A Self Optimizing Control Framework and A Benchmark for Smart Process Control

Abstract

Industry 4.0 requires introducing smart capabilities into the classic process control that makes the system aware of its current health status, modifying its closed-loop controller parameters or references to ensure the optimal performance of a system under acceptable conditions. This paper presents a Self Optimizing Control (SOC) framework using a Real-Time Globalized Constrain Nelder Mead optimization algorithm supported by the system closed-loop performance specification to control a thermal system. A simulation benchmark is designed to assess the SOC controller performance using a normalized First Order Plus Dead Time model of the thermal system. Obtained results show that the SOC controller can reach the desired closed-loop performance after multiple periodic reference executions of the system.