基于网络拓扑协作复杂系统振动主动控制算法
Active vibration control algorithm for complex systems based on network topology cooperation
-
摘要: 为解决复杂系统振动主动控制中传统分散式算法收敛条件严苛和集中式算法计算量庞大的工程难题,将网络拓扑协作策略与自适应FxLMS算法相结合,设计了基于网络拓扑协作的振动主动控制算法,并选取某直升机简化机身模型作为被控对象,开展了规模为20×20(20个作动器和20个误差传感器)的振动主动控制仿真研究。仿真结果表明:网络拓扑协作算法在显著降低计算量的同时能取得与集中式算法基本一致的减振效果,这是分散式算法和集中式算法所不具备的优势,单频控制时振动均值降幅约34.3 dB,多频控制时振动均值衰减约12.6 dB;同时适当简化了次级通道耦合的特性,有利于收敛系数的取值,充分验证了该算法用于直升机复杂系统振动主动控制的有效性和优越性。Abstract: In order to solve the engineering problems of strict convergence condition of traditional decentralized algorithm and huge computation amount of centralized algorithm in vibration active control of complex systems, this paper combines network topology cooperation strategy and FxLMS algorithm to design a novel active vibration control algorithm based on network topology cooperation, and selects a simplified airframe model of a helicopter as the controlled object. The simulation study of active vibration control with a scale of 20×20 (20 actuators and 20 error sensors) is carried out. The results show that the algorithm based on network topology can achieve the same vibration reduction effect as the centralized algorithm while significantly reducing the computation amount, which is an advantage that the decentralized algorithm and the centralized algorithm do not have. The mean vibration decreases about 34.3 dB under single-frequency control, and about 12.6 dB under multi-frequency control. At the same time, the characteristics of secondary path coupling are properly simplified, which is conducive to the value of convergence coefficient, and the effectiveness and superiority of this algorithm for active vibration control of helicopter complex system are fully verified.