending out water to many locations and circulating so that hot water is always close to the outlet points? (large pressure/flow rate changes)
3. Is this a combination of both items 1 and 2?
4. Are the valves you have power to open and power to close without any kind of spring return?
5. Are the valves gate, ball, butterfly, or globe type? (very different flow curves at the same percentage open positions)
6. When you say the valves will be in series are you saying that you want to have a digital open and close contact where the open power feeds open on the main valve and close on the bypass valve and the close signal powers close on the main valve and open on the bypass valve?

Assuming that the valves are power to open and power to close with no spring return then you will be looking to control these valves by bump control where you dedicate close and open outputs from the PLC to both valves independently (4 different outputs). You then will have to write logic to use the PID output to make a decision on how often and how long to apply voltage to either the open or close of either valve to bring your process back into control. I have great reservations in your ability to do this well, or at all, without some sort of position feedback from the valves, but I respect your need and desire to make this work.  If you changed your valves to 4-20ma or 0-10Vdc control then you would not have to have the positioning feedback to try and make this work, but because (I think) you are manipulating valve position by motor power you will have no way to know if you have moved the bypass valve too far or the main valve too little.  When everything is new the motors may travel at pretty much the same speed, but even new motors generally move faster whether closing or opening and will typically have different travel speeds at different points along the open/close of the valve due to different level of resistance to water pressure or the total amount of valve contact to the body of the valve in the different positions (also break away torque requirements will affect everything also).

This level of control theory discussion on what valves and piping structure you should use is way beyond the capabilities of a forum like this to really help you, but we can help with basic PLC component suggestions, generic control thoughts, and specific programming needs. 
 
Best regards
Keith

[ylle]

Hi,

 

This motorized valves are in district heating boilerhouse. Im using Belimo actuators DGR230A-7, with no feedback potensiometer, but with auxiliary switches for series- connection (as it is seen in File 2). Im using Unitronics V1040 and I want to use one PID to control outgoing temperature.

 

1. Is this a closed loop hydronic heating system in which the temperature and pressure changes of the returning water are slower but heat load can be radically different?  - Yes, this is a closed loop heating system, there is a pump which keeps pressure difference between IN&OUT (e.g OUT 3,5 bar, IN 2,5 bar). Returning water to a boilerhouse  is always cooler that outgoing water (e.g returning water is IN=40C, water from boiler 70C, water after mixing OUT=60C).

4. Are the valves you have power to open and power to close without any kind of spring return? - No feedback.

5. Are the valves gate, ball, butterfly, or globe type? (very different flow curves at the same percentage open positions) - Valves are butterfly type.

6. When you say the valves will be in series are you saying that you want to have a digital open and close contact where the open power feeds open on the main valve and close on the bypass valve and the close signal powers close on the main valve and open on the bypass valve? - Yes, as it is seen in File 2.

 

*Actually, all this electrical drawnings and algorithm, these are not final solution, sure ise only that there is no feedback and both valves are having auxiliary switches, actuators are leading butterfly valves, valve positions are manipulated by motor power.

 

 

So, you think I have to use 4 ouputs- 2 for bypass valve and 2 for direct valve (not 2 outputs for both valves)? When valves are having both their own outputs, then I can calculate their virtual positions. If i want raise temperature, I just close little bit bypass and open little bit direct valve. If I want lower temperature, I just open little bit bypass and open little bit direct valve.

 

Ok, if im going to make control of valves to analog which is possible ( relays with  0-10Vdc input), would be there 2 analog outputs, one for bypass valve, one for direct valve? How the logic would be build up? Which analog value would be condition fo close/open relay?

 

 

Greetings,

Ylle

 

[Walkerok]

What model of PLC do you have to use?  Or are you still in the selection process.
 
Here is a substantial document on the selection and use of different types of control valves.

http://www.documentation.emersonprocess.com/groups/public/documents/book/cvh99.pdf

 

Look specifically at page 35 to see the general characteristics of a butterfly valve (and globe valve).  You can see that the first 10% of movement (of an overall arc of 0-90gegrees) there is no flow through a butterfly valve.  There is extremely little flow until you get past 20% of full open and then pretty much maximum flow is achieved at 60% open.  I am trying to tell you that I believe it is important that you have a more precise way to control the actual position of the two valves than by applying voltage to either the open or close inputs of the Belimo motor.

 

It appears that the Belimo would respond to 2-10VDC for a modulating control signal if you buy the Belimo model that has that as a feature.  Depending on overall impedance of the valves you can send one analog voltage signal from the PLC to both valve motors.

 

Keith
 

[lawry]

Since Ylle is using a Vision1040, this topic should be moved away from the M90 / Jazz section