Non-linear Model Predictive Control of Conically Shaped Liquid Storage Tanks

TL;DR

This paper presents a non-linear model predictive control approach for managing liquid levels in conically shaped storage tanks, demonstrating improved tracking and control performance over traditional linear methods through simulation.

Contribution

It introduces a non-linear MPC strategy based on a discretized dynamical model for conical tanks, showing enhanced control accuracy compared to linear MPC.

Findings

NMPC provides better level tracking.

Simulation confirms increased control performance.

Compared with linear MPC, NMPC offers improved accuracy.

Abstract

This paper deals with the analysis and synthesis of a model predictive control (MPC) strategy used in connection with level control in conically shaped industrial liquid storage tanks. The MPC is based on a dynamical non-linear model describing the changes of the liquid level with respect to changes in the inlet flow of the liquid. An Euler discretization of the dynamical system is exploited to transform the continuous time dynamics to its discrete time counterpart, used in the non-linear MPC (NMPC) synthesis. By means of a simulation case study will be shown, that NMPC better tracks the changes of the liquid level, hence provides increased control performance. This paper also compares the traditional approach of optimal control, the linear MPC, with the NMPC strategy.