engleski

Optimization of thruster allocation for dynamically positioned marine vessels

Existing strategies and proposed methods for optimal thruster allocation in dynamic positioning systems of marine vessels are primarily focused on minimization of power consumption with handling of numerous constraints and limitations that should be satisfied at the same time. This particularly applies to thruster saturation and forbidden zones that are usually used in order to avoid or reduce thrust losses due to thruster-thruster interactions. By defining a forbidden zone for an azimuth thruster, generally a non-convex thrust region is left over which consequently can cause the significant deviations in allocated thruster azimuth angles and increase the time required for thrust and azimuth angle allocation. On the other hand, one can notice that thruster interaction effects such as axial and transverse current, thruster-hull interaction, thruster-thruster interaction and ventilation are rarely taken into account. These effects, whether they occur separately or in combinations, can cause significant thrust losses, which without appropriate reallocation strategy would consequently lead to degradation of vessel position and heading, as well as to increase of response time and power consumption. In dynamic positioning systems, nonlinear objective functions, and nonlinear equality and inequality constraints within optimal thrust allocation procedures cannot be handled directly by means of the solvers like industry-standardized quadratic programing, at least not without appropriate linearization technique applied, which can be computationally very expensive. Thus, if optimization requirements are strict and problem should be solved for nonlinear objective function with nonlinear equality and inequality constraints, than one should use some appropriate nonlinear optimization technique. In addition, implementation of thruster interactions in optimal thrust allocation can easily transform the optimization problem into nonlinear and non-convex, which can become very demanding. In this context, two optimization approaches have been proposed. The first one is based on the weighted generalized inverse matrix approach, and the other one on the hybrid approach using sequential quadratic programing and direct search algorithms coupled with generalized regression neural network. The latter one is particularly convenient for handling non-convex and nonlinear optimization issues related to generally nonlinear power-thrust relationship and to non-convex nature of forbidden zones and thruster interaction effects. Moreover, both optimization approaches can handle selected thruster interaction effects with negligible increase in total power consumption, compared to the strategy based on forbidden zones only, while minimizing the tear and wear of thrusters at the same time. Obtained results were evaluated with numerous simulations involving different initial conditions and with direct comparison of analysed strategies that include forbidden zones, i.e. thruster-thruster interactions.

dynamic positioning systems ; thruster interactions ; thrust losses ; thruster allocation ; nonlinear optimization

nije evidentirano