一种逆向估计涡扇发动机主燃油流量的方法

由某型涡扇发动机实时模型获取训练样本,使用支持向量回归机(SVR)的方法,通过离线训练建模,得出了发动机主燃油流量逆向估计模型。训练样本数据包括发动机状态量、输出量以及其它输入量,SVR建模参数则通过交叉验证与网格搜索组合法寻优得到。最后用C++语言进行了计算机仿真验证。结果表明,在飞行包线内的稳态工况或动态工况下,基于SVR的逆向估计模型的输出与发动机主燃油实际输入都基本吻合,且具有较高的精度。

An inverse method of estimating the main fuel-flow of turbo-fan engine

Using support vector regression, an inverse estimation model of engine fuel - flow is built by off - line training. The training samples are got from a real - time turbofan engine model, including engine state data, output data and other input data. The modeling parameters are got by the method of cross validation and grid search. The model is simulated with the C ＋ ＋ language. The results show that the output of the inverse method is basically consistent with the fuel input of engine. This method has high accuracy no matter in steady state or dynamic conditions on the flight envelope.