Higher-Order And Multiscale Local Projective Algorithm And Its Application Research In Mechanical Fault Diagnosis

Rolling bearings and gears have been widely used in a verity of rotating machinery.The phenomenon of the whole equipment downtime due to the failure of key components such as rolling bearings and gears emerge in endlessly.Thus,it is of great significance to research the fault diagnosis technology of them.Aiming at the non-stationary and non-linear characteristics of the vibration signal of rolling bearings and gears,this thesis improved the existing local projection algorithm and proposed the higher-order and multiscale local projective algorithm based on phase space reconstruction.And combined with other signal analysis methods to achieve the fault feature extraction and fault identification of gears and rolling bearings,the main contents of this thesis are described as follows:(1)In order to deal with the centroid selection of local neighborhoods in local projective noise reduction algorithm,the higher-order polynomials were used to estimate the centroid of the neighborhood more precise in the proposed method.In this way,the noise was further suppressed,and the noise reduction performance of local projective method was improved(2)In order to deal with the centroid selection of local neighborhoods in local projective noise reduction algorithm,a multiscale selection method of centroid was proposed.This method considered both the geometrical shape and statistical error of the signal in the high dimensional phase space,which can not only effectively eliminate the noise,but also can better preserve the complete geometric structure of the attractors.(3)A method of mechanical fault feature extraction based on multiscale local projective algorithm and diagonal slice spectrum was presented.Firstly,the multiscale local projective algorithm can filter out noise while maximum retention characteristics related to the fault.Secondly,the diagonal slice spectrum can identify the frequency components of quadratic phase coupling and suppress the absence of the coupled frequency component in the nonlinear signal.Therefore,the combination of the two methods can be more accurate to achieve the fault feature extraction of gears and rolling bearings.(4)From the perspective of multiscale time-frequency analysis and entropy theory,a new method of mechanical fault classification based on multiscale local projection and LMD sample entropy was considered.Firstly,the multiscale local projective algorithm was used to denoise the collected vibration signals.Secondly,the processed signals were decomposed by LMD.Then,the PF components were quantified by using the sample entropy,and the complexity information of the vibration signals at different scales were obtained.Finally,the fault classification of gears and rolling bearings were realized by using the extracted information.