CAJ | 학술논문

针对强背景噪声环境下齿轮箱故障特征信号往往被噪声淹没等问题，提出最小熵反褶积（Minimum entropy deconvolution，MED）和循环域解调的方法提取齿轮箱故障特征。通过仿真信号发现循环自相关函数解调在强背景噪声下不具有免疫性，为了剔除噪声的干扰，提取故障特征信息，先用MED作为滤波器，以最大峭度值作为滤波的终止条件，通过仿真信号验证其强大的降噪功能，同时用提出的方法对强背景噪声下的齿轮箱多故障试验台振动信号进行降噪处理，对降噪后的信号进行循环自相关函数解调分析，成功提取出故障特征，验证此方法的可靠性。
침대강배경조성배경하치륜상고장특정신호왕왕피조성엄몰등문제，제출최소적반습적（Minimum entropy deconvolution，MED）화순배역해조적방법제취치륜상고장특정。통과방진신호발현순배자상관함수해조재강배경조성하불구유면역성，위료척제조성적간우，제취고장특정신식，선용MED작위려파기，이최대초도치작위려파적종지조건，통과방진신호험증기강대적강조공능，동시용제출적방법대강배경조성하적치륜상다고장시험태진동신호진행강조처리，대강조후적신호진행순배자상관함수해조분석，성공제취출고장특정，험증차방법적가고성。
The fault characteristic signals of gearboxes are usually drowned out in strong background noise. So, it is difficult to identify them by using conventional diagnosis methods. In this paper, the method combining MED (minimum entropy deconvolution) with cyclic autocorrelation function was proposed to extract the fault features of the gearboxes. It is found that the cyclic autocorrelation function can be applied to separate out the modulators effectively, but it does not work well in the condition of very strong background noise. Thus, the MED was used as the pre-filter to refine the vibration signals with the maximum steepness value as the ultimate condition of filtering, so that the interference of the strong background noise was eliminated and the fault feature signals could be extracted. The strong denoising function of this method was verified through simulative signals. Finally, this method was applied to the denoise processing of multi-fault vibration signals of a turbine gearbox in a strong noise background. The denoised signals were analyzed with the cyclic autocorrelation function. The fault features of the signals were successfully extracted. The reliability and feasibility of this method were verified.