Hydrodynamic Interaction Research & Optimal Thrust Allocation Algorithm Of DP

Hydrodynamic characteristics of DP propulsion system and feasibility analysis of several optimal thrust allocation algorithms are presented in this article. At first, one simple design method of propulsion system is given out in order to carry out farther work, such as DP model test and numerical simulation. And then based on model test and numerical method, it has researched the hydrodynamic characteristics of DP, including thruster-thruster interaction and thruster-hull interaction, aiming for reducing thrust loss and improving efficiency and safety coefficient of the system, providing theory basis for the DP design.It is proved that thrust loss could reach to 45% due to thrust-hull interaction, and 70% due to thruster-thruster interaction. The main factors causing thruster-hull interaction include Coanda Effect, fiction loss, pontoon effect (semi-platform) and so on. The main methods to reduce thrust loss are to tilt duct of thruster or tilt the whole thruster, and to arrange thrusters with reasonable and effective distance. Besides setting forbidden vector boundary in the optimal thrust allocation logic could greatly improve thruster-thruster interaction characteristics as well.Pseudo-inverse method and Sequential Quadratic Programming are two typical algorithms used to solve optimal thrust allocation problems. In this article the timeliness and stability of the logic are illustrated with examples and some typical nonlinear boundaries are given out. Our aim is to enhance the stability, maneuverability and effectiveness of DP. From the results, it is concluded that Pseudo-inverse could be better to gain the required thrusts, which is not only simple, but very quick and reliability. However, there is a evident deficiency of Pseudo-inverse method used before, that is hard to solve nonlinear constraint problem. And usually it required the azimuth angles of thruster are given out as initial values. Given these problems, I give out a method which could take into account much more complex boundaries to optimal azimuth angles. Several numerical methods are presented. As regards SQP, the main advantages are that it could solve effectively much more complex nonlinear constraint problems, and the results are usually more optimal than Pseudo-inverse method. Of course in some conditions the objective function and boundaries are required to be simplified with Linear Fit method in order to ensure the problem always has feasible iteration solution. In this article another example is given out to contract several different boundaries from the aspect of fuel efficiency, with forbidden vector or without. And analyze how to simplify the objective function and boundaries and analyze the feasible of algorithms in the end.