Performance analysis and troubleshooting of process control loops

Abstract

Typical chemical plants comprise of several hundred to thousand process controllers. The performance of these a??regulatory levela?? controllers must be kept at optimum levels if the benefits of automatic and advanced control are to accrue to the process industry. Even when control loops perform well at the time of commissioning, their performance can deteriorate over time due to changing operating conditions. In such scenarios, and with the abundance of routine operating data, there are good reasons to develop methods and procedures for measuring the performance of control loops and determining the causes for loops exhibiting poor performance. In our research, we propose a detailed data analysis and novel control loop performance assessment framework that would systematically detect, diagnose and suggest rectifications to poorly performing control loops. Results from various simulations and industrial case studies illustrate the usefulness of our methods in successfully improving the performance index of a control loop.