
3Ć9 Theory of Operation: Point Processing
Original Ć June 1990 CE4.2:CL6211
The P/PD algorithm will also allow bias and transfer ramping bias to be
added. Refer to section NO TAG on page NO TAG. In this case, the final
equations become:
IVPĂ +Ă K(SP * DN) ) LOOPĂBIAS ) TBIAS
3.3.1.5 Proportional-Integral-Derivative
The Proportional-Integral-Derivative (PID) PCA provides one, two, or
three mode control capability based upon a positional control algorithm.
The PID PCA allows the proportional and derivative terms to be tuned to
zero, allowing the point to function as a PI, a PID, or an I-only controller.
Like the proportional-derivative with bias PCA, the PID control algorithm
can be viewed as a dual input, single output math function. The point's
process variable value is used as one input, the set point is the second
input, with the resulting output of the math function being the IVP of the
point.
The PID algorithm is based on the following transfer function:
IVP(s)
error(s)
Ă +Ă
K(T
i
Ăs ) 1)(T
d
Ăs ) 1)
T
i
Ăs(aĂT
d
Ăs ) 1)
where: T
i
= integral time constant
T
d
= rate time constant
a = rate action limiter; 0.125
K = proportional gain
By using an intermediate term, DN(s), the transfer function can be
rearranged to:
DN(s)
PV(s)
Ă +Ă
T
d
Ăs ) 1
aĂT
d
Ăs ) 1
IVP(s)
SP(s) * DN(s)
Ă +Ă
K(T
i
Ăs ) 1)
T
i
Ăs
The equations can then be represented in the time domain as:
DN(s)Ă +Ă T
d
dPV
dt
) PV * aĂT
d
dDN
dt
IVPĂ +Ă K[(SP * DN) )
1
T
i
ŕ
R
0
(SP * DN)dt]
Kommentare zu diesen Handbüchern