基于广义预测控制的主动吸收造波

Active Absorption Wave-Making Using Generalized Predictive Control

通过最小二乘法支持向量机对主动吸收造波模型进行在线辨识,基于广义预测控制算法提出一种新颖的主动吸收造波策略.采用高阶边界元法与自由面混合欧拉拉格朗日模型建立二维推板造波数值波浪水槽,将预测控制器的输出作为造波边界条件,进行了不规则波主动吸收造波仿真对比试验与白噪声波谱的验证试验,并采用波浪入、反射分离算法进行分析.结果表明:预测控制器具有较好的鲁棒性和快速响应能力,对于模型试验中常用的波浪具有良好的主动吸收造波效果.

An active absorption wave-making model was identified online with least squares support vector machines （LS-SVM）, and a novel method was proposed based on the strategy generalized prediction con trol algorithm. A piston-type 2D numerical wave flume was developed with the higher-order boundary element method and mixed Eulerian-Lagrangian free surface condition. Meanwhile, the predictive controller was embedded in it as wave-making boundary. An absorbing wave-making contrast simulation for irregular wave and a white noise verification test were conducted in numerical wave flume. Using the separating incident and reflected irregular waves method, the test data was also analyzed. Simulation and verification test results verify that the controller has a strong robusticity and quick response ability, and this system has good active absorbing effect for waves commonly used in model tests.