Dynamic Positioning Systems: Comparison Between Wave Filtering Algorithms and Their Influence on Performance
Wave filtering is an essential function of a Dynamic Positioning System, being responsible for the separation between high-frequency wave induced motions and low-frequency ones, which must be controlled. Low attenuation of the first components may cause oscillatory control action, high fuel consumption and can damage propeller systems. On the other hand, depending on filtering design, high levels of attenuation may be associated with non-admissible delay-times, which may cause instability in the closed-loop system. Traditionally, low-pass or notch-type filters have been used since the first DP systems, due to the simplicity, acceptable performance and possibility of being implemented using analog circuits. Alternatively, observers based on Kalman Filtering Theory have also been used, based on simplified ship models, separating low and high frequency motions. Two wave filters, representing each of those categories, were implemented in a complete DP simulator. All DP components and algorithms are considered and modeled in the simulator, including propellers, thrust allocation, wind-feedforward and different control strategies. Environmental loads are evaluated using fully validated models, including wind, current and wave actions. A pipe-laying barge under typical Campos Basin environmental conditions has been considered as example. Several aspects of the filtering algorithms were analyzed and compared, involving the influence of each filter in the overall DP performance, relationship of design parameters with physical system, ease of commissioning and tuning. The trade off between low frequency tracking (which is associated with the delay time) and wave response suppression is analyzed and discussed for both categories of filters.