北京航空航天大学学报 ›› 2011, Vol. 37 ›› Issue (3): 283-288.

• 论文 • 上一篇    下一篇

粒子群优化在直升机旋翼动平衡调整中的应用

刘红梅1, 吕琛1, 欧阳平超2, 王少萍3   

  1. 1. 北京航空航天大学 可靠性与系统工程学院, 北京 100191;
    2. 中国空间技术研究院 总体部 北京100186;
    3. 北京航空航天大学 自动化科学与电气工程学院, 北京 100191
  • 收稿日期:2010-01-11 出版日期:2011-03-31 发布日期:2011-04-01
  • 作者简介:刘红梅(1978-),女,辽宁沈阳人,讲师,liuhongmei@buaa.edu.cn.
  • 基金资助:

    国家自然科学基金资助项目(61074083,50705005); 国防科技工业技术基础科研项目(Z132010B004)

Helicopter rotor tuning based on neural network and particle swarm optimization

Liu Hongmei1, Lü Chen1, Ouyang Pingchao2, Wang Shaoping3   

  1. 1. School of Reliability and Systems Engineering, Beijing University of Aeronautics and Astronautics, Beijing 100191, China;
    2. China Academy of Space Technology System Department, Beijing 100186, China;
    3. School of Automation Science and Electrical Engineering, Beijing University of Aeronautics and Astronautics, Beijing 100191, China
  • Received:2010-01-11 Online:2011-03-31 Published:2011-04-01

摘要: 传统的直升机旋翼调整方法没有考虑调整参数与振动信号之间的非线性关系,针对这一缺点,提出将广义回归神经网络(GRNN,General Regression Neural Network)和粒子群算法相结合的旋翼调整方法,采用GRNN网络建立旋翼动平衡调整模型,以桨叶的调整参数作为神经网络的输入,以旋翼转轴和机身的三向的加速度测量值作为网络输出,建立调整参数与直升机振动信号间的模型.以直升机振动作为目标函数,采用粒子群优化算法对桨叶的调整参数进行寻优,获得当直升机振动最小时的桨叶的调整量. 飞行实验结果表明,此方法可通过飞行测试获得的新数据对神经网络进行更新,使系统在使用过程中不断完善,并可在较少的飞行调整下完成旋翼的动平衡调整.

Abstract: Considering the drawbacks of traditional rotor adjustment method without calculating possible nonlinear between rotor adjustments and fuselage vibration signals of the helicopter, a new rotor adjustment method based on the general regression neural network (GRNN) and the particle swarm optimization (PSO) was presented. GRNN network was employed to model the relationship of the rotor adjustment parameters and the fuselage vibrations, whose input parameters are rotor adjustment parameters and whose outputs are acceleration measurements along the three axes of rotor shaft and the fuselage. With the helicopter vibration as an objective function, the PSO was used to make a global optimization to find the suitable rotor adjustments corresponding to the minimum vibrations. Flight test results indicate that the neural networks are easily updated if new data becomes available thus allowing the system to evolve and mature in the course of its use.

中图分类号: 


版权所有 © 《北京航空航天大学学报》编辑部
通讯地址:北京市海淀区学院路37号 北京航空航天大学学报编辑部 邮编:100191 E-mail:jbuaa@buaa.edu.cn
本系统由北京玛格泰克科技发展有限公司设计开发