In the May 2012 issue of Control, an article by F. Greg Shinskey, "The Case Against Lambda Tuning," compares the performances of minimum integrated absolute error (IAE) tuning and Lambda tuning. In the article, it also made a reference to our paper (Chia, Lefkowitz) "Robust PID Tuning Using IMC Technology," InTech, Oct-Nov, 1992.
Shinskey stated that the article by Chia and Lefkowitz encouraged the false assumption that load disturbances enter the loop directly at the controlled variable. That certainly was not our intention.
Our article provided a framework of using internal model-based control (IMC) to come up with PID parameters. Indeed, setpoint changes were used and discussed as an example. In IMC-based tuning, different tuning parameters are recommended for setpoint changes and disturbance rejection.
Shinskey's article compares the performances of two different PI controllers, minimum IAE and Lambda tuning, for an integrating process. He pointed out that the use of integral action will unavoidably cause overshoot. Indeed, in IMC-based tuning, not PI, but a PID control, is recommended as derivative action improves the stability. Even though integral action will contribute to overshoot and oscillation, we continue to recommend integral action, so that setpoint will be tracked properly in the presence of sustained disturbances that pass through the integrating process.