Emerson Rosemount 56, Инструкция по эксплуатации

Страница: 68 / 136

Section 7: PID Control Instruction Manual
April 2017 LIQ-MAN-56
58 PID Control
7.1.4 Proportional Plus Integral (Reset)
For the automatic elimination of deviation, Integral mode, also referred to as Reset, is used.
The proportional function is modified by the addition of automatic reset, rather than a
constant Bias value. With the reset mode, the controller continues to change its output until
the deviation between measurement and set point is eliminated.
The action of the reset mode depends on the overall gain. The rate at which it changes the
controller output is based on the overall gain band size and the reset time (I). The reset time
is the time required for the reset mode to repeat the proportional action once. It is expressed
as seconds per repeat, adjustable from 0-3,000 seconds.
The reset mode repeats the proportional action as long as an offset from the set point exists.
Reset action is cumulative. The longer the offset exists, the more the output signal is
increased. If the PV overshoots the setpoint, the reset action will decrease. When the
measurement reaches the setpoint and the proportional control action becomes zero, there
will be an accumulated integral control action to keep the process at the setpoint.
The controller configured with reset continues to change until there is no offset. If the offset
persists, the reset action eventually drives the controller output to its 100% limit - a condition
known as "reset windup".
Once the controller is "wound up", the deviation must be eliminated or redirected before the
controller can unwind and resume control of the measured variable. The integral time can be
cleared and the "windup" condition quickly eliminated by manually overriding the 56 analog
output using the manual mode to reduce the control output and then setting the reset time
to 0 seconds to make integral control action 0%. The reset time can then be changed to
avoid reset windup.
The proportional plus integral control output is given below. Note that the larger the reset
time (I), the slower the integral response will be:
% Output = [Gain
×
100 / (URV – LRV)]
×
[(PV – SP) + 1/I
Σ
(PV
t
– SP)
Δ
t]
Fig. 7-6 Reverse Acting Control with Bias