Abstract: For some faults of refusing action， sticking and jamming under valves operation for a liquid oxygen/liquid hydrogen rocket engine， the condition monitoring technique of valve switching action based on vibration characteristic analysis was investigated. Dual self-optimizing wavelet technology was used to denoise the vibration signals of valves. Short-time Fourier transform method was then adopted to conduct the time-frequency analysis of the denoised signals. The relative instantaneous energy extracted from the signal analysis was regarded as one index to recognize the status of valve switching action. Based on this tool， a portable system which can automatically detect the valve action was explored by using LabVIEW software. The practical application of the system shows that the relative instantaneous energy of the ‘open’ action of the valve is one order of magnitude larger than that of the ‘off’ action， and the relative instantaneous energy of the ‘off’ action is 2 orders of magnitude larger than that of the status without any action. The switching status of the valve can be identified timely and accurately by the relative instantaneous energy.

Key words: Rocket engine valve；Vibration characteristics；Dual self-optimizing wavelet analysis；Relative instantaneous energy；Condition recognition

摘要： 针对液氢液氧火箭发动机阀门在工作过程中出现拒动、卡滞等故障，研究了基于振动特性分析的阀门开关动作状态识别技术。首先应用双自寻优小波分析技术对阀门振动信号进行降噪，再利用短时傅里叶变换对降噪后信号进行时频分析，最后提炼出用信号相对瞬时能量这一指标来识别阀门的开关动作状态。基于此技术，运用LabVIEW开发出便携式阀门开关动作自动识别系统。经工程实际应用，阀门“开”动作比“关”动作相对瞬时能量大1个数量级，“关”动作比无动作时相对瞬时能量大2个数量级。阀门开关动作状态可通过相对瞬时能量及时、准确地识别出。

关键词: 火箭发动机阀门；振动特性；双自寻优小波分析；相对瞬时能量；状态识别