神经网络在液体火箭发动机故障检测中的应用(Ⅰ)非线性辨识技术

NEURAL NETWORK APPROACH TO FAULT DETECTION OF LIQUID ROCKET ENGINE (Ⅰ)NONLINEAR IDENTIFICATION TECHNOLOGY

关键词:  液体推进剂火箭发动机  人工神经元网络  故障检测

Key words:  Liquid propellant rocket engine  Artificial neural network  Fault detection

摘要:

应用ＢＰ神经网络，提出了一种液体火箭发动机故障实时检测系统。它采用非线性辨识技术，在建立发动机数学模型和输出包含故障信息的监视指标信号之后，用阈值线与监视指标比较，从而预报发动机故障。液体火箭发动机启动与稳态过程的试验数据检验表明：基于非线性辨识技术的故障检测系统性能优越。

Abstract:

A fault detection system for a liquid propellant rocket engine was proposed in the realtime conditions using the back propagation(BP)neural network.The engine mathematical model with the nonlinear identification technology was established,and the watching index signal that contains the engine fault message was output at the same time.If the watching index signal became bigger than a given threshold,then there existed a fault during the engine running.While the engine operation was divided into the start and steady state processes,the test data were validated the advanced performances of the fault detection system based on the nonlinear identification technology.Comparatively,the pattern recognition technology will be described in part Ⅱ