| 陈聪1，娄高3，高洁2，陈灏1.基于ABC-RBF神经网络的飞机燃油流量监测与故障诊断[J].航空发动机,2022,48(3):89-93 |
| Aircraft Fuel Flow Prediction and Fault Diagnosis Based on ABC-RBF Neural Network |
| DOI： |
| Key Words:fuel flow prediction ABC algorithm RBF neural network quick access recorder data aeroengine |
|
| 摘要点击次数: 648 |
| 全文下载次数: 577 |
| Abstract:In order to detect the performance and fault of the engine，the fuel flow prediction model was established based on neural network. The Artificial Bee Colony (ABC) algorithm was expanded in three dimensions combined with the prediction demand，and the generalization value and the center value of neural network based on Radial Basis Function (RBF) were optimized respectively. Compared with the classical RBF neural network and K-means clustering algorithm，the after ABC algorithm three dimensional estension had better predic?
tion effect on the "feedback update" of RBF neural network，its calculation average difference and prediction error were smaller，and the time required was shorter. The data of short voyage，medium voyage and long voyage flights were random-ly selected for verification. The results show that the selection of TRA、H、Ma、TAT、N1、N2 and other parameters can reflect the operating conditions of the engine，and the pre?
diction effect is ideal. ABC algorithm has strong ability to update the optimized RBF neural network，which can obtain higher prediction accuracy and reduce mae value. Through the flight fault data，it is verified that the method of fault diagnosis using neural network has great practical value. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
