| 基于L-M网络的涡扇发动机动态过程辨识 |
| |
| 引用本文： | 刘小锋,雷勇.基于L-M网络的涡扇发动机动态过程辨识[J].航空计算技术,2000,30(4):10-15. |
| |
| 作者姓名： | 刘小锋 雷勇 |
| |
| 作者单位： | 中国飞行试验研究院陕西西安 710089 西北工业大学陕西西安 710072 |
| |
| 摘 要： | 介绍了多层前馈网络的Levenberg-Marquardt (LM)算法,对涡扇发动机动态过程的数学模型进行了分析和简化,运用多层前馈网络L-M算法对涡扇发动机进行加力推力变换时的非线性MIMO动态过程的数学模型进行了辨识,辨识所用样本数据和测试数据均为试飞实测数据.辨识结果表明,这种方法收敛快精度高,非常适合于涡扇发动机非线性动态过程的建模.
|
| 关 键 词： | 涡扇发动机 神经网络 Levenberg-Marquardrdt算法 非线性系统辩识 |
| 修稿时间： | 2000年11月8日 |
Dynamical Process Identification of Turbofan Engine Using L-M Feedforward Neural Network |
| |
| LIU Xiao feng,LEI Yong.Dynamical Process Identification of Turbofan Engine Using L-M Feedforward Neural Network[J].Aeronautical Computer Technique,2000,30(4):10-15. |
| |
| Authors: | LIU Xiao feng LEI Yong |
| |
| Abstract: | In this paper,nulti-layer feedforward network using Levenberg-Marquardt algorithm is enployed in nonlinear dynamical process identification of a turbofan engine.L-M algorithm for neural metwork use is briefly introduced.The nonlinear model of the engine is analyzed and simplified.The data used in the paper are from the flight test.It is shown that L-M algorithm is much more efficient than BP algorithm and is very suitable for dynamical process identification of the engine investigated. |
| |
| Keywords: | turbofan engine nueral network Levenberg-Marquardt algorithm nonlinear system identification |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
