零航速减摇鳍自适应主从控制器设计

零航速下鳍上的水动力与鳍角,角速度和角加速度存在记忆非线性关系,同时船舶横摇模型本身亦具有非线性和不确定性.导致常规控制方式无法直接求取控制量,且不能适应变化的模型和海况.本文通过主从控制器来分离减摇鳍系统的输出和输入非线性,对分离后的系统,采用基于径向基网络的逆控制构成主控制器来求取中间控制量,并自适应横摇模型和海况的变化.采用广义回归神经网络来逼近中间控制量到鳍角的映射作为从控制器.仿真结果表明该方法对横摇模型的不确定性具有自适应性,并可提高在变海况和高海况下的减摇效果.

Adaptive master-slave controller for zero-speed fin stabilizer

The relation between the fin hydrodynamics with the fin angle, fin angular velocity and fin angular acceleration is nonlinear in memory in zero-speed condition; meanwhile, the rolling ship model is nonlinear with uncertainties. These cause difficulties in obtaining operation variables directly for conventional control, making the conventional control method inadequate to situations of varying rolling model and sea conditions. To tackle this problem, we employ a masterslave controller to separate the output nonlinearity from the input nonlinearity of the fin stabilizer control system. In the separated model, we make use of the inverse control based on the radial base function (RBF) neural network to build a master controller to obtain intermediate control variables for adapting the varying rolling ship model and sea conditions. A slave controller based on generalized regression neural network (GRNN) is developed to approximate the mapping from intermediate variables to the fin angle. Simulation results show that this method has self-adaptability to the uncertainties of the rolling ship model and improves the results of anti-rolling in varying sea conditions and high sea conditions.