考虑键相丢失的二重逐点Vold-Kalman滤波涡轮泵故障诊断  

Fault diagnosis of turbopump using double point-wise Vold-Kalman filtering considering key-phase loss

作　　者：王帅[1] 孙若斌 翟智 马猛 陈雪峰[1,2] WANG Shuai;SUN Ruobin;ZHAI Zhi;MA Meng;CHEN Xuefeng(School of Future Technology,Xi’an Jiaotong University,Xi’an 710049,China;School of Mechanical Engineering,Xi’an Jiaotong University,Xi’an 710049,China)

机构地区：[1]西安交通大学未来技术学院,西安710049 [2]西安交通大学机械工程学院,西安710049

出　　处：《振动与冲击》2025年第3期210-220,229,共12页Journal of Vibration and Shock

基　　金：国家自然科学基金资助项目(52105118)。

摘　　要：液体火箭发动机涡轮泵在高转速、高温度梯度、高压的非平稳工况下极易发生故障。Vold-Kalman滤波方法能够从复杂时变振动信号中检测出涡轮泵转子故障,但由于涡轮泵振动传递路径复杂,该方法依赖于所采集振动信号的载波的高采样率高精度的相位信息,在键相信号丢失和采样频率低(一圈一个脉冲)的实际应用场景下存在故障检测精度偏低的问题;且Vold-Kalman滤波使用批量式优化的方法,求解缓慢,无法在箭载计算机上实现在线检测故障。针对上述两个问题,为实现毫秒级的涡轮泵故障实时诊断,提出了一种滤波诊断方法——二重逐点Vold-Kalman滤波器(double point-wise Vold-Kalman filter,DPVKF)。DPVKF首先建立各阶次分量状态转移和状态观测的时变线性高斯模型;然后,从低精度的转速脉冲和振动信号中准确重构相应载波的高精度瞬变相位;随后,在重构相位的指导下,得到各阶次复包络的最优线性无偏估计;最终,在复杂激励干扰下提取到涡轮泵转子的故障特征。故障模拟试验和某型号涡轮泵低温轴承运转试验表明,提出的方法可实现高实时性、高可靠性的涡轮泵转子故障诊断。Turbopump of liquid rocket engines is prone to malfunction under non-stationary conditions of high speed,high temperature gradient and high pressure.Vold-Kalman filtering method can detect rotor faults in turbopump from complex time-varying vibration signals.However,due to complex transmission path of turbopump vibration,this method relies on high sampling rate and high-precision phase information of carrier waves of the collected vibration signals.In practical application scenarios of key-phase signals loss and low sampling frequency with one pulse per circle,there is a problem of low fault detection accuracy,and Vold-Kalman filtering using batch optimization method makes it be slow to solve and unable to realize online fault detection with rocket-borne computer.Here,aiming at the above 2 problems,a filtering diagnosis method called double point-wise Vold-Kalman filtering(DPVKF)was proposed to realize real-time diagnosis of turbopump faults in milli-seconds.With DPVKF,firstly,time-varying linear Gaussian models for state transfer and state observation of various order components were established,and then high-precision transient phases of corresponding carrier waves were correctly reconstructed from low precision rotating speed pulses and vibration signals.Subsequently,guided by the reconstructed phases,optimal linear unbiased estimations of various order complex envelopes were obtained,and finally fault features of turbopump rotor were extracted under complex excitation interference.Fault simulation experiments and low-temperature bearing operation experiments of a certain type of turbopump showed that the proposed method can realize high real-time and high reliability turbopump rotor fault diagnosis.

关 键 词：二重逐点Vold-Kalman滤波(DPVKF) 键相信号丢失 涡轮泵 故障诊断 

分 类 号：TH212[机械工程—机械制造及自动化] TH213.3