A study on the fault tolerant control of a board machine
Miao Yu,Alexey Zakharov,Sirkka-Liisa Jämsä-Jounela
2015-01-01
Abstract:This paper aims to propose and evaluate the practical schemes of Fault Tolerant Control (FTC) for modern control systems, combining the Model Predictive Control (MPC) and the traditional PI feedback control loops. In contrast to the existing FTC literature, mostly focusing either on MPC or solely on PI loops, the following three cases are studied. Firstly, we consider faults in the measurement of a controlled variable of a PI loop, which is a measured disturbance of an MPC at the same time. Secondly, faults in the measurement of the controlled variable of a PI loop are assumed, which is an input of an MPC at the same time. Thirdly, we study faults in the actuator of the control loop, receiving its setpoint from an MPC. To demonstrate the proposed FTC schemes, the MPC of a large-scale board machine is studied, which controls the key quality variables, the basis weight and the moisture content, by coordinating the operations of the stock preparation and the drying section. Thus, many PI loops of these sections are critical for the successful MPC operations. The proposed FTC system is tested on a board machine simulator model of the Stora Enso’s Kaukopää mills implemented in Advanced Process Simulation Environment (APROS). Three fault scenarios are selected for the studies, including a consistency sensor fault in the stock preparation, a sensor fault in the pressure measurement in the drying group 5, and a valve blockage in the drying group 5.