Using PID basic functions
7.3 TCONT_CP
PID control
Function Manual, 03/2017, A5E35300227-AC
223
In phase 2, the process value attempts to detect the point of inflection with a constant
manipulated variable. This method prevents the point of inflection from being found too early
as a result of process variable noise.
With the pulse controller, the process variable is averaged over N pulse cycles and then
made available to the controller stage. There is a further averaging of the process variable in
the controller stage: Initially, this averaging is inactive; in other words, averaging always
takes place over 1 cycle. As long as the noise exceeds a certain level, the number of cycles
is doubled.
The period and amplitude of the noise are calculated. The search for the point of inflection is
canceled and phase 2 is exited only when the gradient is always smaller than the maximum
rise during the estimated period. TU and T_P_INF are, however, calculated at the actual
point of inflection.
Tuning, however, is only ended when the following two conditions are met:
1. The process value is more than 2*NOISE_PV away from the point of inflection.
2. The process value has exceeded the point of inflection by 20%.
Note
When exciting the process using a setpoint step change, tuning is ended at the latest
when the process value exceeds 75% of the setpoint step change (SP_INT
-PV0) (see
The phases 3, 4 and 5 last 1 cycle each.
In Phase 3, the valid PI/PID parameters are saved before the optimization and the process
parameter is calculated.
In Phase 4, the new PI/PID parameters are calculated.
In Phase 5, the new manipulated variable is calculated and the controlled system is given.
The process type is inspected in Phase 7, because TCONT_CP always changes to
automatic mode after optimization. The automatic mode starts with LMN = LMN0 +
0.75*TUN_DLMN as a manipulated variable. The testing of the process type occurs
with the recently recalculated controller parameters and ends at the latest
0.35*TA (equilibrium time) after the point of inflection. If the process order deviates strongly
from the estimated value, the controller parameters are newly calculated and STATUS_D is
counted up by 1, otherwise, the controller parameters remain unchanged.
Then the optimization mode is complete and TCONT_CP is back in PHASE = 0. At the
STATUS_H parameter, you can identify whether the tuning was successfully completed.
Premature cancellation of the optimization
In Phase 1, 2 or 3, you can cancel the optimization by resetting TUN_ON = FALSE without
calculating new parameters. The controller starts in the automatic mode with LMN = LMN0 +
TUN_DLMN. If the controller was in manual mode before the tuning, the old manual
manipulated variable is output.
If the tuning is canceled in Phase 4, 5 or 7 with TUN_ON = FALSE, the determined
controlled parameters are contained until then.
To Next Page
Loading...
