NF-3风洞神经网络自适应稳风速控制系统研制  

The development of wind velocity adaptive control system based on neural networks for NF-3 wind tunnel

作　　者：吕鹏涛[1] 惠增宏[1] 

机构地区：[1]西北工业大学翼型叶栅空气动力学国家重点实验室,西安710072

出　　处：《实验流体力学》2009年第4期82-86,共5页Journal of Experiments in Fluid Mechanics

基　　金：翼型叶栅空气动力学国家级重点实验室基金资助项目(9140C4203020705)

摘　　要：介绍西北工业大学NF-3大型低速直流式翼型风洞改进后的稳风速控制系统的结构、控制原理和性能特点。针对低速直流风洞风速控制系统建模困难,易受外界干扰影响,且自身参数时变不确定,控制难度大的特点,新系统采用了多线程、数字式数据采集方式,带死区控制的PSD神经网络自适应控制算法。通过在NF-3风洞三元试验段和二元试验段的实际应用,与正在使用的模糊控制系统相比较,新系统的适应性和鲁棒性更强,调试时间大为缩短,风速控制精度在风速大于10m/s时从国军标的合格指标0.3%提高到先进指标0.1%,稳定时间减少了10%左右,人机交互也更加友好。This paper describes the structure and the characters of wind velocity control system for NF-3 low speed wind tunnel which is located at Northwestern Polytechnical University. Due to the difficulties of establishing precise mathematical model, the complicated nonlinear characteristics of the system and the change of parameters of the wind tunnel itself, the velocity system is not easy to control in the traditional control theory. In this new system we applied the multithreading digital data acquiring method, PSD neural networks with dead line control algorithm to solve these problems. Compared with .the former fuzzy control systems, this system is not only easier to operate,less time to stable, but also have higher control performance from 0.3 % to 0.1% when the velocity is greater than 10m/s.

关 键 词：风洞 PSD神经网络 多线程 自适应控制 风速 

分 类 号：V211.752[航空宇航科学与技术—航空宇航推进理论与工程]