Convex optimization of thrust allocation in a dynamic positioning simulation system

Vessels conducting dynamic positioning (DP) operations are usually equipped with thruster configurations that enable the generation of resultant force and moment in any direction. These configurations are deliberately redundant in order to reduce the consequences of thruster failures and increase the safety. On such vessels a thrust allocation system must be used to distribute the control actions determined by the DP controller among the thrusters. The optimal allocation of thrusters' settings in DP systems is a problem that can be solved by several convex optimization methods depending on criteria and constraints used. The paper presents linear programming (LP) and quadratic programming (QP) methods adopted in the DP control model which is being developed at the Maritime University of Szczecin for ship simulation purposes.