Abstract
Dynamic Positioning (DP) systems play a crucial role in oil and gas drilling and production floaters used globally for deep-water operations. Drilling operations need to maintain automatic positioning of the platform in the horizontal-plane within the safe zone. Operating DP systems typically require highly responsive control systems when encountering prevailing weather conditions. However, DP incident analysis demonstrates that control and thruster failures have been the leading causes of accidents for the past two decades, according to the International Marine Contractors Association (IMCA). In this paper, a Predictive Neural Network (PNN) strategy is proposed for thruster allocation on a platform; it has been developed by predicting the platform response and training the network to transform the required force commands from a nonlinear Proportional Integral Derivative (PID) motion controller for each thruster. The strategy is developed for increasing safety and zone keeping of DP-assisted-drilling operations in harsh weather. This is done by allowing the platform to recover the position more rapidly whilst decreasing the risk of losing the platform position and heading, which can lead to catastrophic damage. The operational performance of the DP system on a drilling platform subjected to the North Sea real environmental conditions of wind, currents and waves, is simulated with the model incorporating the PNN control algorithm, which deals with dynamic uncertainties, into the unstable conventional PID control system for a current drilling semi-submersible model. The simulation results demonstrate the improvement in DP accuracy and robustness for the semi-submersible drilling platform positioning and performance using the PNN strategy.
Advanced Intelligent Control Strategy in Dynamic Positioning (DP) System Applied to a Semi-Submersible Drilling Platform in the North Sea
