I am trying to improve the control of a pdic that is only turned on when things go wrong on a particular product, so it is only active a couple unplanned hours at a time. Any suggestions for using limited historian or Insight data to improve tuning?
Thanks,
Danaca
You can estimate the dead time from the time it takes the PID PV to change direction after a change in PID output direction when the PDIC is active. For a large change in PID output, you can approximate an integrating process gain as the new rate of change minus the old rate of change (including sign) in %/sec divided by the % change in PID output yielding a gain in %/sec/% or 1/sec. You can then use Lambda tuning rules for integrating processes. I suggest a Lambda equal to 3x the dead time to help account for nonlinearities. Even if you cannot estimate the integrating process gain, you can check the reset setting based on just knowing the dead time. Most pressure and differential pressure loops have too much integral action (too small of a reset time in seconds or too large of a reset setting in repeats/minute).
I suggest you try using the auto tuner relay oscillation method for a quick better estimate of dynamics. You can set the change in output to be just 5x the valve dead band, check the integrating process option, and go for just 2 cycles. If the valve dead band is 0.5% (typical for a reasonably good valve), the amplitude in PID output for the relay method can be as small as 2.5%. The relay method should be able to keep the PV near the setpoint for small load upsets while you are running the test but the estimates will be affected. You can manually check the integrating process tuning settings using a lambda equal to the 3x the dead time.
The assumption here is that you have a process where the primary time constant is more than 4x the dead time making it a near integrating process. If this is not the case, the reset time will be too large but this will be in the conservative direction.
Greg
Greg is a retired Senior Fellow from Monsanto-Solutia and an ISA Fellow. Greg was inducted into the Control “Process Automation Hall of Fame” in 2001 and received the ISA Life Achievement Award in 2010.