Model predictive control of an overhead crane system

Abstract

This project explores and studies the application of Model Predictive Control (MPC) to an overhead crane system. Overhead crane is a machine used in industrial site such as construction and manufacturing site, in order to move hazardous materials or heavy loads from current place to another desired location. While transporting the payload to desired place, the payload oscillation must be minimized for a safety operation. This makes an overhead crane to be an under-actuated system that need to control more process variables with less manipulated variables. Due to this complexity of the dynamical system, it is very challenging to reduce or eliminate the payload swing angle during the trolley positioning. In addition, constraint is needed to be concerned when designing a controller for the overhead crane system. Therefore, MPC which has the advantage of dealing with constraint, is proposed to have more precise trolley positioning and low payload oscillation during the crane motion. The project starts by deriving the mathematical model using the Euler-Lagrange equation of an overhead crane system. Then, Optimal Predictive Control (OPC), which is one of the type of MPC, was selected for MPC design and then applied to the overhead crane system in a simulation. The result shows that all constraints were satisfied when the overhead crane system was controlled by using OPC. Subsequently, the MPC design was implemented on a laboratory scale crane to investigate the real-time implementation and performance of the controller. The result shows that a desired steady-state value can be achieved in the experiment. However, for transient response, there was a slight deviation for the system responses between the simulation and experiment, which may be due to the deviation between model in simulation and the laboratory crane system.