Model-based control designs for offshore hydraulic winch systems

Abstract Model-based control designs have an advantage in controlling complex systems, and in assuring global system stability, in comparison to ordinary linear controllers such as PID controllers. This paper attempts to promote the use of model-based control designs in industrial applications, particularly in the marine system industry. A simple pendulum system is used as an example help to compare a backstepping sliding mode controller with a PID controller, and to display the main steps in deriving such a model based control design in a consistent manner, and to highlight the benefits using model-based control laws in comparison to standard PID based control laws. The results from this example indicate that the model based controller is barely affected by changes in system parameters in comparison to the PID controller which must be retuned each time the system parameters are changed. This indicates that a lookup table of controller gains and interpolation is needed if a PID controller is to be used for controlling a system with changing system parameters and/or highly nonlinear dynamics. Furthermore, an offshore hydraulic winch system is used as an in-depth case study in which the system is thoroughly analyzed. Based on simple fluid dynamics, mechanics and dynamics in general, a simplified state space model, but a highly nonlinear model that only describes the main characteristics of the hydraulic system is developed and used in the derivation of a speed controller and a torque controller for the hydraulic winch system. A process plant model describing the hydraulic winch system, together with a lumped wire-load model, is used in simulations to test and verify the derived control laws, which seem to be well-suited for controlling the hydraulic system. Based on the observations and results obtained in this work, it is indicated that a PID controller would not be able to control the hydraulic winch system with a large stability region, as well as a model-based controller, if a huge amount of work was not done in tuning and making lookup tables and interpolation algorithms for controller gains. This argues for the use of model-based control designs in industrial applications containing nonlinear dynamics and/or varying system parameters.