航空学报 > 2005, Vol. 26 Issue (1): 32-35

基于支持向量机的涡扇发动机起动性能估算研究

李应红, 刘建勋   

  1. 空军工程大学 工程学院, 陕西 西安 710038
  • 收稿日期:2003-12-29 修回日期:2004-06-28 出版日期:2005-02-25 发布日期:2005-02-25

Turbofan Engine Starting Characteristics Estimation Based on Support Vector Machines

LI Ying-hong, LIU Jian-xun   

  1. Engineering Institute, Air Force Engineering University, Xi'an 710038, China
  • Received:2003-12-29 Revised:2004-06-28 Online:2005-02-25 Published:2005-02-25

摘要: 针对不同大气条件尤其是高原和高、低温条件,及不同起动机和负载扭矩特性下涡扇发动机起动过程的数值模拟问题,给出了通用起动模型状态空间描述形式,研究了基于支持向量机的起动模型辨识和起动性能仿真方法。以某型涡扇发动机为例,给定某些条件下的起动试验数据,采用模型逐次递推的方式,估算了其它条件下的起动性能。提出的方法能够保证估算结果的准确度,适合于工程应用。

关键词: 航空, 航天推进系统, 涡扇发动机, 支持向量机, 起动, 数学模型, 起动性能

Abstract: To simulate the starting characteristics of turbofan engine in different atmospheric environments, particularly in plateau regions and at high or low temperature, a generalized description of engine starting mathematical model based on state space is given. The model considers the atmospheric condition, the starter and the load torque characteristics. A method for developing starting model via Support Vector Machines (SVM) is presented, and simulation techniques of starting characteristics are suggested and investigated. For a certain turbofan with given experimental data of engine starting at some temperature, the starting characteristics at other temperature are estimated by utilizing the recursive relations of the identified SVM model itself. The results show that the method can assure the estimation accuracy. It is fit for practical application.

Key words: aerospace propulsion system, turbofan, SVM, start, mathematical model, starting characteristics

中图分类号: