A Framework for the Control Room of the Future: Human-in-the-loop MPC

A general framework and preliminary results for the Control Room of the Future are presented. The inclusion of latest technology (e.g., natural language processing, facial recognition, big-data analytics) in control rooms, Model Predictive Control (MPC), and integration of human-in-the-loop is proposed. The paper focuses on process start-ups and uses a two-tank interacting nonlinear system to demonstrate the dynamic changes in the network topology of the plant. Three sets of simulations are conducted to highlight the MPC performance under human inputs to the system. A multiple-model MPC approach is used to solve the nonlinear control problem and the benefits of using it in human-in-the-loop systems are discussed. Future work includes using the network formulation as a supervisory fault detection layer, providing suggestions to the personnel via the Smart Control Room and implementing advanced models for predicting human decisions for safe process start-ups and transient operations. (C) 2019, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.