The invention discloses a rough surface-based three-dimensional contact stiffness calculation method for a spur gear. The method comprises the steps of firstly by adopting hexahedral mesh generation, requiring all square meshes on a tooth surface to be equal in area; calculating contact pressure distribution of the smooth tooth surface based on a finite element method, and extracting node pressure intensity of a contact region; calculating normal contact stiffness and tangential contact stiffness of a single square mesh based on a fractal theory; and calculating the contact stiffness of the tooth surface. According to the method, the precision of a gear contact stiffness calculation model is further improved; the defect of smooth contact according to a Hertz theory in an original calculation process is overcome; an influence law of a roughness parameter on a gear stiffness characteristic in a rough tooth surface contact process and a calculation method are disclosed; and meanwhile, a finite element-based three-dimensional contact stiffness calculation model is more accurate than an original two-dimensional calculation model or calculation formula, and the influence of errors in actual assembling and manufacturing processes of the gear is considered, so that the gear contact stiffness is more accurate and a foundation is laid for gear dynamics analysis.

Rough surface-based three-dimensional contact stiffness calculation method for spur gear

The invention relates to a method for determining the amplitude of self-adaptation ensemble average empirical mode decomposition EEMD assistance noise. The method comprises the following steps that (1), a maximum value emax and grid numbers N are added to the initialized noise, the amplitude of the noise corresponding to each grid i (i =1, 2, ...N) meets a formula ei=i/N*emax; (2) the Gaussian white noise ni (t) with the amplitude being ei is sequentially added to a signal x (t); (3) the position Max (j) where a maximum value point of a signal xi (t) is located and the position Min (k) where the minimum value point is located are calculated to obtain a new sequence ai (j) and bi (j), and standard deviation Sai of the new sequence ai and standard deviation Sbi of the new sequence bi are calculated to obtain the corresponding curvilinear relationship between the noise distribution character Qi and ei; (4) er corresponding to a point where Qi descends fast along with ei is the optimal value of the amplitude of self-adaptation ensemble average empirical mode decomposition EEMD assistance noise. The method reduces the blindness of artificial parameter selection.

Method for determining amplitude of self-adaptation ensemble average empirical mode decomposition EEMD assistance noise

The invention relates to a detection method of tooth root crack fault characteristics of a planetary gear or a sun gear and belongs to the field of fault diagnosis technology and signal processing andanalysis technology. By combining windowing vibration separation, synchronization averaging and narrow-band demodulation, and observing whether phase position demodulation and amplitude value demodulation result is subjected to periodically sudden change in the same angle position, whether the planetary gear or the sun gear is subjected to tooth root crack faults is determined, so detection of the tooth root crack faults of the planetary gear or the sun gear of a planetary gear box is achieved.

Detection method of tooth root crack fault characteristics of planetary gear or sun gear

The invention discloses a planet gear box vibration signal separation method and system for fault diagnosis. The method comprises steps that (1) to-be-diagnosed planet gear box vibration signals are separated by utilizing an order tracking method to obtain harmonic components of gear box vibration signals, and the harmonic components comprise gear engagement vibration signals; (2) the gear engagement vibration signals are decomposed to a plurality of planet gears synchronously and averagely according to a continuous window function to obtain engagement vibration of a single planet gear so as to diagnose distributed abrasion of the planet gears; and (3) the harmonic components are subtracted from the to-be-diagnosed planet gear box vibration signals to obtain residual signals, the residualsignals only comprise impact signals and noise caused by local damage of the gears and bearings, the impact components in the residual signals are extracted through filtering, and monitoring and diagnosis of local damages of the parts are achieved. The method is advantaged in that gear box fault diagnosis accuracy can be improved.

Planet gear box vibration signal separation method and system for fault diagnosis

A gear pitting modeling method based on probability distribution comprises: first, establishing a tooth face probability pitting plane distribution model, with pitting randomly distributed along toothwidth and normally distributed along tooth height; second, acquiring meshing line length variation by means of image processing; third, establishing one-to-one correspondence relation between the meshing line length variation and pressure angle; fourth, comprehensively considering hertz contact stiffness, bending stiffness, shear stiffness and axial compression stiffness to precisely calculate gear time-varying mesh stiffness and solve internal excitation of a gear system. The gear pitting modeling method based on probability distribution has the advantages that the defects are overcome that,for instance, a traditional pitting model is greatly deviated from the actual and complex pitting distribution is unable to be handled; model calculation precision is improved; internal excitation ofa pitting gear system is precisely calculated; gear pitting modeling based on probability distribution is completed.

Gear pitting modeling method based on probability distribution

Provided is a peeled-off gear meshing model based on meshing stiffness. Force conditions of peeled-off gear teeth in gear engagement is considered, and a meshing process is divided into three phases, and meshing stiffness of the peeled-off gear teeth in the three meshing phases is respectively calculated. Defects of an existing peeled-off gear model that just two kinds of conditions of a peeled-off meshing region and a non-peeled-off meshing region are considered are overcome. In the model, influence of the root of gear teeth is considered, and a gear tooth is equivalent to a cantilever bam on a root circle, and peeled-off gear meshing stiffness is accurately calculated, and precision is improved, so the more accurate peeled-off gear meshing model is established. The model actually and accurately reflects a meshing process of a peeled-off gear, and the model can be effectively applied in dynamics modeling research of peeled-off gears.

Peeled-off gear meshing model based on meshing stiffness

Provided is an epicyclic gearbox signal separating and diagnostic method independent of time domain average. A cosine power ascending function is selected as a separation window function. A frequency band near engage frequency is selected for carrying out band-pass filtering on epicyclic gearbox vibration signals collected by a sensor, then filtering signal envelopes are obtained, a moment T corresponding to the envelope signal maximum value is found, and the moment is the moment when a certain epicyclic gear is closest to the sensor; and T + ifc/p is used as time shifting t' to be put into the separation window function, the separation window function is multiplied with the epicyclic gearbox vibration signals collected by the sensor, epicyclic gearbox vibration separating signals are obtained, the separating signals are subjected to Fourier transformation, if fault feature frequency sideband exists near engage frequency, a corresponding type of fault exists in an epicyclic gearbox, according to the method, the fault feature information of the epicyclic gearbox is effectively extracted, and effective diagnosis on epicyclic gearbox fault is achieved.

Epicyclic gearbox signal separating and diagnostic method independent of time domain average

The invention discloses a noise determination method disintegrated and assisted by the ensemble average empirical mode. Fast Fourier transform is performed on the white Gaussian noise expressed by randn to acquire the frequency spectrum, the frequency spectrum is multiplied by sine function, the product is expressed asY(f) =A*X(f), fast Fourier transform is performed onY(f), a time series of a noise signal that amplitude is of sinusoidal variation with frequency is acquired, the constructed noise replaces the white Gaussian noise and is added to the original signal for disintegration of the ensemble average empirical mode. According to the invention, the problem of pattern mix caused by adding of the white Gaussian noise, of which the amplitude does not change with frequency, is solved, the effective diagnosis of failures of mechanical equipment is realized, and the analyzing result is relatively correct.

Noise determination method disintegrated and assisted by ensemble average empirical mode

Disclosed is a meshing phase difference-based phenomenological modeling method for a normal single-stage epicyclic gear train. The meshing phase difference-based phenomenological modeling method includes calculating meshing phase differences between a plurality of planet gears and a central gear, namely a sun gear as well as the planet gears and a gear ring, and the meshing phase differences between one planet gear and the sun gear as well as the gear ring; analyzing influences of different transferring paths to signals during multiple meshing processes; establishing a simulation model of a vibration signal in the normal state of a planet gear train to achieve a vibration response expression. The meshing phase difference-based phenomenological modeling method breaks the limits that vibration influence of just one pair of meshing in the gear system is considered and comprehensively considers phase relations between multiple pairs of meshing in the epicyclic gear train and time-varying transferring path caused by revolution of the planet gears. According to the phenomenological model, the vibration response signal of the normal single-stage epicyclic gear train is achieved and verified by experiments, providing the basis for fault diagnosis of the single-stage epicyclic gear train.

Lin Jing

Papers

Peeled-off gear meshing model based on meshing stiffness

Epicyclic gearbox signal separating and diagnostic method independent of time domain average

Noise determination method disintegrated and assisted by ensemble average empirical mode

Gear pitting modeling method based on probability distribution

Meshing phase difference-based phenomenological modeling method for normal single-stage epicyclic gear train