BP神经网络在涡轴发动机参数换算中的应用

陈江明; 贾锐; 段辉

BP神经网络在涡轴发动机参数换算中的应用

投稿时间：2020-09-11 修订日期：2020-09-18 点此下载全文

引用本文：

摘要点击次数: 114

全文下载次数: 0

作者	单位	邮编
陈江明	中国航发湖南动力机械研究所	410000
贾锐^*	中国航发湖南动力机械研究所	412002
段辉	中国航发湖南动力机械研究所

中文摘要:带自由涡轮的涡轴发动机的参数换算关系难以表达为具体的函数形式。将神经网络应用于发动机参数换算，针对功率、燃气温度、燃油流量和空气流量换算分别建立四个BP神经网络模型。使用经过预处理的试验数据，对四个BP神经网络进行了训练。将训练后的神经网络应用于两台发动机的性能参数换算，与传统换算公式相比：BP神经网络换算模型的精度和稳定性都远高于传统换算公式。该方法拓展了神经网络的应用范围，有良好的工程价值，可供其它型号航空发动机研制参考。

中文关键词:BP神经网络航空发动机参数换算精度稳定性

Application of BP Neural Network in Performance Parameter Correction of Turboshaft Engine

Abstract:The correction function for a turboshaft engine with free turbine is difficult to be given directly with specific form. The neural network is applied to engine parameter correction and four BP neural network models for the correction of power, gas temperature, fuel flow and air flow were established. Then these BP neural networks were trained using pre-processed data. Finally, trained neural networks were applied to two engines. The results show that, the precision and stability of BP neural network models were much higher than that of traditional correction formulas. The method is proved to expand the application scope of neural network and have a bright prospect in engineering practice, which can give references for other aero engines development.

keywords:BP neural network aero engine parameter correction precision stability

查看全文 查看/发表评论 下载pdf阅读器