Evaluation of the performance of a self-tuning controller: Project study

• Main Authors: Po, Crispin, Pua, Henry S., Sotto, Rodrigo V.
• Format: text
• Language: English
• Published: Animo Repository 1987
• Subjects: Automatic control; Automatic control > Data processing; Digital control systems; Chemical process control; Chemical Engineering
• Online Access: https://animorepository.dlsu.edu.ph/etd_honors/27
• Institution: De La Salle University

Feedback control is an essential part of chemical engineering. Over the years, new and better technology has been introduced in the design of industrial feedback controllers. The past decade saw the implementation of digital controllers in many areas of chemical process industries because of the advantages which the former offer over conventional pneumatic and analog counterparts. More recently, engineers have started to use the microcomputer for actual process control: Electronics signals are fed into a personal computer through the use of an analog-to-digital converter, and these signals are manipulated by the computer and sent back to the process through a digital-to-analog converter. The introduction of microcomputer-based controllers has made possible the inclusion of self- tuning as part of the controller software. Self- tuning involves getting the initial optimal controller settings. This study develops a PID-controller program to which three tuning algorithms --- Ziegler-Nicolas, Cohen-Coon, and Shinskey --- are individually attached as subroutines. The performance of each tuning algorithm is tested on two analog RC circuits. Results show that the Ziegler-Nichols algorithm is suited for processes that require fast response time but can tolerate high overshoot the Shinskey algorithm, for processes that require low overshoot and the Cohen-Coon algorithm, for general cases of control.