摘要
为解决传统频域积分对非平稳信号处理误差过大问题,根据水轮发电机组振动信号的特点,提出了一种利用EM D的自适应滤波特性对实测加速度信号频域积分方法的修正算法。首先,对振动信号进行频域积分,得到相对应的速度和位移时程数据;然后,利用EM D的自适应的分析能力,将时域数据分解成不同特征时间尺度的固有模态函数,并按环境噪声以及水电站信号的特征频率进行滤波处理。仿真试验表明,在3种方法中,该方法的相对误差最小,相关系数最大。将该方法应用于景洪水电站振动信号处理,积分所得数据符合实际振动状况。结果表明,该方法求速度和位移是可行的,能够对水电站振动信号进行处理。
To solve the obvious errors of the frequency-domain integration when dealing with the measured acceleration data,a modified algorithm is provided to use the EMD′s adaptive features for the frequency-domain integration of the measured acceleration signal,according to the traits of vibration signals of the hydro-generating unit.In the method,the frequency domain integral of the vibration signal is used to calculate the time history data about velocity and displacement.The time history data are divided into many intrinsic mode functions(IMFs) by the EMD′s adaptive features,and then these IMFs are filtered according to the environmental noise and the characteristic frequency of the signal of hydro-generating unit.The simulation results show that the relative error is minimal and the correlation coefficient is maximal.The result of processing vibration signals of Jinghong Hydropower shows that integration results accord with the facts.The above analyses conclude that the method is feasible for dealing with acceleration signal and is reliable for hydropower vibration signal post-treatment.
出处
《振动．测试与诊断》
EI
CSCD
北大核心
2011年第5期610-613,665-666,共4页
Journal of Vibration,Measurement & Diagnosis
基金
国家自然科学基金资助项目(编号:50679009)
关键词
水轮发电机组
振动信号
时域积分
频域积分
经验模态分解
water turbine-generator unit vibration signal time-domain integral frequency-domain integral empirical mode decomposition(EMD)
