| 摘要: |
| 研究了用于液体火箭发动机（ＬＲＥ）故障仿真与故障检测的神经网络ＢＰ（ＢａｃｋＰｒｏｐａｎａｔｉｏｎ）算法。在ＢＰ算法中采用了加噪声等技术来避免系统误差陷入局部极小，训练出精度高（误差小于０．０２）的神经网络。试验表明：神经网络ＢＰ算法成功地用于故障仿真与故障检测。 |
| 关键词: 液体推进剂火箭发动机 故障诊断专家系统 故障模拟 故障检测 噪声干扰 |
| DOI： |
| 分类号:V448.15 |
| 基金项目:国家自然科学基金 |
|
| FAILURE SIMULATION AND DETECTION OF A LIQUID ROCKET ENGINE BASED ON ARTIFICIAL NOISE BP NEURAL NETWORK |
|
Huang Minchao, Zhang Yulin, Feng Xin
|
|
Department of Aerospace Technology, NationalUniversity of Defense Technology, Changsha, 410073
|
| Abstract: |
| BP (Back Propagation) neural network used in liquid rocket engine(LRE)failure simulation and detection was described. In order that the system errordid not fall into partial minimum, additional noise was recommended in BP algorithm, and an accurate neural network was trained (with error below 0. 02). It wasshown that the BP neural network could be successfully employed in simulation anddetection of LRE failures. |
| Key words: Liquid propellant rocket engine Fault diagnosis expert system Fault simulation Fault detection Noise jamming |
