CN104849037A - Rotation machinery fault diagnosis method based on complex signal double-side spectrum analysis - Google Patents

Abstract

The invention relates to a rotating machinery fault diagnosis method based on complex signal bilateral spectrum analysis. The method comprises the following steps: 1) signal acquisition and preprocessing; 2) complex signal synthesis and analysis; 3) FFT transformation and axis track synthesis. This method is not only equivalent to the holographic spectrum in essence, but also simpler and more effective than the holographic spectrum technology. It only needs to do one FFT transformation and one spectrum correction, and there is no need to analyze the X and Y directions separately; The precession circle parameter directly embodies Bently's full-spectrum technology; it uses negative frequency information that is not used by holographic spectrum technology and full vector spectrum technology, making the synthesis of fault parameters simpler and more practical. By performing FFT transformation on the complex signal, the FFT transformation of the X and Y direction signals can be obtained at one time, which improves the calculation speed, enhances the real-time performance of signal analysis, and also makes the signal analysis accuracy on X and Y consistent.

Description

A Fault Diagnosis Method for Rotating Machinery Based on Bilateral Spectrum Analysis of Complex Signals

technical field

The invention belongs to the technical field of mechanical fault diagnosis, and relates to a fault diagnosis method for rotating machinery based on complex signal bilateral spectrum analysis.

Background technique

Rotating machinery is widely used in aviation, aerospace, petroleum, chemical, transportation, electric power and other industrial production fields. Its reliable operation not only involves the economic benefits of the enterprise, but also affects the safety and continuity of production. Therefore, the failure prediction of rotating machinery It is of great significance to ensure the smooth operation of equipment and reduce or even avoid major safety accidents through the related research on fault diagnosis and timely and accurate identification of fault initiation and evolution.

Equipment fault diagnosis technology was formed in the UK in the early 1970s, and has been valued by enterprises and government authorities because of its practicability and the benefits it brings to society and enterprises. With the progress and development of science and technology, especially the rapid development and popularization of computer technology, fault diagnosis has gradually formed a relatively complete emerging interdisciplinary engineering discipline. As far as the fault diagnosis technology of rotating machinery is concerned, the vibration diagnosis method covers the most extensive areas, has the most solid theoretical foundation, and has the most research results and is the most mature. The analysis techniques for vibration signals mainly include correlation analysis for stationary signals, statistical feature analysis, spectrum analysis and time-frequency analysis for non-stationary signals, fractal theory, wavelet transform, order analysis and other new analysis methods.

The key to fault diagnosis is signal processing and fault feature extraction. In order to extract the characteristic information of faults, many advanced signal processing methods have been developed successively, among which Fast Fourier Transform (FFT) is the main means of feature extraction in fault diagnosis. The classical spectral analysis method based on Fourier transform has become the most important and basic technology in information processing. At present, almost all dynamic analysis instruments use it as the core for signal processing. However, the main disadvantage of these methods is that the phase and amplitude are separated, and the analysis results are not intuitive; only the information of a single sensor in the same section is used as the research object, and the analysis results are inaccurate. In response to this problem, scholars at home and abroad have successively proposed analysis methods based on homologous information fusion, including holographic spectrum analysis technology, full spectrum analysis technology and full vector spectrum analysis technology. Compared with single-channel information, the fused information Therefore, the fault characteristic information is more accurate and reliable.

The guiding ideology of the full-spectrum analysis method is: the eddy track of the rotor under the combination of various harmonics is an ellipse that can be decomposed into two positive and negative precession circles with the same frequency. The abscissa of the full-spectrum spectrum represents the frequency, which is different from the traditional spectrogram in that its frequency can be divided into positive and negative frequencies; its vertical coordinate represents the radius of the precession circle at each positive and negative frequency. A positive frequency indicates positive precession, and a negative frequency indicates anti-precession. The full-spectrum analysis method can quickly identify the precession direction, and then get the whole picture of the rotor vibration. The resolution of the map is high. Therefore, the vibration intensity under each harmonic cannot be accurately expressed, and energy analysis is also inconvenient.

The guiding ideology of the holographic spectrum technology is: the whirl of the rotor is formed by the combined action of each harmonic frequency, and the real motion trajectory of each harmonic frequency is an ellipse. The core method of holographic spectrum is to perform FFT transformation on the X and Y direction signals in a supporting surface of the rotor respectively, and use its unilateral spectrum to integrate the amplitude spectrum and phase spectrum of the two directions to obtain its axis trajectory. Judging from the final spectrum obtained by the holographic spectrum analysis method, the figure is very intuitive and vivid, and it can more truly describe the whirl conditions of the rotor at each harmonic frequency. However, the holographic spectrum analysis method also has its limitations. Since the holographic spectrum is used to represent the whirl of the rotor in the form of an ellipse, we can see ellipses one by one in the spectrum. It is possible to draw too many ellipses, that is to say, the resolution of the spectrum obtained by using the holographic spectrum analysis method will not be too high; in addition, if the vibration at a certain frequency is relatively severe or there are many characteristic frequency points, that is to say, the characteristic frequency difference When it is small, there will be a lot of disorderly ellipses in the obtained spectrum, and they may overlap together, which makes it difficult to distinguish the characteristic frequency, and it is also difficult to analyze the energy.

The full vector spectrum inherits the strengths of the full spectrum and the full vector spectrum well, and effectively makes up for the deficiencies of the two. The guiding ideology of the full vector spectrum is: the whirl phenomenon of the rotor is produced by the combined action of each harmonic, and the whirl intensity of the rotor at each harmonic frequency is the basic basis for its fault diagnosis and identification. Its vortex trajectory is a series of ellipses, defining the long semi-axis of the ellipse as the main vibration vector, the short semi-axis as the auxiliary vibration vector, the angle between the long semi-axis and the x-axis is the vibration vector angle, and the axis of the rotor along the The phase angle when the elliptical trajectory moves is the sagittal phase. The core method of the full vector spectrum is to first synthesize the signals in the X and Y directions of the rotor into a single-complex signal, and then perform FFT transformation on the complex signal, and extract the single-sided spectrum in the X and Y directions from the complex signal amplitude spectrum phase spectrum , and synthesized. The spectrum of the full vector spectrum has the same good frequency resolution and dynamic characteristics as the traditional spectrum, and can also be used for energy analysis. However, holographic spectrum and full vector spectrum have a common disadvantage. That is, only the unilateral spectrum of X and Y is used, and only its positive frequency spectrum is synthesized, which makes it very difficult to obtain some fault characteristic parameters and the calculation is complicated.

Negative frequency information has clear physical meaning, and it is very important information for rotating machinery. Therefore, the present invention proposes a rotating machinery fault diagnosis technology based on complex signal bilateral spectrum analysis, which is characterized in that the vibration signal is processed into a complex signal, and FFT is directly performed on it to obtain an asymmetric bilateral spectrum. The fault parameters are obtained directly by using the positive and negative half-axis spectrum. After several experiments and theoretical studies, it is found that using the complex signal bilateral spectrum technology can obtain all the existing FFT fault diagnosis parameters more simply and directly.

Contents of the invention

In view of this, the object of the present invention is to provide a method for diagnosing faults of rotating machinery based on complex signal double-sided spectrum analysis. Positive and negative semi-axis spectrum information, so that it is more convenient and direct to obtain fault parameters, and improve the practicability and effectiveness of fault diagnosis.

To achieve the above object, the present invention provides the following technical solutions:

A method for diagnosing faults of rotating machinery based on complex signal bilateral spectrum analysis, comprising the following steps:

Step 1: Signal acquisition and preprocessing:

Two sensors with the same specifications are installed in two directions perpendicular to each other, namely the x direction and the y direction. The installation conditions, signal transmission path, sampling frequency and sampling start time must be consistent;

Carry out signal preprocessing on the signals in x direction and y direction, add Hanning window to suppress interference and noise, and improve frequency identification ability;

Step 2: Complex signal synthesis and analysis:

The motion of each particle on the rotary machine can be uniquely characterized by a complex signal Z{n}=X{n})+i*Y{n};

If the rotor motion is expressed as: z=x+i*y (1)

Combined with the rotor dynamics model, the disc differential equation can be expressed as:

$\stackrel{<span\; class="notranslate">\; \&CenterDot;</span><span\; class="notranslate">\; \&CenterDot;</span>}{\mathrm{<span\; class="notranslate">\; z</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&omega;</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; z</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

Its solution is: $z=B_1{e}^{i{\mathrm{\&omega;}}_{\mathrm{no}}t}+B_2{e}^{-i{\mathrm{\&omega;}}_{\mathrm{no}}t}---\left(3\right)$

Both B ₁ and B ₂ are complex numbers. The first item is the counterclockwise movement with a radius of |B ₁ |, which is in the same direction as the rotational angular velocity Ω, which is called positive precession; the second item is the motion with a radius of |B ₂ | Clockwise motion, opposite to the rotational angular velocity Ω, is called anti-precession. The vortex at the center of the disk is the synthesis of these two precessions. Due to different initial conditions, the movement of the center of the disk may appear in the following situations:

①B ₁ ≠0, B ₂ =0; whirl is positive precession, the trajectory is a circle, and its radius is |B ₁ |.

②B ₁ ＝0, B ₂ ≠0; whirl is anti-precession, the trajectory is a circle, and its radius is |B ₂ |.

③B ₁ ＝B ₂ ; the track is a straight line, doing linear simple harmonic motion.

④B ₁ ≠B ₂ ; the trajectory is an ellipse, and when |B ₁ |>|B ₂ |, it will do positive whirl. When |B ₁ |<|B ₂ |, reverse whirl.

Step 3: FFT transformation and axis trajectory synthesis:

The axis trajectory is uniquely determined by B ₁ , B ₂ and ±ω _n , where ω _n represents the positive precession frequency, and -ω _n is the reverse precession frequency; perform FFT transformation on z to determine B ₁ , B ₂ and ±ω _n , because ω _n has positive and negative points, so it is called a double-sided spectrum; the axis trajectory at a specific frequency can be synthesized through a spectrum correction, and the corresponding fault parameters can be obtained according to the axis trajectory.

The beneficial effects of the present invention are:

1) This method is not only equivalent to the holographic spectrum in essence, but also simpler and more effective than the holographic spectrum technology. It only needs to do one FFT transformation and one spectrum correction, and there is no need to analyze the X and Y directions separately; in addition, the bilateral spectrum can be directly viewed The positive precession and anti-precession circle parameters directly reflect the full-spectrum technology of Bently; through the use of negative semi-axis frequency information, the synthesis of fault parameters is simpler and more practical.

2) By performing FFT transformation on the complex signal, the FFT transformation of the X and Y direction signals can be obtained at one time, which improves the calculation speed, enhances the real-time performance of signal analysis, and also keeps the signal analysis accuracy on X and Y consistent.

Description of drawings

In order to make the purpose, technical scheme and beneficial effect of the present invention clearer, the present invention provides the following drawings for illustration:

Fig. 1 is a schematic flow sheet of the method of the present invention;

Fig. 2 is X, Y axis time domain signal in the embodiment;

Fig. 3 is the original axis locus figure in the embodiment;

Fig. 4 is the bilateral spectrum of complex signal Z in the embodiment;

Fig. 5 is the holographic spectrum under the specific frequency in the embodiment;

Fig. 6 is the single sided spectrum of X, Y direction signal in the embodiment.

Detailed ways

The preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

Fig. 1 is a schematic flow chart of the method of the present invention, as shown in the figure, the rotating machinery fault diagnosis method based on complex signal bilateral spectrum analysis of the present invention comprises the following steps: Step 1: signal acquisition and preprocessing; Step 2 : Complex signal synthesis and analysis; Step 3: FFT transformation and axis trajectory synthesis.

Signal acquisition and preprocessing:

Two sensors with the same specifications are used in the x direction and the y direction, and their installation conditions (perpendicular to each other), the signal transmission path, sampling frequency and sampling start time must be consistent to ensure data consistency and unity;

Carry out signal preprocessing on the signals in the x direction and y direction, add the Hanning window to suppress interference and noise, and improve the ability of frequency identification.

Complex signal synthesis and analysis:

The motion of each particle on the rotary machine can be uniquely characterized by a complex signal Z{n}=X{n})+i*Y{n};

If the rotor motion is expressed as: z=x+i*y (1)

Combined with the rotor dynamics model, the disc differential equation can be expressed as:

$\stackrel{<span\; class="notranslate">\; \&CenterDot;</span><span\; class="notranslate">\; \&CenterDot;</span>}{\mathrm{<span\; class="notranslate">\; z</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&omega;</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; z</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

Its solution is: $z=B_1{e}^{i{\mathrm{\&omega;}}_{\mathrm{no}}t}+B_2{e}^{-i{\mathrm{\&omega;}}_{\mathrm{no}}t}---\left(3\right)$

Both B ₁ and B ₂ are complex numbers. The first item is the counterclockwise movement with a radius of |B ₁ |, which is in the same direction as the rotational angular velocity Ω, which is called positive precession; the second item is the motion with a radius of |B ₂ | Clockwise motion, opposite to the rotational angular velocity Ω, is called anti-precession. The vortex at the center of the disk is the synthesis of these two precessions. Due to different initial conditions, the movement of the center of the disk may appear in the following situations:

①B ₁ ≠0, B ₂ =0; whirl is positive precession, the trajectory is a circle, and its radius is |B ₁ |.

②B ₁ ＝0, B ₂ ≠0; whirl is anti-precession, the trajectory is a circle, and its radius is |B ₂ |.

③B ₁ ＝B ₂ ; the track is a straight line, doing linear simple harmonic motion.

④B ₁ ≠B ₂ ; the trajectory is an ellipse, and when |B ₁ |>|B ₂ |, it will do positive whirl. When |B ₁ |<|B ₂ |, reverse whirl.

FFT transformation and axis trajectory synthesis:

The axis trajectory is uniquely determined by B ₁ , B ₂ and ±ω _n , where ω _n represents the positive precession frequency, and -ω _n is the reverse precession frequency; perform FFT transformation on z to determine B ₁ , B ₂ and ±ω _n , because ω _n has positive and negative points, so it is called a double-sided spectrum; the axis trajectory at a specific frequency can be synthesized through a spectrum correction, and the corresponding fault parameters can be obtained according to the axis trajectory.

Example:

In this embodiment, the fault data of a fan running at 11180 rpm (186.34 Hz) is selected as a case for analysis. Among them, the sampling frequency is 2000Hz, and the data length is 1024 points. Figure 2 is the time-domain signal of x and y, and Figure 3 is the original axis trajectory diagram. It can be seen that the original axis trajectory in the time domain cannot judge the fault. Therefore, it is necessary to process this signal, mainly to process the trajectories of specific frequencies such as 0.27X, 0.42X, 1X, 2X, 3X, 4X, etc., and extract the corresponding fault parameters. First, use the method in this paper to perform FFT transformation on the complex signal z=x+i*y synthesized by x and y to obtain its bilateral spectrum, as shown in Figure 4. From Figure 4, we can see the radii of the positive and negative precession circles of each main frequency point, and at the same time, we can judge the elliptical hand direction of each frequency point. This is the full spectrum proposed by bently. In addition, the hologram in Fig. 5 can be obtained by drawing the circles of corresponding frequency points (positive and negative frequency superposition) together. It can be seen that the eccentricity of the double frequency is relatively large, and there is a fault of rotor misalignment.

At the same time, the holographic spectrum technology is used for analysis, and the process is as follows: Firstly, FFT is performed on the x and y signals to obtain the unilateral spectrum, as shown in Figure 6. Then calculate the radius of the major axis and the radius of the minor axis of the ellipse, and then calculate the radius of the positive and negative precession circles to synthesize the ellipse (as shown in Figure 5). The process is compared with this method as shown in Table 1.

Table 1 method of the present invention and holographic spectrum comparison

Finally, it should be noted that the above preferred embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail through the above preferred embodiments, those skilled in the art should understand that it can be described in terms of form and Various changes may be made in the details without departing from the scope of the invention defined by the claims.

Claims (1)

1. based on a rotary machinery fault diagnosis method for the bilateral analysis of spectrum of complex signal, it is characterized in that: comprise the following steps:

Step one: signals collecting and pre-service:

Adopt two sensors of same specification, be arranged in orthogonal both direction, be respectively x direction and y direction, its mounting condition, signal transmission path, sample frequency and sampling initial time need be consistent;

Signal Pretreatment is carried out to the signal in x direction and y direction, adds Hanning window, suppress interference and noise, improve frequency estimation ability;

Step 2: complex signal synthesis and analysis:

On rotating machinery, the motion of each particle can uniquely characterize its characteristics of motion by a complex signal Z{n}=X{n}+i*Y{n};

If rotor motion is expressed as: z=x+i*y (1)

In conjunction with dynamical model of rotor, the disk differential equation can be expressed as:

$\stackrel{\&CenterDot;\&CenterDot;}{z}+{\mathrm{\&omega;}}_n^{2}z=0---\left(2\right)$

Its solution is: $z=B_1{e}^{i{\mathrm{\&omega;}}_{n}t}+B_2{e}^{-i{\mathrm{\&omega;}}_{n}t}---\left(3\right)$

B ₁, B ₂be all plural number, to be radius be Section 1 | B ₁| counterclockwise motion, with rotational angular velocity Ω in the same way, be called positive precession; Section 2 is radius | B ₂| clockwise motion, reverse with rotational angular velocity Ω, be called backward whirl; The whirling motion of disc centre is exactly the synthesis of these two kinds of precession, and because initial conditions are different, following several situation may appear in the motion of disc centre:

1. B ₁≠ 0, B ₂=0; Whirling motion is positive precession, and track is circle, and its radius is | B ₁|;

2. B ₁=0, B ₂≠ 0; Whirling motion is backward whirl, and track is circle, and its radius is | B ₂|;

3. B ₁=B ₂; Track is straight line, does straight line simple harmonic motion;

4. B ₁≠ B ₂; Track is oval, when | B ₁| >|B ₂| time, do forward whirling motion.When | B ₁| <|B ₂| time, do backward whirling;

Step 3: FFT converts and orbit of shaft center synthesis:

Orbit of shaft center is by B ₁, B ₂and ± ω _nuniquely determine, wherein ω _nrepresent positive precession frequency ,-ω _nit is backward whirl frequency; FFT conversion is carried out to z, determines B ₁, B ₂and ± ω _n, due to ω _nthere is positive and negative dividing, so be called bilateral spectrum; By once composing the orbit of shaft center corrected and just can synthesize under characteristic frequency, just corresponding fault parameter can be obtained according to orbit of shaft center.