摘要
针对氢气压缩机的振动信号成分相对复杂、易受噪声干扰以及压缩周期难以自动提取的问题,本文提出了基于变分时域分解、Hilbert变换和自相关函数相结合的液驱往复式氢气压缩机的压缩周期参数非直接测量方法。首先建立了液驱往复式压缩机活塞冲击仿真信号的数学模型,模拟活塞撞击气缸左、右极限位置时的振动信号,利用变分时域分解算法将模拟的活塞冲击信号从整个仿真信号中分离出来,然后利用Hilbert变换对分离的活塞冲击仿真信号进行包络解调并绘制包络谱,对包络谱的幅值序列进行自相关运算,最后自动检索自相关函数曲线最大峰值对应的延迟点数,计算压缩周期参数值。仿真信号分析结果表明:该方法可以自动提取获得正确的压缩周期参数值,当冲击信号受噪声影响时仍然具有高鲁棒性,其对9种不同信噪比仿真信号的提取成功率为88.89%。最后,将该方法应用到真实的液驱往复式氢气压缩机上,分别采集了不同电机转频、进气压力和二级排气压力共12种稳态运行工况下的排气阀振动信号。实验台数据分析结果表明,该方法提取的成功率为100%,与频谱、倒频谱、包络谱以及自相关函数的结果对比表明,该方法在所有5种方法中的提取成功率最高,进而验证了本文所提方法的有效性和优越性,为氢气压缩机状态监测、故障诊断与寿命预测提供了基础运行工况参数。
Objective This study aims to address significant challenges associated with the complex vibration signals in hydrogen compressors,which are susceptible to noise interference.This interference complicates the extraction of accurate compression cycle parameters.It develops and validates a novel non-direct measurement method that enhances the reliability and efficiency of parameter extraction in liquid-driven reciprocating hydrogen compressors,which are crucial for operational efficiency and diagnostic accuracy in hydrogen fuel systems.Methods The proposed method integrates Variational Time-Domain Decomposition(VTDD),Hilbert transform,and autocorrelation functions to enhance signal analysis and parameter extraction in noisy environments.The approach commences with developing an elaborate mathematical model that simulates dynamic interactions within the compressor,specifically focusing on piston impacts at cylinder limit positions.This simulation is pivotal for determining the intrinsic mechanical behaviors characteristic of the compression cycle.The model facilitates the extraction of clean piston impact signals from a noisy background using the VTDD algorithm,isolating pure signal components essential for subsequent analyses.The next step involves applying the Hilbert transformation to these isolated signals for envelope demodulation,enhancing signal clarity by amplifying its amplitude modulations,which are pivotal for accurately identifying cycle parameters.The amplitude sequences derived from the envelope are then examined using autocorrelation functions.This assists in precisely identifying piston impact periodicity and enables the accurate calculation of compression cycle parameters.Results and Discussions The effectiveness of the proposed method is rigorously validated through simulations and real-world applications in a liquid-driven reciprocating hydrogen compressor under various operational conditions.The method in simulation tests demonstrated an 88.89%success rate in precisely extracting compression cycle
作者
陈志鹏
刘志亮
张潇楠
曾学兵
蒋兴文
顾小明
CHEN Zhipeng;LIU Zhiliang;ZHANG Xiaonan;ZENG Xuebing;JIANG Xingwen;GU Xiaoming(School of Mechanical and Electrical Engineering,University of Electronic Science and Technology of China,Chengdu 611731,China;Chengdu Andisoon Measure Company Limited,Chengdu 610207,China;Houpu Clean Energy Group Company Limited,Chengdu 610097,China)
出处
《工程科学与技术》
EI
CAS
CSCD
北大核心
2024年第6期82-92,共11页
Advanced Engineering Sciences
基金
四川省科技计划资助项目(2023YFG0351,2024JDHJ0057)。
关键词
压缩周期
液驱往复式压缩机
振动分析
氢能
compression cycle
liquid-driven reciprocating compressor
vibration analysis
hydrogen energy
