Multi-criteria Loop Quality Assessment: A Large-Scale Industrial Case Study

This paper is concerned with Control Performance Assessment (CPA) of an industrial nitric fertilizer production installation. As many as more than 200 single or cascaded loops based on the PID control algorithm are considered. Effectiveness of various control loop quality measures is compared: integral indexes, factors of different probabilistic density functions and persistence fractal measures are taken into account. Finally, the most informative ones are integrated into a single radar plot being the common platform for comparison. Analysis is accompanied with PID settings analysis showing further tuning directions. As each of the installation elements is in different status, the results enable to point out necessary steps for future plant improvements.