摘要
为了解决某型涡扇发动机慢车转速以下的数学模型难以建立,无法进行起动性能数值计算的问题,提出了用支持向量机辨识起动模型,依据辨识结果估算起动性能的方法。采用发动机起动试验数据作为学习样本,建立了基于支持向量机的非线性动态起动模型。根据该型发动机起动供油量调整试验得到的供油压力数据,利用所建立的模型对起动性能进行了估算,给出了估算结果与试验数据的对比情况。研究表明,将支持向量机用于起动模型的辨识是可行的,能够较好地解决某型发动机起动性能计算的难题。
In order to solve the problem of simulation for the transient performance of turbofan engine under idle speed,a nonlinear mathematical model of aeroengine starting was presented.The model Support Vector Machines in establishing the identification model of engine starting by using engine ground experimental data as learning samples was employed.To check the generalizing capability of the SVM model, starting characteristics of the turbofan engine was estimated by utilizing the developed mathematical model and making use of the experimental data measured in starting performance adjustment experiments. Comparisons of the results were given.It shows that Support Vector Machines can be effectively applied to identify aeroengine starting mathematical model. The method can solve the difficult problem that turbofan engine faces in starting characteristics calculation.
出处
《推进技术》
EI
CAS
CSCD
北大核心
2004年第5期401-404,480,共5页
Journal of Propulsion Technology
基金
国家自然科学基金(60304004)。
关键词
涡轮风扇发动机
起动
数学模型
支持向量机
仿真
Turbofan engine
Starting
Mathematical model
Support vector machine^+
Emulation
