Improved methods for processing position information for dynamic positioning.

Abstract

Dynamic Positioning (DP) has been used on drill ships, semisubmersible platforms, and offshore support vessels since the early 1960's. Beginning in 1977, these control systems have used modern optimal control with various types of Kalman Filtering to extract the low-frequency motions from the noisy position sensor outputs. As oil and gas exploration moves to deeper water, some of the traditional position sensors become useless while others provide significantly less accuracy and reliability. Future DP systems will have to deal with simultaneous, multiple sensors; i.e., monitoring sensor performance, detecting loss of performance, and then making decisions about which measurements to use and how to estimate the low frequency position error from them. The conceptual design of such a "Third Generation" DP system is studied. The fault detection and accommodation concepts are introduced into the DP system. Binary Phase Detection Filters are evaluated for this application. The Binary Phase Detection Filter concept is adapted and extended to a closed-loop stochastic state estimation feedback DP system. The proposed concept is designed for a Gotaverken GVA4000 semisubmersible platform with radar and Global Positioning System (GPS) position inputs, duplicate heading gyros, and a rate gyro. Supervisory logic is developed to monitor the Binary Phase Detection filters which are designed to detect bias and complete inoperation errors in these sensors. Complete system feasibility and performance are evaluated through simulation. The work demonstrates the feasibility of this DP system concept and provides the theoretical and numerical basis for their design.