Determination of switching value
Setting parameters for the function block begins with the determination of switching value trig_err. This parameter determines when the function block is automatically changed from PD to PI operation.
When the absolute value of system deviation ERR = SP - PV is smaller than the switching value trig_err, the controller switches automatically from PD operation into PI operation.
When the system deviation ERR is larger than the switching value trig_err, the controller switches automatically form PI operation into PD operation.
It then follows that:
Each controller type is linked to a parameter set, which must be configured as well. The control algorithm changeover is practically a switch from one parameter set to the other. The changeover is bumpless.
PD controller parameter settings are accomplished by configuration of the proportional action coefficient gain_d and rate time td.
For PD controller operation the D component is delayed by the time constant value td_lag. The td/td_lag ratio is termed the differential gain, and is generally selected between 3 and 10. The D component directly determined by the system deviation ERR, such that for reference variable fluctuations (variations at input SP) a jump attributed to the D component is produced.
The D component can be disabled by setting td = 0.
PI controller parameter settings are accomplished by configuration of the proportional action coefficient gain_i and reset time ti.
In general during the start-up procedure with the PD algorithm, the proportional action coefficient is set considerably higher than in the practically stationary operation with the PI algorithm. This circumstance is conceded to by the designation of two independent proportional action coefficients.
The I component can be disabled by setting ti = 0.
Manipulated variable limiting
The limits ymax and ymin retain the manipulated variable within the prescribed range.
It therefore holds that: ymin ≤ Y ≤ ymax
The outputs qmax and qmin signal that the manipulated variable has reached a limit, and thus been limited:
-
qmax = 1, if Y ≥ ymax
-
qmin = 1, if Y ≤ ymin
For limiting the manipulated variable, the upper limit ymax should be greater than the lower limit ymin.
Should limiting of the manipulated variable take place while the PI control algorithm is active, the anti-windup reset should ensure that the I component "cannot go berserk". Anti-windup measures are taken only for I component values other than 0. Anti-windup limits are identical to those for the manipulated variable.
The anti-windup reset measure corrects the I component such that:
There are three operating mode selectable through the man and halt parameter inputs:
|
Operating mode
|
man
|
halt
|
Meaning
|
|
Automatic
|
0
|
0
|
The manipulated variable output Y is determined through the discrete PI or PD closed-loop control algorithms, based on the controlled variable PV and reference variable SP. The manipulated variable is limited by ymax and ymin. The controller output limits also serve as limits for the antiwindup reset.
|
|
Manual mode
|
1
|
0 or 1
|
The manual manipulated value YMAN is passed on directly to the manipulated variable Y. The manipulated variable is limited by ymax and ymin. Internal variables will be so manipulated, that the controller changeover from manual to automatic can be bumpless.
|
|
Halt
|
0
|
1
|
The manipulated variable remains unchanged, the block does not influence the manipulated variable Y. Internal variables will be manipulated in such a manner that the controller can be driven smoothly from it's current position. Manipulated variable limits and antiwindup measures are as those in automatic mode Halt mode is also useful in allowing an external operator device to adjust control output Y, whereby the controller's internal components are given the chance to continuously react to the external influence.
|
