A Study on Dynamic Positioning System Robustness with Wave loads predictions from Deep Belief Network

Abstract

The dynamic positioning (DP) system is a sophisticated control system which is vital for station keeping, auto-pilot, way-point tracking and other navigation functions. The safe and efficient operation of DP system depends upon the control system, vessel reference system, motion reference and heading measurement sensors. In spite of this state-of-the-art sensor technologies, a control system which is not robust and not stable could jeopardise the performance of the entire DP system. Therefore, robustness of the DP control system is a fundamental basic requirement for any offshore marine activities. The external disturbances that influence the performance and robustness of the control system are due to environmental loads from wind, wave and currents. Till date, these disturbances are treated as bounded constant disturbances. However, these loads are not constant and are much more dynamic in nature. In this work, we use the predicted wave characteristics and wave loads from Deep Belief Networks (DBN) to analyse the robustness of the DP control system. H ∞ robust control system is used to determine the asymptotic internal stability, sensitivity function of the DP closed-loop control system. Only wave loads which are predominant is considered in this work. The simulation results with and without the knowledge on wave loads are examined and it indicates that the robustness of the DP control system has enhanced with the knowledge of dynamic environmental loads.