82 ◾  Distillation Control, Optimization, and Tuning
The two main types of disturbances introduced into a controller for closed loop tuning are:
1. A step change in controller setpoint
2. A step change in load
A step change in load can be simulated by first placing a controller in manual output; second, introducing a step change
in the controller output (valve position); and third, placing the
controller back in automatic.
9.2 Ultimate Gain Tuning of Control Loops
The Ziegler–Nichols 1 closed-loop tuning method was published in 1942 and has become a classic method of tuning control loops. Many suppliers of controllers provide the
Ziegler–Nichols method of tuning in the user guide for their
controllers. The method consists of first turning off the integral
(reset) and derivative (rate) action in the controller and, second, finding the ultimate proportional gain (UG) that will just
sustain continuous cycling of the process variable and measuring the ultimate peak-to-peak time period (UTP). These two
values (UG and UTP) characterize the closed-loop response.
When only P and I (proportional and integral) actions are
used in a controller, the recommended P (proportional gain)
is a fraction of the ultimate gain, 0.45 UG. Some controllers
use proportional band, which is 100 divided by the proportional gain.
The value of I (integral action) is set by T i (integral time
constant) in minutes in some controllers, or the reciprocal, 1/T i , as reset rate in repeats/minute, in other controllers.
Ziegler and Nichols recommended the value for integral time
constant to be set as a fraction of the ultimate peak-to-peak
time period, T i = 0.83 UTP, in units such as minutes. The
controller action from the proportional gain times error will