[email protected]
Member
Hi Guys,
Looking through this forum it seems I'm in the right place to ask this kind of questions because some of you have much more experience than me with PID control.
I have an AB ML1400 and a valve that I need to control with a PID loop. So far I know that the valve has a slow response time and to open it from 0% to 100% it would take about 6 sec. The requests in the valve control are to command the valve to open to 10% before to start the actual control, and very important, I need the CV variation to be really slow (can't accept fast variations of the CV).
The PV in my loop is very sensitive and seems to be an instantaneous value.
The way that I'm thinking to control it is to have a PID instruction in an STI interruption and to have some filter on the PV (about 4 or 5 sec low pass filter).
I not quite sure what would be an appropriate PID update time (time between interruptions) and I'm thinking to have something like 0.5 or 1sec. Would that be right?
To have a very slow variations of the CV probably very slow PI coefficients would help, along with some "deadband" settings.
One other thing I'm not sure is the "Feed Forward Bias" setting. Can this help me to open the valve to 10% and to start the control from here? If not may be somebody would help me to understand what is this for.
From AB documentation about the gains, Kc, Ti & Td it seems that the lower integral gain is the higher variable value, respectively 32767. So, when I setup my loop in order to have a low integral gain I need to have for Ti a value pretty close to 32767? I just need a confirmation for this to make sure not doing the right thing there.
I really appreciate your thoughts on this matter,
Adi
Looking through this forum it seems I'm in the right place to ask this kind of questions because some of you have much more experience than me with PID control.
I have an AB ML1400 and a valve that I need to control with a PID loop. So far I know that the valve has a slow response time and to open it from 0% to 100% it would take about 6 sec. The requests in the valve control are to command the valve to open to 10% before to start the actual control, and very important, I need the CV variation to be really slow (can't accept fast variations of the CV).
The PV in my loop is very sensitive and seems to be an instantaneous value.
The way that I'm thinking to control it is to have a PID instruction in an STI interruption and to have some filter on the PV (about 4 or 5 sec low pass filter).
I not quite sure what would be an appropriate PID update time (time between interruptions) and I'm thinking to have something like 0.5 or 1sec. Would that be right?
To have a very slow variations of the CV probably very slow PI coefficients would help, along with some "deadband" settings.
One other thing I'm not sure is the "Feed Forward Bias" setting. Can this help me to open the valve to 10% and to start the control from here? If not may be somebody would help me to understand what is this for.
From AB documentation about the gains, Kc, Ti & Td it seems that the lower integral gain is the higher variable value, respectively 32767. So, when I setup my loop in order to have a low integral gain I need to have for Ti a value pretty close to 32767? I just need a confirmation for this to make sure not doing the right thing there.
I really appreciate your thoughts on this matter,
Adi