This item is available under a Creative Commons License for non-commercial use only
Proportional-integral (PI) or proportional-integral-derivative (PID) controllers are still by far the most common form of automatic feedback control employed in the process industries at present. It is estimated that more than 90% of control loops in the process control industry are of the PID type. Unfortunately, it is a common occurrence, in the majority of industries where this type of control is employed, for the performance of PI/PID controlled processes to be poor. There are a number of factors influencing this poor performance, namely controller tuning problems, controller equipment design restrictions, deficiencies in control strategy design, etc. The research was conducted with the intention of developing an affective strategy for assessing the performance and robustness of closed loop systems controlled by PI and PID controllers. Subsequent to an initial investigation into the most effective performance assessment measures, a Matlab based software tool to automate the evaluation process was developed. This software tool incorporated an identification stage followed by an evaluation stage and was developed using Matlab TM 6.5. A comparison of two distinctive identification techniques was conducted with the intention of identifying the most efficient means of extracting the most constructive information with minimal disturbance to the system under test. The identification techniques investigated included a relay based techniques and a Pseudorandom binary sequence testing scheme. A thorough investigation was carried out into identifying the most efficient means of applying these testing schemes. With respect to the Pseudorandom binary sequence based identification technique, a comparison of a number of technique parameter selection methods was carried out and a guideline for applying a pseudorandom binary test signal was developed. Also, a PI(D) tuning rule database was created. The effectiveness of the overall evaluation strategy, as well as the capabilities of the identification techniques was investigated in order to validate the merit of the strategy developed.
O'Connor, Niall (Thesis), "Performance and robustness metrics for the compensation of processes with time delay" (2005). Masters. Paper 15.