CAJ | 학술논문

振动信号功率谱分析常用方法是周期图法及其改进算法，存在方差性能与分辨力性能的矛盾，当振动信号采样点数较少时，应用存在局限。为了解决这一问题，引入了最小方差谱估计算法，并通过信息论准则确定其最优阶次，以及通过加最优窗进一步提高其方差性能。结合工业现场实测信号，基于MATLAB环境对其实际性能进行了分析与仿真，并与经典谱估计算法以及AR模型算法进行了对比。理论分析与仿真结果表明，最优阶次下的最小方差谱估计算法，性能优于经典算法，分辨力性能与AR模型算法相当，但方差性能较好，有利于振动信号的分析。
진동신호공솔보분석상용방법시주기도법급기개진산법，존재방차성능여분변력성능적모순，당진동신호채양점수교소시，응용존재국한。위료해결저일문제，인입료최소방차보고계산법，병통과신식론준칙학정기최우계차，이급통과가최우창진일보제고기방차성능。결합공업현장실측신호，기우MATLAB배경대기실제성능진행료분석여방진，병여경전보고계산법이급AR모형산법진행료대비。이론분석여방진결과표명，최우계차하적최소방차보고계산법，성능우우경전산법，분변력성능여AR모형산법상당，단방차성능교호，유리우진동신호적분석。
The methods that often used in power spectrum analysis of vibration signal are Periodogram and its improved algorithms,which have the conflict between variance performance and resolution performance,so the analysis effect is worse when the sample length of tur-bine’ s vibration signal is relatively short. Introduce MVSE algorithm to solve this problem,and through the AIC determining its optimal order times,the same time by adding the variance of the best window to further improve its performance. Combined with the turbine’ s vi-bration signals collected in industrial site,the actual performance of the MVSE method is simulated and analyzed through MATLAB. And the performance of the method is compared with the performance of classical methods and AR model method. The theoretical analysis and simulation result show that MVSE algorithm with optimal order is better than the classical algorithm,and the resolution performance of MVSE and AR model algorithm is fairly,but the variance performance of MVSE algorithm is better,which is advantageous to the vibra-tion signal analysis.