基于神经网络的风洞尾支杆减振系统

风洞试验时,由于气流的影响,测试用悬臂式尾支杆容易产生大幅度低频振动,这会严重影响测试精度,甚至损坏自身结构。为了有效抑制尾支杆的振动,本文设计了基于压电组件的主动减振系统,并将人工神经网络应用于PID控制,提出了神经网络PID智能控制算法。对尾支杆进行有限元分析,获取其模态参数。然后设计试验测试减振系统的性能,将神经网络PID与经典PID的控制效果进行对比。试验结果表明:在连续载荷的作用下,采用经典PID控制算法与神经网络PID均可达到有效控制(减振幅度70%以上),且神经网络PID在保证减振效果的情况下实现控制参数自整定,具有良好的鲁棒性。

Damping System for Sting Used in Wind Tunnel Based on Neural Network

In the wind tunnel tests, due to the influence of airflow, large-amplitude and low-frequency vibration is easily produced on the cantilever sting used for testing, which would seriously affect the accuracy of tests and even destroy the structure. In order to effectively reduce the vibration of the sting, this paper designs an active damping system based on piezoelectric components and applies the artificial neural network to PID control, then proposes a neural network PID (NNPID) intelligent control algorithm. The sting is analyzed by the finite element method, and its modal parameters are obtained. Meanwhile, experiments are carried out to test the performance of the damping system and the effects of NNPID and general PID algorithm are compared. Results indicate that under continuous loads, the general PID control and NNPID both have good performance (over 70% amplitude of vibration reduced) in controlling the first modal vibration of the structure. Furthermore, NNPID achieves the goal of the self-adjusting of parameters under the condition of ensuring the damping effect, and possesses good robustness.