
SINUS PENTA
PROGRAMMING
INSTRUCTIONS
164/452
22.3.3.
D
E RIV A TIV E
A
CTIO N
(D)
Symbol
Tuning function
Main goal
Td
An input variance (error) generates an output
variance proportional to the variance rate
Decreases the response time for the return to
the tuning point
The derivative action set with Td increases the stability of the system, thus increasing the transient response. The
derivative action tends to get an earlier response, but it increases the system sensitivity to the disturbance overriding
the error signal.
22.3.4.
T
UNING
A
CTIO NS A T
C
ONS TA NT
S
P E E D
When the system is operating at constant speed, the system response shall be the most accurate as possible
(minimum error) and shall adjust any little reference variations.
When at constant speed, if the system does not promptly respond to little reference variations, a shorter integral time
may solve this problem. Otherwise, when little and long-lasting oscillations affect the reference value, setting a longer
integral time could be the right solution.
2 2.4.
Anti-windup
The major benefit of the integral action is to ensure null errors at steady speed. However, just like the derivative action,
the integral action shall be applied with caution to avoid worse performance.
A case in point is the output saturation occurring at the same time as an excessive integral action. When the output
saturates,
the control action is limited, so the error is still remarkable. If the error persists, the actuator will saturate,
because the longer the time the error persists, the stronger the integral action is; this phenomenon is called “windup”.
In case of output saturation, the integral term can reach very high values; as a result, the error shall have opposite sign
for a long period before exiting from saturation.
The PID regulator of the Penta drive is provided with an Anti-windup function which compensates the effect described
above. This Anti-windup action is described below (P=proportional term; I=integral term; D=derivative term).
The output is always calculated as follows:
OUT
P + I + D
When output saturation occurs:
OUT
OUTsat
The integral term is forced based on the following:
I
OUTsat – P – D
(which is the Anti-windup function).
This prevents the integral term from reaching very high values; the integral term is then kept constantly in line with the
saturated output value OUTsat that is present at each moment; any variations of the error (i.e. the P) that allows exiting
from saturation have immediate effect to the output, without having to wait for a long time before discharging the
integral term itself.
The effect of the Anti-windup can be adjusted with parameter
P260
; if
P260
<1, the effect is reduced and the system is
less sensitive to error variations; if
P260
=0, the effect is cancelled.
The value of
P260
=1 is correct for the applications requiring to quickly exit from saturation.
On the other hand, reducing
P260
can be useful when output variations are to be avoided for negligible error variations.