Search

The outer ring spalling failure angle position of the rolling bearing has an important influence on the vibration size and service life of the rotating machine. It is necessary to judge the outer ring spalling failure angle within the load-bearing interval. Based on the peak change law of the acceleration vibration signal, a fault location method was proposed for ball bearing outer ring spalling. This method established the Lagrangian dynamic equation with time-varying impact force. By analyzing the change trend of the contact force in the bearing area of the ball bearing, the law of the peak ratio of the absolute value of the vibration acceleration signal was obtained by analogy, and the impact point of the acceleration vibration signal was analyzed. For nearby locations, a new judgment index was proposed, and the specific location of the fault was obtained. Simulation and experiment results show that there is a corresponding functional relationship between the horizontal vertical synchronization peak ratio of the vibration signals at different positions of the ball bearing outer ring spalling fault. Simulation and experiment verify the accuracy of the positioning law. It provides a theoretical basis for the diagnosis method to determine the location of the ball bearing outer ring spalling fault location.

To study the contact effect of the rotor system with curvic couplings under rolling bearing supports, the combination of 3D finite element and analytical method were used to derive the curvic coupling shaft section stiffness matrix, and the dynamics model of rotor-rolling bearing system was established, which was verified by Ansys. By comparing with the continuous rotor, it was found that the curvic coupling structure reduced the first-order critical rotor speed by 5% and increased the amplitude by 3%. The results show that the nonlinear bifurcation trend of the curvic coupling rotor is nearly the same as that of the continuous rotor, however, the discontinuous characteristic of the curvic coupling structure will make the rotor access or leave the periodic motion state significantly earlier. Obviously, the contact effect needs to be considered when analyzing the rotor vibration characteristics with curvic couplings. The research can provide analysis methods and design ideas for the prediction of nonlinear vibration characteristics and the operating speed design for rotor system with curvic coupling structure.

The data-driven intelligent diagnosis of rolling bearing status suffers from low recognition rate due to the poor quality of learning samples in the process of identification model construction. To address this problem, a method is proposed to improve the recognition rate of the rolling bearing intelligent diagnosis model by selecting learning samples using the deep belief neural network. First, aiming at the time-varying modulation characteristics of the rolling bearing vibration signal, a finite number of intrinsic mode function components were decomposed using variational mode decomposition, which represents different components of the original signal. Secondly, according to the fault energy fluctuations and correlation of each component, the proportion of fault information in each intrinsic mode function components is quantitatively evaluated. On the basis of this, the vibration signal is screened and reconstructed to obtain learning samples. Finally, the obtained samples are used as the input of the deep belief network, and the fault identification model of the rolling bearing is constructed accordingly. Experimental results show that the proposed method is capable of screening out the intrinsic mode function components of the rolling bearing vibration signal, which contains the main components of the fault and realizes the construction of learning sample sets. Moreover, it improves the fault recognition rate of the rolling bearing state identification model based on vibration data.

A rolling bearing performance degradation assessment algorithm based on single ring theorem in the random matrix theory is proposed to solve the problem that the traditional feature extraction algorithm is not sufficient to extract the useful information of the bearing under the industrial big data environment, and the constructed index is difficult to effectively characterize the bearing degradation process. Firstly, through segmentation, randomization, amplification and dimensional reconstruction of the bearing sampling data, a high-dimensional random matrix of the bearing at the current moment is established; Secondly, given the good high-dimensional data processing ability of random matrix theory, the distribution of eigenvalues of bearing random matrix is investigated by using the single ring theorem. Finally, the average spectral radius and the distribution of eigenvalue discrete points are used as degradation indexes to describe the degradation history of rolling bearings. The application research of the proposed method was carried out through the IMS rolling bearing full life test data, the research results show that the two quantitative indexes established by using the average spectral radius and the number of discrete points proposed based on the single ring theorem, can effectively describe the degradation process of rolling bearings and are more sensitive to the early anomaly state of rolling bearing.

In view of the shortcoming that the dimensivity reduction effect of local linear embedding（LLE） is greatly affected by the uniformity of sample distribution, the uniform distance is designed to replace the Euclidean distance in LLE, so the uniform distance local linear embedding（UDLLE） method is proposed, and the effectiveness of UDLLE is verified by typical curves. On this basis, a rolling bearing fault diagnosis method based on UDLLE was proposed, and the diagnosis example was verified. The experimental results show that UDLLE can obtain more effective low-dimensional features than some other methods, reduce the difficulty of subsequent diagnosis and improve the fault diagnosis accuracy.

In order to solve the problem of the lack of life characteristic characterization ability in the prediction of the remaining life of the rolling bearing,a method of the residual life prediction of the rolling bearing based on multi-resolution singular value decomposition and ECNN-LSTM is proposed. Firstly,the time-frequency information of vibration signals in onedimensional multi-space scale is extracted by using the multi-resolution singular value decomposition method,and the health stage is divided according to the standard deviation at the initial time. Secondly,a high-efficiency channel attention mechanism module was added to the two-layer one-dimensional convolutional neural network structure,and the convolution kernel was adaptively adjusted for multi-channel interaction without dimension reduction,so as to fully extract bearing degradation characteristics and establish effective life degradation indicators. Finally,MSE loss function is used to achieve the residual life prediction on LSTM.The feasibility and effectiveness of the proposed method are verified by Cincinnati whole life data.

Smart bearing is a structural unit of bearing set sensor. It has the characteristics of self-sensing,self-diagnosis,self-adaptation and self-regulation. It is one of the development directions of high-end bearing development at home and abroad.Firstly,Expounded the concept and key technologies of smart bearings. Secondly,the advantages and disadvantages of external smart bearing and embedded intelligent bearing are summarized. It is considered that embedded structure is the main direction of future research of smart bearing. According to the requirements of rolling bearing condition monitoring and diagnosis,the current application of smart bearing sensors,such as acceleration sensors,temperature sensors,lubricating oil quality detection sensors,torque sensors,and thin film sensors,are introduced. Finally,the development trend of smart rolling bearing monitoring technology is summarized and forecasted.

Adjusting the value of nonlinear system parameters will affect the output Signal-to-Noise Ratio（ SNR）,so Signalto-Noise Ratio（ SNR） was used as the output evaluation index of stochastic resonance. A method combining Multipoint Optimal Minimum Entroy Deconvolution Adjusted（ MOMEDA） with Cuckoo search adaptive Stochastic Resonance（ SR） was proposed to extract weak fault feature frequencies.Simulation analysis shows that using MOMEDA as the stochastic resonance pretreatment can significantly improve the weak fault signal,and experimental example verification further shows that the combination of MOMEDA method and stochastic resonance can effectively extract the characteristic frequency of weak fault signal from the signal with strong noise,so as to realize the weak fault diagnosis of rolling bearings.

Aiming at the problem that it is difficult to extract subtle fault features in the process of rolling bearing fault identification,this paper proposes a rolling bearing fault diagnosis method based on fusion convolutional neural network and support vector machine based on particle swarm optimization algorithm. In this method,the bearing vibration signal is used as the input signal of the one-dimensional convolutional neural network and the two-dimensional convolutional neural network,and the extracted fault information is fused in the convergence layer. Finally,the optimized classifier improves the accuracy of fault recognition. In order to verify the diagnostic performance of this method,a comparison will be made between the 1 D convolutional neural network and the 2 D convolutional neural network with the same specifications of the fusion convolutional neural network.Experimental results show that this method can not only improve the accuracy of fault identification,but also maintain good diagnostic performance when the signal is polluted by noise.

Aiming at the problem that the fault feature information of rolling bearing is weak under the strong background noise environment,and the single use of the complete ensemble empirical mode decomposition with adaptive noise（ CEEMDAN）method is not effective in extracting the fault feature,a method based on parameter optimized multi-point optimal minimum entropy deconvolution adjusted（ POMOMEDA） and CEEMDAN was proposed. Since the filter effect of MOMEDA is greatly affected by its parameters — fault period T and filter length L,the variable step size search method was proposed to optimize them. Firstly,the fault period T and filter length L of MOMEDA were selected by using multi-point kurtosis and permutation entropy to realize adaptive MOMEDA noise reduction for the original signal. Then,the CEEMDAN method was used to decompose the de-noised signal,and the signal reconstruction was carried out by selecting the intrinsic mode function（ IMF） containing rich fault information according to the weighted kurtosis（ WK） index. Finally,the reconstructed signal was analyzed by envelope spectrum and the fault feature information was extracted. The analysis results of simulated signals and measured signals show that the proposed method is able to extract the weak fault feature frequency of rolling bearings and has certain reliability.

In order to solve the problem of poor accuracy and low efficiency of traditional rolling bearing life prediction method. A prediction method based on bi-directional stack simple recurrent unit is proposed. Multiple time and frequency features are extracted from the original signal to construct a multi-dimensional data set to enhance the information representation and avoid the insufficient reflection of a single feature on the bearing degradation ability. Relying on the ability of recurrent neural network to process time series data,the SRU model is constructed,the traditional RNN serial operation structure is reconstructed,and the training efficiency is improved. On the basis of single layer SRU,it is stacked in reverse,so that the model can learn two-way information in time dimension,improve the accuracy of prediction. The experimental results show that the prediction accuracy and training efficiency of this method are improved.

In order to accurately diagnose the fault of rolling bearing,a wavelet base construction method based on Gaussian function is proposed. In this method,the superimposed Gaussian function is used as the initial model to construct the wavelet basis function. The coefficients of the wavelet basis function are determined by fitting the initial model with the vibration signal of rolling bearing. Through the analysis of fault signals of outer ring,inner ring and rolling element of 6205 rolling bearing,the results show that the new wavelet basis can effectively suppress the noise component in the signal,extract fault features,and realize accurate fault diagnosis of rolling bearing. Under the same decomposition conditions,the new wavelet basis can diagnose the running state of rolling bearing more accurately than Gaussian wavelet,so it has higher application value.

This paper takes the support structure of a low-pressure rotor system of a marine gas turbine as the background.Considering the heel angle of the rotor system,a dynamic model of the rotor supported by asymmetric rolling bearing is established. The variable-step fourth-order Runge-Kutta method was used to solve the system’ s dynamic equations,and the nonlinear theory were used to analyze the simulation results. The study shows that the heel angle has no effect on the stability of the system at low rotational speeds. At high rotational speed,the heel angle mainly affects the 1/2 harmonic of the system,and the system has a single chaotic island or two chaotic islands. When the left or right tilt angle of the rotor system is the same,the heel angle has the same effect on the nonlinear dynamics of the system. The research results can provide a theoretical basis for the design,monitoring and fault diagnosis of the rotor system.

The change in the internal contact state caused by the failure of the rolling bearing is an important factor that causes the change in its vibration characteristics. According to the contact form of the fault and the rolling element,this paper considered the different sizes of the fault evolution process,and established the dynamic finite element model of the multi-body contact of the cylindrical roller bearing with different degrees of failure. The effectiveness of the model was verified by the experiment on the outer ring fault of the rolling bearing. Based on the explicit dynamics method,the dynamic contact force between the rolling element and each component and the contact strain of the fault area was analyzed,then the influence law of different fault degrees on the internal dynamic characteristics of the bearing and the change law of the contact deformation under the fault condition were revealed. The results show that the contact force of the rolling elements with the components is reduced when the defects are crossed. As the degree of failure increases,the fluctuation of the internal contact force of the bearing becomes more and more intense,but the fluctuation frequency and amplitude of the contact strain decrease. In the loading area,the rolling elements of faulty bearing are in contact with both sides of the cage pocket,which causing a negative contact force. The fault tends to expand along the rolling direction of the rolling element. The research conclusion can provide a theoretical basis for fault diagnosis and residual life prediction of rolling bearings.

Aiming at the problems of traditional bearings fault diagnosis methods had such shortcomings as largely dependent on expert prior knowledge and difficulty in fault feature extraction,a method based on compressive sensing（CS） and improved deep wavelet neural network（DWNN） was proposed. Firstly,the collected vibration data of bearings were de-noised and compressed by CS. Secondly,the improved wavelet auto-encoder was designed to construct the DWNN,and the " crosslayer" connection was introduced to alleviate the gradient disappearance of the network. Finally,unsupervised pre-training of DWNN was performed using a large amount of unlabeled compressed data and supervised and fine-tuned with a small amount of tagged data to realize fault discrimination. The experimental results show that the method can effectively identify the bearings with multiple fault types and multiple fault severities,which is less affected by prior knowledge and subjective knowledge and avoids complex artificial feature extraction process. The feature extraction ability and recognition ability of proposed method are superior than artificial neural network,deep belief network,deep sparse auto-encoder and so on.

In order to extract effective features of complex signals,a fault diagnosis method combining short-time Fourier transform and two-dimensional convolution neural network is proposed. First,a short-time Fourier transform is performed on the rotating mechanical vibration signal to obtain a time-frequency map. Then,the time-frequency map is input into a two-dimensional convolution neural network for identification,and a final classification result is obtained. The method is applied to the fault diagnosis of rolling bearing and gearbox,and has achieved good results in the data of the Case Western Reserve University and the PHM2009 dataset.The correct rate is better than the direct comparison of the original signal into CNN,which verifies the superiority of the method.

Rolling bearing as a key component of rotating equipment, its performance seriously affect the safe operation of the equipment. As the equipment condition is complex, the impact component of the fault feature is often submerged by the noise signal, therefore the fault feature cannot be extracted effectively. A method was proposed based on improved complete ensemble empirical mode decomposition with adaptive noise（CEEMDAN） and kurtosis index by this paper. Firstly, the improved CEEMDAN method is used to add adaptive white noise to each signal in the decomposition process, a unique residue was computed to obtain each intrinsic model function（IMF）, compared with ensemble empirical mode decomposition（EEMD）, the decomposition is complete. Secondly, the kurtosis index is calculated of each IMF to select the reconstructed IMF component and the kurtosis index is used to select the most suitable reconstructed signal. Finally, the bearing fault feature is obtained by envelope demodulation. The results confirm that this method has better decomposition effect, better adaptability and highlight the impact of the bearing fault.

Bearing is one of the most critical and the most vulnerable parts of mechanical transmission system,Accurate identification of its faults is the key to achieve stable operation of equipment. The five degree of freedom dynamic model of rolling bearing system with local fault is established by considering the nonlinear factors such as oil film and the sliding of rolling element. An improved model of pitting failure is proposed, and the corresponding models of different fault parameters are analyzed. The system vibration signal of the inner ring and the outer ring with pitting fault is simulated, and the corresponding fault characteristics are analyzed. Through the comparative analysis of the simulation signal and experimental signal, the correctness of the model is verified, which can help to realize the early diagnosis of bearing fault.

Aiming at the problem that the fault information of rolling bearing was weak and susceptible to noise, which was difficult to judge the damage degree of bearing failure, a new method was proposed. Bearing fault location and damage diameter size could be obtained using the method based on third-order cumulant diagonal slice spectrum and gray correlation theory. Bearing fault location information could be determined using the characteristics of third-order cumulant diagonal slice spectrum to suppress the Gaussian noise and to identify the secondary phase coupling signal. And gray relational theory was used to analyze the kurtosis index, the pulse index and the margin index of the third-order cumulant diagonal section data to obtain the curve of the damage degree. Finally, actual signal is analyzed to obtain the bearing fault location and the damage diameter, which indicates the feasibility and practicability of the method.

Aiming at the fact that the vibration signal of rolling bearing would exactly display non-stationary characteristics and fault features hard to extracted,a fault method of rolling bearing based on local characteristic-scale decomposition（ LCD）and multifractal detrended fluctuation analysis（ MF-DFA） was proposed. Firstly,the vibration signals was decomposed into several intrinsic scale components（ ISC）. Secondly,the intrinsic features hidden in each major ISC were extracted by using MFDFA,among which the generalized Hurst exponents are selected as fault feature. Thirdly,liner local tangent space alignment（ LLTSA） was applied to compress the high-dimension features into low-dimension features which insensitive to fault. Finally,the support vector machine（ SVM） was employed to diagnosis fault. Experiment results of rolling bearing show that the proposed method can classify typical fault of rolling bearing exactly and has certain superiority.

Aiming at the problem that accuracy of orthogonal locality preserving projections（OLPP） for fault diagnosis is not high enough,a fault diagnosis method based on semi-supervised neighborhood adaptive orthogonal locality preserving projections（SSNA-OLPP） for dimension reduction is proposed.In this method,fault features that can represent the fault state is firstly constructed based on local characteristic-scale decomposition（LCD） and time-frequency domain feature.And then,the SSNA-OLPP is used to compress the high-dimension feature into low-dimension feature which has better discrimination.Finally,the low-dimension feature are input support vector machine（SVM） to identification fault.SSNA-OLPP can adaptive adjust the neighborhood with the guidance of local cluster coefficient,at the same time,information of some labeled samples are also used to adjust the weight matrix among all samples in the original characteristic space,as a result,better fault diagnosis accuracy can achieved.The experiment results of rolling bearing fault diagnosis verified the effectiveness of the method.

Aiming at the problem that the Local mean decomposition（LMD） method is difficult to draw early weak fault,a fault diagnosis method for the roller bearing based on maximum correlated kurtosis deconvolution（MCKD） and LMD was proposed.Firstly,the fault signal was de-noised and meantime periodic impact components were enhanced by MCKD method,Then,that result is decomposed by LMD to get PF,the correlation coefficient between the PF and the signal is used as the standard of judgment,so that the redundant low-frequency PF can be rejected.Finally,the effective PF is selected to analyze the spectrum and extract the fault feature.The experiment of the simulation data and the actual roller bearing fault diagnosis data show that the method can effectively extract the feature frequency information of incipient fault and has a certain reliability.

Based on the advantages of Bat Algorithm（ BA） that is easily to be realized,fast convergence speed,high efficiency and high universality,combining BA and blind source separation,a new method of mechanical failure blind source separation based on BA is proposed（ BA-BSS）. The BA-BSS method,as the target function of the sum of the absolute value of the kurtosis,seeks the maximum of the target function by BA,and then determines the optimal separation matrix. Simulation results show that the BA-BSS method is significantly superior to the traditional mechanical failure blind source separation based on Genetic Algorithm（ GA-BSS） in separation performance,algorithm convergence and operation speed. Finally,the BA-BSS method is successfully applied to the actual rolling bearing inner and outer mixing fault blind source separation,and the separation effect is very effective

In order to solve the problem that the fault feature information of rolling bearing is difficult to be separated,a new method of fault feature extraction is presented,which is based on the complementary ensemble empirical mode decomposition（ CEEMD） and fast independent component analysis（ Fast ICA）. First,analyze the CEEMD vibration signals,decompose them into intrinsic mode function（ IMF） components signal of different scales; then through the sensitivity evaluation algorithm,decompose and recombine the signals,and use Fast ICA to reduce their noise; in the end,conduct Hilbert envelope spectrum analysis to the signals separated by the Fast ICA,to obtain the fault feature information. This method is applied to the fault analysis of rolling bearing vibration signal,and was proved to be valid.

The surface roughness data of test-pieces with transverse profiles were firstly measured by a roughness profile measuring analyzer. With the help of the Fourier nonlinear approximation the measured data was used to form a roughness function which was superimposed to film thickness equation and a mixed thermal elastohydrodynamic lubrication model for the rolling bearing was then established. Based on the model,45 sets of numerical calculations were carried out by changing the roughness values under other conditions of the rolling bearing were fixed. Then the two quantitative relationships between the average oil film at the contact point and the roughness as well as the maximum subsurface principal stress and the roughness are respectively established through the regression analysis of the calculation results.

Aiming at the no stationary characteristic of a gear fault vibration signal,a method based on Ensemble local mean decomposition and Kernel density estimation is proposed in this paper. First,the vibration signal is decomposed to be a series PF component by ELMD,calculating RMS、kurtosis、skewness coefficient of PF components,which contains main fault information,then they are combined into a feature vector,the Classification based on kernel density estimation is proposed,multiple sets of vibration signal feature vectors are used to train and test,identify their fault condition. The results showed that this method can effectively identify the fault of rolling bearing,and it is better than the LMD method

Considering the early fault feature information of rolling bearings is difficult to identify,and form the frequency bands after wavelet packet decomposition can not be effectively determined and adaptive to extract the resonance band,the concept of amplitude entropy of frequency band is proposed.On this basis,the wavelet packet transform and Teager energy spectrum was combined,a rolling bearing early fault feature extraction method is proposed based on wavelet packet transform adaptive Teager energy spectrum.Firstly,the vibration signal was decomposed by wavelet packet,and the frequency amplitude entropy of each subband was calculated.Then,on the basis of kurtosis index to determine the best entropy and the optimal decomposition level of wavelet packet,thus,the resonance band was extracted adaptively and effectively.Finally,the Teager energy spectrum analysis was performed to identify the frequency of the bearing fault.Through the signal simulation and experimental data analysis it verifies the effectiveness of the proposed method.

Load distribution characteristics of rolling bearing at high speed have important effects on the rotor system’s dynamic characteristics. In order to simulate the actual load distribution,the dynamic analysis model of rolling bearing is established by quasi-dynamic,simultaneously considering radial load,axial load,centrifugal force and gyroscopic moment acted on the rolling bearing,and study the rolling bearing load distribution under different radial load,axial load and speed conditions by MATLAB programming for solving nonlinear equations,the effect rules of contact angle and deformation are analyzed by program.

For the problem of Local Mean Decomposition（ LMD） was easily affected by noise interference when in the extraction of fault features,a rolling bearing fault diagnosis method which based on LMD and Independent Component Analysis（ ICA） was proposed. Firstly,original signal was decomposed into a series of production functions（ PF） by the LMD method.Secondly,the estimate of PF was obtained after the PF components had been separated by ICA method,and the noise was effectively removed. Then,mutual information,correlation coefficient and approximate entropy which were extracted from the estimate of PF components were grouped together as a feature vector. Finally,the fault feature vectors were classified by SVM.The results of the feature extraction and fault diagnosis experiments show that the fault recognition rate of LMD-ICA method is significantly better than the traditional LMD method.

Reduce the work wear and prolong the service life of rolling bearings is crucial problem in industrial production.Sums up the basic types of rolling bearing and wear problem,the research of rolling bearing wear process at home and abroad,at present,the experimental study of rolling bearing wear,numerical simulation method,system engineering are introduced in the failure of the rolling bearing analysis method and the method of repair is commonly used in the actual production and the development trend of rolling bearing and wear problems and several aspects of the need to strengthen research were also discussed.

Based on the bearing contact mechanics and heat transfer theory,thermal-structure coupling model of hollow cylindrical roller bearing for cone bit was established. Temperature,stress and deformation distribution of the bearing were analyzed,and influences of parameters on bearing’s thermal-structure coupling field were studied,such as hollow ratio,bit pressure,friction factor,revolving speed,grease temperature and well temperature. The results show that the maximum temperature of bit’s hollow cylindrical roller bearing system appears on the thrust bearing,the maximum equivalent stress occurs on the tooth claw journal,the maximum displacement occurs on the cone,and the maximum contact stress occurs on the hollow roller contact pair. With the increasing of hollow ratio,the maximum temperature,contact stress and deformation of the hollow roller gradually reduce,while the maximum equivalent stress decrease firstly and then increase. The equivalent stress of the hollow roller with hollow ratio 55% is minimal. Temperature,stress and deformation of the bearing parts increase with the increasing of drilling pressure,friction factor,revolving speed,grease temperature and well temperature. These results can provide useful references in designing and optimizing the bearing of cone bit,and benefit the application of cone bit in high temperature deep well.

Aiming at feature identification of weak signal for rolling bearing fault,an adaptive redundant lifting scheme packet（ ARLSP） diagnostic method of undecimated operation was proposed to solve the faults distortion that caused by traditional wavelet packet or lifting wavelet transform using sample operation.ARLSP was based on the principle of lifting scheme.Initial undecimated prediction and update operator were calculated by Lagrange interpolating subdivision,and then constructed an ARLSP decomposition and reconstruction algorithm.Simulation signals were tried out by denoising and anti frequency aliasing,results show that ARLSP method was superior to the traditional wavelet packet in denoising capabilities,and there are no frequency aliasing.Moreover,taking an experiment on an asynchronous motor,the vibration signals of the rolling bearings 6205 under conditions of no fault,the inner ring fault,outer ring fault and rolling fault were measured.ARLSP method has been successfully used to extract characteristic frequency and frequency multiplication for various faults,and has got a higher diagnostic accuracy than traditional wavelet packet.

Considering the reflection and influences of nonlinear touch rolling bearing towards dynamic,the theme of this paper focus on the slight axe experiment flat roof and comprise pedestal as well as rolling bearing. Use the way of variable step size "Runge-kutta"to make a solution through the model of dynamic. On the other side,the bifurcation diagrams also investigate the mass eccentricity for different rotors and its rotate speed. By the way,the evident change of shaft stiffness and supporting damp appear more obvious. The orbits plot,Poincare pictures and its frequency spectrum analyzed the signal features under the various situations. Through the experiment we can safely draw the conclusion that with the increase of the speed,vibrancy is higher up,the area of periodic locomotion will become narrow,so keep clear of the chaos sedulously when it happen to add up speed. The shaft stiffness have a great influence to the rotors,bulky stiffness can improve the stabilization to the rotors. But it will pass the vibration to the rotors. Meanwhile,avoid the bulky mass eccentricity is necessary if you want to keep a normal area and lower the impact speed. Supporting damp makes difference to the rotors vibration,with the addition of supporting damp,vibrancy will be reduced,the area of Chaos locomotion will become narrow,the area of periodic locomotion will be higher up. As a result,the system will be stabilized.