Predictive monitoring and control

Khan, Mohammad Aminul Islam (2013) Predictive monitoring and control. Masters thesis, Memorial University of Newfoundland.

[English] PDF - Accepted Version Available under License - The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission. Download (6MB)

Abstract

This thesis investigates the effectiveness of model predictions in two important fields of process operation: process monitoring and process control. Both monitoring and control are essential for the safe and optimal operation of a chemical process. Process monitoring is necessary to notify the operator about an undesired condition, whereas a controller directs a process to desired condition. In Part I of the thesis, a novel model-based predictive technique is proposed for early warning generation to the operator. In Part II of the thesis, an in-depth simulation-based investigation is carried out to evaluate t he performance between two control structures: 'model predictive controller (MPC) cascaded to the proportional-integral-derivative (PID) controller', and 'PID-free MPC'. -- The proposed early warning generation method uses open-loop process model and disturbance model. Process feedbacks are used to correct prediction bias. This method exploits t he controller limitations in dealing with process time delay and actuator constraints. An warning in form of alarm is issued if the open-loop predictions exceed the threshold within the process time-delay. Actuators also plays a major role in controlling processes. If at any point actuators do not have sufficient capacity, controller will fail to regulate t he process. Using t he process gain, measurements, and constraint information a linear programming algorithm is used to check t he existence of a feasible solution that meets all constraints. An alarm is generated if no feasible solution can be found (i.e., the actuator does not have sufficient capacity). The proposed method is applied to a single-input-single-output (SISO) and a multi-input-multi-output (MIMO) continuous stirred tank heater (CSTH) system. The method gave early warnings compared to the existing safety systems. Also the method demonstrated robustness under small disturbance in the process. -- The comparative study between the 'MPC-cascaded to PID' and the 'PID-free MPC' structure is carried out on a simulated CSTH system. In the cascaded structure the flow-loops are regulated by the PID controller. On top of that a dynamic matrix controller (DMC) manipulates the set-points of the flow-loops to control tank temperature and level. The 'PID-free MPC' structure uses a dynamic matrix controller (DMC) to manipulate the valve positions directly. The study reveals that the PID-free MPC structure outperforms the cascade structure in both disturbance rejection and set-point tracking. However, the PID-free MPC structure demands more control action and has more computational load. Integrated square error (ISE) is used to quantify the performance.

Actions (login required)

View Item