论文部分内容阅读
液体火箭发动机振动检测涉及部件振动数据的收集、振动特征的抽取与度量以及度量结果的决策。基于模糊神经网络提出了一种发动机振动故障检测的基本系统。这种技术的吸引力在于：神经网络采用可变模糊集代表发动机工作模式，自然地提供了反映故障程度的有用信息；神经网络的离线学习算法可以从训练样本中提取振动知识；神经网络的监测算法不仅能正确预报故障，同时也能对新的振动信息进行在线学习。实验研究结果表明：模糊神经网络可以成功地用于泵压式液体火箭发动机热试车的振动故障检测。
Vibration detection of liquid rocket engine involves the collection of vibration data of components, extraction and measurement of vibration characteristics and decision-making of measurement results. Based on fuzzy neural network, a basic system of engine vibration fault detection is proposed. The attractiveness of this technique lies in the fact that the neural network uses the variable fuzzy set to represent the working mode of the engine and naturally provides useful information reflecting the extent of the fault. The offline learning algorithm of the neural network can extract the vibration knowledge from the training samples. The neural network monitoring The algorithm can not only predict the fault correctly, but also learn the new vibration information online. The experimental results show that the fuzzy neural network can be successfully used to detect the vibration of pump-driven liquid rocket engine.