CAJ | 학술논문

针对轴向柱塞液压泵故障引起的振动信号非线性强、故障信息湮灭在噪声干扰的问题，提出一种基于加权形态非抽样小波分解（WMUWD）的振动信号预处理方法。首先，在形态非抽样小波分解的一般框架下，提出 WMUWD方法，利用特征能量因子表征形态非抽样各分解层近似信号对故障特征的贡献量，并以此为依据进行加权融合，以提高有用信息比重，便于特征提取；在此基础上，对 WMUWD 方法的初始参数设置进行了分析，给出了一套比较系统的优选组合方法；最后，利用仿真信号以及液压泵实测振动信号验证了该方法的有效性。
침대축향주새액압빙고장인기적진동신호비선성강、고장신식인멸재조성간우적문제，제출일충기우가권형태비추양소파분해（WMUWD）적진동신호예처리방법。수선，재형태비추양소파분해적일반광가하，제출 WMUWD방법，이용특정능량인자표정형태비추양각분해층근사신호대고장특정적공헌량，병이차위의거진행가권융합，이제고유용신식비중，편우특정제취；재차기출상，대 WMUWD 방법적초시삼수설치진행료분석，급출료일투비교계통적우선조합방법；최후，이용방진신호이급액압빙실측진동신호험증료해방법적유효성。
In allusion to the problem that the vibration of an axial piston pump is of strong nonlinearity,and the fault feature information is affected seriously by noises,a novel method for vibration signal preprocessing based on the weighted morphological un-decimated wavelet decomposition (WMUWD)was proposed.The WMUWD method was presented under the general frame of the morphological un-decimated wavelet decomposition.In order to increase the useful feature information content,approximate signals of various decomposition layers were weightedly fused according to their contributions to fault features,they were measured with the feature energy factors.On this basis,the initial indexes of WMUWD were analyzed and a systematic method for optimal selection was provided.The validity of the method was testified by using the data of simulated signals and real vibration signals.