MPC implementation in a PLC based on Nesterov's fast gradient method

Abstract

This work presents a linear quadratic model predictive controller (MPC) implemented in an industrial programable logic controller (PLC). The control law is calculated by solving on-line the quadratic programming problem derived from the optimization control problem of MPC. Nesterov's fast gradient algorithm has been used to solve the corresponding linear quadratic MPC problems with input constraints. The predictive controller has been implemented in a Schneider Electric M340 using the standard structured language from the IEC 1131.1 norm. The memory and computational time requirements of the proposed implementation have been characterized through extensive simulations. In addition, the properties of the proposed controller have been experimentally demonstrated using a test-bed in which the MPC running in the M340 controls a model of a quadruple-tank process simulated in Matlab/Simulink through an OPC server.