CN102798462B - No-time-mark order tracking method based on self-demodulation transform - Google Patents

Abstract

The invention provides a no-time-mark order tracking method based on self-demodulation transform, comprising the following steps of: firstly, adsorbing an accelerated speed or speed sensor on a tested rotary machine and collecting a vibration signal; utilizing fractional order Fourier transform to carry out time frequency rotation on the vibration signal and projecting to a frequency axis; searching a corresponding Fourier transform order alpha obtained by projecting the extreme great value of a characteristic frequency component; then extracting a stable projection amount of the characteristic frequency component by adopting a band-pass filtering method; carrying out alpha order fraction Fourier transform to obtain an engaged frequency component; then carrying out Hilbert transform to obtain an instantaneous phase of the engaged frequency component; and finally, utilizing the instantaneous phase to carry out the self-demodulation transform on the original vibration signal to obtain an order spectrum of rotary equipment, so as to finish the no-time-mark order tracking. According to the no-time-mark order tracking method based on the self-demodulation transform, the dependence on a time mark can be get rid, so that the no-time-mark order tracking method not only can be suitable for a weak rotary speed fluctuation condition, but also can be suitable for online of the rotary machine under a large rotary speed fluctuation condition of starting and stopping the machine; and the real-time order tracking is realized.

Description

Based on explain that modulation changes by oneself without markers Order Tracking

Technical field

The invention belongs to rotary machinery fault diagnosis and control field, particularly based on explain that modulation changes by oneself without markers Order Tracking.

Background technology

It is analysis means the most frequently used in rotary machine fault diagnosis that order is followed the tracks of, and is also to solve in the fluctuation of speed situation gordian technique of rotary machine fault diagnosis.Order Tracking, from essence, is that the vibration information of rotary machine or voice signal are converted into equiangular sampling sequence from the sampling of equal time sample sequence, thereby effectively suppresses the impact of the fluctuation of speed on spectrum analysis and fault diagnosis.

The key problem that order is followed the tracks of is how accurately to extract the angle domain information of rotary machine, to realize the angle domain resampling of vibration signal.At present, existing order tracking has following several: hardware order is followed the tracks of; Calculating order is followed the tracks of; Based on the Order Tracking of instantaneous Frequency Estimation, how without timing signal in the situation that, realize rotary machine online, order is followed the tracks of and is not only had theory significance in real time, and has very large industrial application value.

Number of patent application is that 201110169763.5,201120172140.9 patent has provided respectively implementation method and corresponding hardware unit that order is followed the tracks of, but above two kinds of patents are all to need timing signal, it is speed probe, number of patent application is that 201110026078.7 patent has proposed one and utilizes ratio of gear parameter to generate analog pulse, and the Order Tracking of vibration signal being carried out to equal angles resampling, but the method still need to be equipped with speed probe at high speed axle head.

Summary of the invention

In order to overcome the shortcoming of above-mentioned prior art; the object of the present invention is to provide based on explain that modulation changes by oneself without markers Order Tracking; break away to time target dependence; be not only applicable to faint fluctuation of speed operating mode, and be applicable to online, the in real time order tracking of rotary machine under the large fluctuation of speed operating modes such as start and stop.

In order to achieve the above object, the technical scheme that the present invention takes is:

Based on explain that modulation changes by oneself without markers Order Tracking, comprise the following steps:

Step 1, degree of will speed up or speed pickup are adsorbed in tested rotary machine, and its vibration signal is gathered, and vibration signal is designated as to x (t);

Step 2, adopts the Fourier Transform of Fractional Order shown in formula (1) to carry out time-frequency rotation to vibration signal, and to frequency axis projection, search makes the projection of characteristic frequency component get the corresponding Fourier transform exponent number of maximum value α, the signal after conversion is denoted as: X _α(u);

$X_{\mathrm{\α}}\left(u\right)={\∫}_{-\∞}^{\∞}x\left(t\right)K_{\mathrm{\α}}(t,u)\mathrm{dt}=\sqrt{\frac{1-j\cot \mathrm{\α}}{2\mathrm{\π}}}{e}^{j\frac{u^2}{2}\cot \mathrm{\α}}{\∫}_{-\∞}^{+\∞}x\left(t\right)e^{j\frac{u^2}{2}\cot \mathrm{\α}}{e}^{\mathrm{jut}\csc \mathrm{\α}}\mathrm{dt}---\left(1\right)$

Wherein: x (t)---vibration signal;

K _αthe kernel function of (t, u)---Fourier Transform of Fractional Order;

α---fractional order;

X _α(u)---the signal after conversion;

Step 3, adopts band-pass filtering method from X _α(u) in, extract the steady projection amount of characteristic frequency composition, and be denoted as X _{α s}(u);

Step 4, to X _{α s}(u) carry out α rank fractional Fourier inverse transformation, obtain meshing frequency composition, be designated as X _m(n);

Step 5, to X _m(n) carry out Hilbert and convert the instantaneous phase that obtains meshing frequency composition;

Step 6, utilizes instantaneous phase, original vibration signal is explained by oneself to modulation and change, and obtains the order spectrum of apparatus for rotating, thus complete without time target order follow the tracks of.

The present invention, than prior art, has following beneficial effect:

A) proposed by the invention without markers Order Tracking, can break away from traditional Order Tracking to time target requirement, therefore there is range of application more widely.

B) implementation algorithm of this method, based on Fast Fourier Transform (FFT), has higher counting yield, therefore can realize on-line analysis and the diagnosis of mechanical equipment fault.

C), aspect test macro exploitation: the method for the invention also can effectively reduce the hardware cost of existing test macro, improve the market competitiveness of test macro.

Brief description of the drawings

Fig. 1 is embodiment experiment table structural representation.

Fig. 2 is the climb curve of embodiment drive motor output shaft.

Fig. 3 is the frequency spectrum of embodiment original vibration signal.

Fig. 4 is the instantaneous phase of meshing frequency.

Fig. 5 is the order spectrum that this method obtains.

Fig. 6 is that this method obtains order and composes the partial, detailed view between 55-75 order.

Embodiment

Below in conjunction with drawings and Examples, the present invention is described in detail.

Boosting velocity procedure order with gear experiment table is tracked as example, this gear case experiment table is by drive motor 1, the first gear 2, the second gear 3, the 3rd gear 4, the 4th gear 5, detent 6 parts compositions, as shown in Figure 1, the output shaft of drive motor 1 is connected with the first gear 2, the first gear 2 and the second gear 3 engage, the second gear 3 and the 3rd gear 4 are arranged on same transmission shaft, the 3rd gear 4 and the 4th gear 5 engage, the transmission shaft of the 4th gear 5 is connected with detent 6, in test, vibration acceleration sensor A is adsorbed in the bearing (ball) cover position near the first gear 2.In test, the rotating speed of drive motor 1 output shaft rises to 730RPM from 610RPM, as shown in Figure 2.

The frequency spectrum of the original vibration of gear experiment table as shown in Figure 3.Can find out, due to the existence of the fluctuation of speed, vibration signal frequency spectrum is fuzzy serious, cannot differentiate the feature meshing frequencies at different levels of gear case from this figure.

In order to identify the vibration characteristics of gear case in boosting velocity procedure, adopt the present invention to carry out order tracking to raw data.

Based on explain that modulation changes by oneself without markers Order Tracking, comprise the following steps:

Step 1, degree of will speed up sensors A is adsorbed in the bearing (ball) cover position of tested gear case near the first gear 2, and its vibration signal is gathered, and vibration signal is designated as to x (t);

Step 2, adopts the Fourier Transform of Fractional Order shown in formula (1) to carry out time-frequency rotation to vibration signal, and to frequency axis projection, search makes the projection of characteristic frequency component get the corresponding Fourier transform exponent number of maximum value α, the signal after conversion is denoted as: X _α(u);

$X_{\mathrm{\α}}\left(u\right)={\∫}_{-\∞}^{\∞}x\left(t\right)K_{\mathrm{\α}}(t,u)\mathrm{dt}=\sqrt{\frac{1-j\cot \mathrm{\α}}{2\mathrm{\π}}}{e}^{j\frac{u^2}{2}\cot \mathrm{\α}}{\∫}_{-\∞}^{+\∞}x\left(t\right)e^{j\frac{u^2}{2}\cot \mathrm{\α}}{e}^{\mathrm{jut}\csc \mathrm{\α}}\mathrm{dt}---\left(1\right)$

Wherein: x (t)---vibration signal;

K _αthe kernel function of (t, u)---Fourier Transform of Fractional Order;

α---fractional order;

X _α(u)---the signal after conversion;

Step 3, adopts band-pass filtering method from X _α(u) in, extract the steady projection amount of characteristic frequency composition, and be denoted as X _{α s}(u);

Step 4, to X _{α s}(u) carry out α rank fractional Fourier inverse transformation, obtain meshing frequency composition, be designated as X _m(n);

Step 5, to X _m(n) carry out instantaneous phase that Hilbert conversion obtains meshing frequency composition as shown in Figure 4;

Step 6, utilizes instantaneous phase, original vibration signal is explained by oneself to modulation and change, and obtains the order spectrum of apparatus for rotating, as shown in Figure 5, thereby complete without time target order follow the tracks of.

The local detail Fig. 6 composing between 55-75 order by order spectrogram 5 and order can find out, line structure clear and definite of the present invention, can truly reflect the health status of gear box structure, effectively overcome due to the caused frequency spectrum blooming of the raising speed stage fluctuation of speed, therefore, the present invention is to without under markers, non-stationary operating mode, and the monitoring and fault diagnosis of gear box arrangement has great importance.

Claims (1)

1. Based on explain that modulation changes by oneself without markers Order Tracking, it is characterized in that, comprise the following steps:

   Step 1, degree of will speed up or speed pickup are adsorbed in tested rotary machine, and its vibration signal is gathered, and vibration signal is designated as to x (t);

   Step 2, adopts the Fourier Transform of Fractional Order shown in formula (1) to carry out time-frequency rotation to vibration signal, and to frequency axis projection, search makes the projection of characteristic frequency component get the corresponding Fourier transform exponent number of maximum value α, the signal after conversion is denoted as: X _α(u);

   $X_{\mathrm{\α}}\left(u\right)={\∫}_{-\∞}^{\∞}x\left(t\right)K_{\mathrm{\α}}(t,u)\mathrm{dt}=\sqrt{\frac{1-j\cot \mathrm{\α}}{2\mathrm{\π}}}{e}^{j\frac{u^2}{2}\cot \mathrm{\α}}{\∫}_{-\∞}^{+\∞}x\left(t\right)e^{j\frac{u^2}{2}\cot \mathrm{\α}}{e}^{\mathrm{jut}\csc \mathrm{\α}}\mathrm{dt}---\left(1\right)$

   Wherein: x (t)---vibration signal;

   K _αthe kernel function of (t, u)---Fourier Transform of Fractional Order;

   α---fractional order;

   X _α(u)---the signal after conversion;

   Step 3, adopts band-pass filtering method from X _α(u) in, extract the steady projection amount of characteristic frequency composition, and be denoted as X _{α s}(u);

   Step 4, to X _{α s}(u) carry out α rank fractional Fourier inverse transformation, obtain meshing frequency composition, be designated as X _m(n);

   Step 5, to X _m(n) carry out Hilbert and convert the instantaneous phase that obtains meshing frequency composition;

   Step 6, utilizes instantaneous phase, original vibration signal is explained by oneself to modulation and change, and obtains the order spectrum of apparatus for rotating, thus complete without time target order follow the tracks of.

Images