An improved variable working condition fault diagnosis method and system

Abstract

The application discloses an improved variable working condition fault diagnosis method and system, and belongs to the field of mechanical equipment fault diagnosis.The method comprises the following steps: acquiring vibration signal data and rotating speed signal data under a steady working condition, calculating time domain waveform signal data, transforming the time domain waveform signal data to obtain frequency spectrum data and envelope waveform data, transforming the envelope waveform data to obtain envelope spectrum data, calculating a phase discrimination time scale, resampling the vibration signal data and order signal data under a variable speed working condition according to the phase discrimination time scale, transforming the order signal data to obtain order spectrum data and order envelope waveform data, transforming the order spectrum data to obtain order envelope spectrum data, and performing fault analysis and diagnosis on the order spectrum data and the order envelope spectrum data under the variable speed working condition to obtain signal fault analysis results.The application can establish a bridge between signal analysis under a steady working condition and signal analysis under a non-steady working condition.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of mechanical equipment fault diagnosis, and particularly relates to an improved variable working condition fault diagnosis method and system. BACKGROUND

[0002] At present, with the continuous development of intelligent manufacturing, the intelligent degree of industrial systems is getting higher and higher, and at the same time, the industrial systems are becoming more and more complex, and the loss caused by equipment damage is also getting larger and larger. Due to the complexity and nonlinearity of industrial systems, it is difficult to establish an accurate mathematical model. And due to the rapid development of information technology, a large amount of operation data is generated in industrial systems, which contains a large amount of valuable equipment state information. For complex systems with high integration, data-driven fault diagnosis methods are proved to be more effective than manual models based on artificial experience.

[0003] The existing patent No. CN201210194724.5, GIS fault diagnosis system and method based on vibration signal spectrum analysis, comprises vibration acceleration sensor, charge amplifier, data acquisition instrument and PC machine which are connected in series. The vibration acceleration sensor comprises three vibration sensors fixed on the X, Y and Z axis directions of the outer surface of the GIS box. The diagnosis method of the system is to collect the GIS vibration signals in the X, Y and Z axis directions by using the three vibration sensors, to carry out wavelet denoising on the vibration signals intercepted in the whole period, and then to carry out frequency spectrum analysis, to calculate the energy at 50Hz and the energy sum at 100Hz, 200Hz and 300Hz, and to compare the calculation values with the normal state, and to determine the GIS fault according to the comparison result and the threshold value. The method is simple in engineering implementation, obvious in characteristics, and can effectively diagnose the GIS fault.

[0004] Although the acceleration and frequency spectrum analysis of the vibration signal can effectively reveal the frequency components of the signal in the whole process, it cannot reflect the law of frequency change with time. For variable speed working conditions of some equipment, especially the low speed end, a long time vibration signal needs to be collected. If the traditional frequency domain analysis method is still used for the long time collected signal, the frequency spectrum energy will be dispersed, and the spectrum line will be blurred. The order analysis can be effectively applied to the variable working condition signal, but since the frequency spectrum analysis cannot reveal the law of signal frequency spectrum change with time, once the speed is unstable, the frequency spectrum of the vibration signal will appear "blurred" phenomenon, which makes the accuracy of the fault analysis result also decrease. SUMMARY

[0005] Problems to be solved

[0006] In view of the problem that the existing frequency spectrum analysis cannot reveal the law of signal frequency spectrum change with time, so once the speed is unstable, the frequency spectrum of the vibration signal will appear "blurred" phenomenon, the present application provides an improved variable working condition fault diagnosis method and system.

[0007] Technical solution

[0008] To solve the above problems, the present invention adopts the following technical solution.

[0009] An improved fault diagnosis method under varying operating conditions comprises the following steps:

[0010] Step 1: Use sensors to synchronously collect vibration signal data and rotational speed signal data under steady-state conditions;

[0011] Step 2: Using the speed signal data as discrete points, the curve fitting technique is used to fit the functional relationship between speed, angle, and time, and the time-domain waveform signal data is calculated.

[0012] Step 3: Perform Fast Fourier Transform and Hilbert Transform on the time-domain waveform signal data to obtain the spectrum data and envelope waveform data. Then perform Fast Fourier Transform on the envelope waveform data to obtain the envelope spectrum data.

[0013] Step 4: Combine the time-domain waveform signal data with the vibration signal data to calculate the time value of the vibration signal angular domain resampling, and use it as the phase detection time marker;

[0014] Step 5: Under variable speed conditions, resampling is performed according to the phase detection time scale to obtain synchronously acquired vibration signal data and equal angle sampling signal data. The equal angle sampling signal data is used as the order signal data.

[0015] Step 6: Perform Fast Fourier Transform and Hilbert Transform on the order signal data to obtain order spectrum data and order envelope waveform data. Then perform Fast Fourier Transform on the order spectrum data to obtain order envelope spectrum data.

[0016] Step 7: Using the steady-state spectrum data and envelope spectrum data as reference signal data, perform fault analysis and diagnosis on the order spectrum data and order envelope spectrum data under variable speed conditions, and obtain the signal fault analysis results.

[0017] Preferably, the curve fitting technique is to obtain the rotational speed curve of the reference shaft by using first-order, second-order, and multi-order polynomial piecewise fitting.

[0018] Preferably, the resampling based on the phase detection time scale under variable speed conditions is performed by interpolating and resampling the vibration signal using the speed signal, thereby converting the equal-time sequence vibration signal into an equal-angle sequence vibration signal.

[0019] Furthermore, when the maximum analysis order Omax = 1 for the equal-angle sampled signal data, the Nyquist sampling theorem is used for sampling, and the sampling order Forder ≥ 2.56 * Omax, that is, the number of sampling points in one 2 * pi period is Forder.

[0020] Furthermore, the angle through which the reference axis rotates between two adjacent sampling points is delta_theta, calculated using the following formula:

[0021] delta_theta=2*pi / FOrder

[0022] Furthermore, when resampling is performed based on the phase detection time scale under variable speed conditions, and the instantaneous change in the power frequency of the rotational speed is a linear function of time, the calculation formula for the constant angle sampled signal data is as follows:

[0023] ft=5*t+2

[0024] X = cos(2*pi*ft.*t)

[0025] Where: X is the equal-angle sampled signal data, and t is the sampling duration.

[0026] Furthermore, the formula for calculating the phase detection time marker tn is as follows:

[0027] ft(tn)*tn=n / FOrder,for n=1,2,...K

[0028] Where K is the total number of resampling points.

[0029] Furthermore, when resampling is performed based on the phase detector time scale under variable speed conditions, and the instantaneous change in the power frequency of the rotational speed is a quadratic function of time, the formula for calculating the equal-angle sampled signal data X is as follows:

[0030] ft = 5 * t. 2 +2

[0031] AM = cos(2*pi*m*ft.*t)

[0032] SM1 = 1 * cos(2 * pi * i * ft. * t)

[0033] SM2 = 2*cos(2*pi*n*ft.*t)

[0034] X = AM + SM1 + SM2

[0035] Where m, i, and n are all the analysis orders of the signal data X, and the maximum analysis order Omax is the largest of these analysis order values.

[0036] Furthermore, in the case of resampling based on phase detector time scales under variable speed conditions, where the instantaneous change in rotational speed frequency is a quadratic function of time, and the gear meshing signal is modulated on the carrier wave, generating sidebands, the formula for calculating the equal-angle sampled signal data X is as follows:

[0037] ft = 5 * t. 2 +2

[0038] AM = 1 + cos(2*pi*m*ft.*t)

[0039] SM = cos(2*pi*i*ft.*t)

[0040] X = AM.*SM

[0041] Where m and i are both the analysis order of the signal data X.

[0042] An improved fault diagnosis system for variable operating conditions includes:

[0043] The steady-state signal acquisition module is used to synchronously acquire vibration signal data and rotational speed signal data under steady-state operating conditions.

[0044] The curve fitting module is used to fit the function relationship between rotational speed, rotational angle, and time using the rotational speed signal data as discrete points, and obtain time-domain waveform signal data.

[0045] The steady-state data conversion module is used to convert time-domain waveform signal data into spectrum data and envelope waveform data, and to convert envelope waveform data into envelope spectrum data.

[0046] The timing calculation module is used to calculate the phase detection time marker by combining time-domain waveform signal data with vibration signal data.

[0047] The variable speed signal acquisition module is used to synchronously acquire vibration signal data and equal angle sampling signal data under variable speed conditions according to the phase detection time scale.

[0048] The variable speed data conversion module is used to convert equal-angle sampled signal data into order spectrum data and order envelope waveform data, and to rotate the order envelope waveform data to obtain order envelope spectrum data.

[0049] The transformation analysis module is used to perform fault analysis and diagnosis on the order spectrum data and order envelope spectrum data under variable speed conditions, and to obtain the signal fault analysis results.

[0050] An improved fault diagnosis method and system for variable operating conditions is proposed. This method acquires vibration and speed signal data under steady-state conditions, calculates time-domain waveform signal data, transforms the time-domain waveform signal data to obtain spectrum data and envelope waveform data, further transforms the envelope waveform data to obtain envelope spectrum data, and calculates a phase detection timescale. Under variable speed conditions, resampling is performed based on the phase detection timescale to acquire vibration and order signal data. The order signal data is then transformed to obtain order spectrum data and order envelope waveform data, and further transformed to obtain order envelope spectrum data. Fault analysis and diagnosis are performed on the order spectrum data and order envelope spectrum data under variable speed conditions to obtain signal fault analysis results. This establishes a bridge between signal analysis under steady-state and non-steady-state conditions, improving the accuracy of fault analysis results.

[0051] Beneficial effects

[0052] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0053] (1) This invention is applied to gearboxes with a defined structure. Even if the rotational speed changes, the proportional relationship between the characteristic frequencies of the components in the gearbox will not change. By sampling at equal angles, the time-domain waveform is converted into an order waveform, and the equal-time sequence vibration signal is converted into an equal-angle sequence vibration signal. That is, no matter how much the rotational speed is, the number of sampling points within a specific angle through which the rotating shaft rotates is equal. This can effectively avoid frequency ambiguity in spectrum analysis and establish a bridge between signal analysis under steady-state conditions and signal analysis under non-steady-state conditions.

[0054] (2) This invention uses Fast Fourier Transform and Hilbert Transform to convert time-domain waveform signal data into spectrum data and envelope waveform data, converts equal-angle sampled signal data into order spectrum data and order envelope waveform data, uses Fast Fourier Transform to convert envelope waveform data into envelope spectrum data, and converts order spectrum data into order envelope spectrum data. By using the spectrum data and envelope spectrum data under steady-state conditions as reference signal data, fault analysis and diagnosis are performed on the order spectrum data and order envelope spectrum data under variable speed conditions to obtain signal fault analysis results. By synchronously converting the data, the data error rate is reduced and the analysis results are more reliable. Attached Figure Description

[0055] To more clearly illustrate the technical solutions in the embodiments or examples of this application, the accompanying drawings used in the embodiments or examples will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other drawings can be obtained according to these drawings without creative effort.

[0056] Figure 1 This is a schematic diagram of the steps of the present invention;

[0057] Figure 2 This is a schematic diagram of the process of the present invention;

[0058] Figure 3 The rotational speed and frequency diagram for Example 2;

[0059] Figure 4 The isochronous sampling diagram and spectrum diagram of Example 2;

[0060] Figure 5 This is the phase-time marker diagram for Example 2;

[0061] Figure 6 These are the isochronous sampling diagram and isoangular sampling diagram of Example 2;

[0062] Figure 7 The order spectrum of Example 2;

[0063] Figure 8 The rotational speed and frequency diagram for Example 3;

[0064] Figure 9 The isochronous sampling diagram and spectrum diagram of Example 3;

[0065] Figure 10 These are the isochronous sampling diagram and isoangular sampling diagram of Example 3;

[0066] Figure 11 The order spectrum of Example 3;

[0067] Figure 12 The rotational speed and frequency diagram for Example 4;

[0068] Figure 13 The isochronous sampling diagram and spectrum diagram of Example 4;

[0069] Figure 14 These are the isochronous sampling diagram and isoangular sampling diagram of Example 4;

[0070] Figure 15 The order spectrum of Example 4;

[0071] Figure 16 This is a diagram of the order envelope signal in Example 4;

[0072] Figure 17 The order spectrum and order envelope spectrum are from Example 4. Detailed Implementation

[0073] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of this application, but not all embodiments. Generally, the components of the embodiments of this application described and shown in the accompanying drawings can be arranged and designed in various different configurations.

[0074] Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0075] Example 1

[0076] like Figure 1 As shown, an improved fault diagnosis method under varying operating conditions employs the following steps:

[0077] First, sensors are used to synchronously acquire vibration and rotational speed signal data under steady-state conditions. The rotational speed signal data is used as discrete points, and curve fitting technology is employed to fit the functional relationship between rotational speed, rotational angle, and time, calculating the time-domain waveform signal data. Fast Fourier Transform and Hilbert Transform are then performed on the time-domain waveform signal data to obtain the spectrum and envelope waveform data. A Fast Fourier Transform is then performed on the envelope waveform data to obtain the envelope spectrum data. The time-domain waveform signal data is combined with the vibration signal data to calculate the time value of the vibration signal angular domain resampling, which is used as the phase detection time stamp during speed change. Under operating conditions, resampling is performed based on the phase detection time marker to acquire synchronously acquired vibration signal data and equal-angle sampled signal data. The equal-angle sampled signal data is used as order signal data. Fast Fourier Transform and Hilbert Transform are performed on the order signal data to obtain order spectrum data and order envelope waveform data. Fast Fourier Transform is then performed on the order spectrum data to obtain order envelope spectrum data. The spectrum data and envelope spectrum data under steady-state operating conditions are used as reference signal data. Fault analysis and diagnosis are performed on the order spectrum data and order envelope spectrum data under variable speed operating conditions to obtain signal fault analysis results.

[0078] Vibration signal data and rotation speed signal data are collected synchronously by sensors under steady-state conditions. The rotation speed signal data is used as discrete points, and the function relationship between rotation speed, rotation angle and time is fitted by curve fitting technology to calculate the time-domain waveform signal data. The curve fitting technology uses first-order, second-order and multi-order polynomial piecewise fitting to obtain the rotation speed curve of the reference shaft.

[0079] Perform Fast Fourier Transform and Hilbert Transform on the time-domain waveform signal data to obtain spectral data and envelope waveform data. Then perform Fast Fourier Transform on the envelope waveform data to obtain envelope spectrum data.

[0080] The time value of the vibration signal angular domain resampling is calculated by combining the time domain waveform signal data with the vibration signal data, and it is used as the phase detection time mark. Under variable speed conditions, resampling is performed based on the phase detection time mark by interpolating and resampling the vibration signal using the speed signal, thereby converting the equal time series vibration signal into an equal angle series vibration signal.

[0081] Under variable speed conditions, resampling is performed according to the phase detection time scale to obtain synchronously acquired vibration signal data and equal-angle sampled signal data. The equal-angle sampled signal data is used as the order signal data. When the maximum analysis order Omax of the equal-angle sampled signal data is 1, the Nyquist sampling theorem is used for sampling. The sampling order Forder ≥ 2.56*Omax, that is, the number of sampling points in one 2*pi cycle is Forder.

[0082] The angle through which the reference axis rotates between two adjacent sampling points is delta_theta, calculated using the following formula:

[0083] delta_theta=2*pi / FOrder

[0084] When resampling is performed based on the phase detection time scale under variable speed conditions, and the instantaneous change in the power frequency of the rotational speed is a linear function of time, the calculation formula for the constant angle sampled signal data is as follows:

[0085] ft=5*t+2

[0086] X = cos(2*pi*ft.*t)

[0087] Where: X is the equal-angle sampled signal data, and t is the sampling duration.

[0088] The formula for calculating the phase detection timescale tn is as follows:

[0089] ft(tn)*tn=n / FOrder,for n=1,2,...K

[0090] Where K is the total number of resampling points.

[0091] When resampling is performed based on the phase detector time scale under variable speed conditions, and the instantaneous change in the power frequency of the rotational speed is a quadratic function of time, the formula for calculating the X of the constant angle sampled signal data is as follows:

[0092] ft = 5 * t. 2 +2

[0093] AM = cos(2*pi*m*ft.*t)

[0094] SM1 = 1 * cos(2 * pi * i * ft. * t)

[0095] SM2 = 2*cos(2*pi*n*ft.*t)

[0096] X = AM + SM1 + SM2

[0097] Where m, i, and n are all the analysis orders of the signal data X, and the maximum analysis order Omax is the largest of these analysis order values.

[0098] Under variable speed conditions, resampling is performed based on the phase detection time scale. The instantaneous change in the power frequency of the rotational speed is a quadratic function of time. Furthermore, the gear meshing signal is modulated on the carrier wave, generating sidebands. The formula for calculating the equal-angle sampled signal data X is as follows:

[0099] ft = 5 * t. 2 +2

[0100] AM = 1 + cos(2*pi*m*ft.*t)

[0101] SM = cos(2*pi*i*ft.*t)

[0102] X = AM.*SM

[0103] Where m and i are both the analysis order of the signal data X.

[0104] Fast Fourier Transform and Hilbert Transform are performed on the order signal data to obtain order spectrum data and order envelope waveform data. Then, Fast Fourier Transform is performed on the order spectrum data to obtain order envelope spectrum data. The spectrum data and envelope spectrum data under steady-state conditions are used as reference signal data to perform fault analysis and diagnosis on the order spectrum data and order envelope spectrum data under variable speed conditions, and obtain signal fault analysis results.

[0105] As described above, in this example, by acquiring vibration signal data and speed signal data under steady-state conditions, time-domain waveform signal data is calculated. The time-domain waveform signal data is then transformed to obtain spectrum data and envelope waveform data. The envelope waveform data is then transformed again to obtain envelope spectrum data. A phase detection timescale is then calculated. Under variable speed conditions, resampling is performed based on the phase detection timescale to acquire vibration signal data and order signal data. The order signal data is then transformed to obtain order spectrum data and order envelope waveform data. The order spectrum data is then transformed again to obtain order envelope spectrum data. Fault analysis and diagnosis are performed on the order spectrum data and order envelope spectrum data under variable speed conditions to obtain signal fault analysis results.

[0106] Example 2

[0107] likeFigures 3-7 As shown, the rotational speed is linearly frequency-modulated to simulate the linear change of rotational speed from a low speed of 120 rpm to a high speed of 720 rpm. The sampling frequency is 1000 rpm, the sampling duration is 2 seconds, and the instantaneous power frequency change is a linear function of time.

[0108] ft=5*t+2

[0109] X = cos(2*pi*ft.*t)

[0110] As the rotating shaft accelerates, the time required for it to traverse a 2*pi angle decreases with increasing rotational frequency, leading to a corresponding reduction in the number of sampling points per period. Directly extracting the FFT spectrum results in frequency ambiguity, such as... Figure 4 .

[0111] To perform order analysis, the vibration signal needs to be interpolated and resampled using the rotation speed signal to convert the time-series vibration signal into an angle-series vibration signal. That is, regardless of the rotation speed, the number of sampling points within a specific angle through which the rotating shaft rotates is equal. This can effectively avoid frequency ambiguity in spectrum analysis.

[0112] Since X = cos(2*pi*ft.*t), the maximum analysis order of the signal is Omax = 1. Similar to the Nyquist sampling theorem, we take the sampling order FORder ≥ 2.56*Omax = 2.56. To improve accuracy, we take a slightly larger one: FORder = 20.

[0113] That is, a 2*pi period sampling FORder = 20 points, and the angle delta_theta between two adjacent sampling points is 2*pi / FOrder.

[0114] Calculate the time value of angular domain resampling of vibration signal: phase detection timescale tn

[0115] ft(tn)*tn=n / FOrder,for n=1,2,...K

[0116] (5*tn+2)*tn=n / FOrder,for n=1,2,...K

[0117] As shown in the image, the time mark is as follows Figure 5 As shown.

[0118] Based on the phase detection time marker tn, an equal-angle signal is obtained by interpolation in the original equal-time sampling signal, such as... Figure 6 As shown.

[0119] Taking the FFT of the above order signal yields the order spectrum: the horizontal axis represents the order, such as... Figure 7As shown, regardless of how the original rotational speed changes, the order spectrum always concentrates at the position Order=1.

[0120] Example 3

[0121] like Figures 8-11 As shown, the rotational speed is frequency-modulated twice to simulate the on-site rotational speed signal as an approximate quadratic function, which contains multiple orders of signals.

[0122] ft = 5 * t. 2 +2

[0123] AM = cos(2*pi*m*ft.*t)

[0124] SM1 = 1 * cos(2 * pi * i * ft. * t)

[0125] SM2 = 2*cos(2*pi*n*ft.*t)

[0126] X = AM + SM1 + SM2

[0127] m, i, and n are of order 1, 5, and 10 respectively, with Forder = 80, and the maximum order being 10.

[0128] The speed and power frequency variation curves are as follows: Figure 8 As shown.

[0129] Original signal isochronous sampling and FFT spectrum as follows Figure 9 As shown, it is difficult to obtain frequency information from it.

[0130] Equal angle sampling to obtain order signals, such as Figure 10 As shown.

[0131] Calculating the order spectrum using the order signal, such as Figure 11 As shown, the energy is concentrated at orders 1, 5, and 10.

[0132] Example 4

[0133] like Figures 12-17 As shown, the rotational speed is frequency-modulated twice, and the simulated gear meshing signal is modulated on the carrier wave to generate a sideband.

[0134] ft = 5 * t. 2 +2

[0135] AM = 1 + cos(2*pi*m*ft.*t)

[0136] SM = cos(2*pi*i*ft.*t)

[0137] X = AM.*SM

[0138] m and i are the 1st and 10th orders respectively, AM is the low-frequency signal, SM is the 10th order carrier signal, and i is the 10 teeth in the meshing signal.

[0139] Speed ​​signal such as Figure 12 As shown, the original signal and its spectrum are as follows: Figure 13 As shown, the original signal and the order signal are as follows: Figure 14 As shown, the order spectrum is as follows Figure 15 As shown, it can be seen that two sideband signals (9th and 11th order) are generated centered around the 10th order. The envelope of the order signals is taken as follows. Figure 16 As shown, the order spectrum and the order envelope spectrum are as follows: Figure 17 As shown, the envelope spectrum mainly contains first-order signals.

[0140] Example 5

[0141] like Figure 2 As shown, an improved fault diagnosis system for variable operating conditions includes:

[0142] The steady-state signal acquisition module is used to synchronously acquire vibration signal data and rotational speed signal data under steady-state operating conditions.

[0143] The curve fitting module is used to fit the function relationship between rotational speed, rotational angle, and time using the rotational speed signal data as discrete points, and obtain time-domain waveform signal data.

[0144] The steady-state data conversion module is used to convert time-domain waveform signal data into spectrum data and envelope waveform data, and to convert envelope waveform data into envelope spectrum data.

[0145] The timing calculation module is used to calculate the phase detection time marker by combining time-domain waveform signal data with vibration signal data.

[0146] The variable speed signal acquisition module is used to synchronously acquire vibration signal data and equal angle sampling signal data under variable speed conditions according to the phase detection time scale.

[0147] The variable speed data conversion module is used to convert equal-angle sampled signal data into order spectrum data and order envelope waveform data, and to rotate the order envelope waveform data to obtain order envelope spectrum data.

[0148] The transformation analysis module is used to perform fault analysis and diagnosis on the order spectrum data and order envelope spectrum data under variable speed conditions, and to obtain the signal fault analysis results.

[0149] As described above, in this example, the steady-state signal acquisition module acquires vibration signal data and rotational speed signal data under steady-state conditions. The curve fitting module uses the rotational speed signal data as discrete points to obtain time-domain waveform signal data. The steady-state data conversion module converts the time-domain waveform signal data into spectrum data, envelope waveform data, and envelope spectrum data. The time calculation module calculates the phase detection timescale. The variable speed signal acquisition module synchronously acquires vibration signal data and constant angle sampling signal data under variable speed conditions. The variable speed data conversion module converts the constant angle sampling signal data into order spectrum data, order envelope waveform data, and order envelope spectrum data. The transformation analysis module performs fault analysis and diagnosis to obtain signal fault analysis results.

[0150] The embodiments described above are merely preferred embodiments of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications, improvements, and substitutions without departing from the concept of the present invention, and these all fall within the protection scope of the present invention.

Claims

1. An improved fault diagnosis method under varying operating conditions, characterized in that, The following steps are adopted: Step 1: Use sensors to synchronously collect vibration signal data and rotational speed signal data under steady-state conditions; Step 2: Using the speed signal data as discrete points, the curve fitting technique is used to fit the functional relationship between speed, angle, and time, and the time-domain waveform signal data is calculated. Step 3: Perform Fast Fourier Transform and Hilbert Transform on the time-domain waveform signal data to obtain the spectrum data and envelope waveform data. Then perform Fast Fourier Transform on the envelope waveform data to obtain the envelope spectrum data. Step 4: Combine the time-domain waveform signal data with the vibration signal data to calculate the time value of the vibration signal angular domain resampling, and use it as the phase detection time marker; Step 5: Under variable speed conditions, resampling is performed according to the phase detection time scale to obtain synchronously acquired vibration signal data and equal angle sampling signal data. The equal angle sampling signal data is used as the order signal data. Step 6: Perform Fast Fourier Transform and Hilbert Transform on the order signal data to obtain order spectrum data and order envelope waveform data. Then perform Fast Fourier Transform on the order spectrum data to obtain order envelope spectrum data. Step 7: Using the steady-state spectrum data and envelope spectrum data as reference signal data, perform fault analysis and diagnosis on the order spectrum data and order envelope spectrum data under variable speed conditions, and obtain the signal fault analysis results.

2. The improved fault diagnosis method under varying operating conditions according to claim 1, characterized in that: The curve fitting technique uses first-order, second-order, and multi-order polynomial piecewise fitting to obtain the rotational speed curve of the reference shaft.

3. The improved fault diagnosis method under varying operating conditions according to claim 2, characterized in that: The resampling based on the phase detection time scale under variable speed conditions involves interpolating and resampling the vibration signal using the speed signal, converting the equal-time sequence vibration signal into an equal-angle sequence vibration signal.

4. An improved fault diagnosis method under varying operating conditions according to claim 3, characterized in that: When the maximum analysis order Omax = 1 for the equal-angle sampled signal data, the Nyquist sampling theorem is used for sampling, and the sampling order Forder ≥ 2.56 * Omax, that is, the number of sampling points in one 2 * pi period is Forder.

5. An improved fault diagnosis method under varying operating conditions according to claim 4, characterized in that: The angle through which the reference axis rotates between two adjacent sampling points is delta_theta, calculated using the following formula: delta_theta = 2*pi / FOrder.

6. An improved fault diagnosis method under varying operating conditions according to claim 5, characterized in that: When resampling is performed based on the phase detection time scale under variable speed conditions, and the instantaneous change in the power frequency of the rotational speed is a linear function of time, the calculation formula for the constant angle sampled signal data is as follows: ft=5*t+2 X = cos(2*pi*ft.*t) Where: X is the equal-angle sampled signal data, and t is the sampling duration.

7. An improved fault diagnosis method under varying operating conditions according to claim 6, characterized in that: The formula for calculating the phase detection time marker tn is as follows: ft(tn)*tn=n / FOrder,for n=1,2,...K Where K is the total number of resampling points.

8. An improved fault diagnosis method under varying operating conditions according to claim 7, characterized in that: When resampling is performed based on the phase detection time scale under variable speed conditions, and the instantaneous change in the power frequency of the rotational speed is a quadratic function of time, the formula for calculating the equal-angle sampled signal data X is as follows: ft=5*t. 2 +2 AM = cos(2*pi*m*ft.*t) SM1 = 1 * cos(2 * pi * i * ft. * t) SM2 = 2*cos(2*pi*n*ft.*t) X = AM + SM1 + SM2 Where m, i, and n are the analysis orders of the signal data X, and the maximum analysis order Omax is the largest of these analysis order values.

9. An improved method for diagnosing faults under varying operating conditions according to claim 8, characterized in that: When resampling is performed based on the phase detection time scale under variable speed conditions, the instantaneous change in the power frequency of the rotational speed is a quadratic function of time, and the gear meshing signal is modulated on the carrier wave, generating sidebands, the formula for calculating the equal-angle sampled signal data X is as follows: ft=5*t. 2 +2 AM = 1 + cos(2*pi*m*ft.*t) SM = cos(2*pi*i*ft.*t) X = AM.*SM Where m and i are both the analysis order of the signal data X.

10. An improved fault diagnosis system for variable operating conditions, characterized in that, include: The steady-state signal acquisition module is used to synchronously acquire vibration signal data and rotational speed signal data under steady-state operating conditions. The curve fitting module is used to fit the function relationship between rotational speed, rotational angle, and time using the rotational speed signal data as discrete points, and obtain time-domain waveform signal data. The steady-state data conversion module is used to convert time-domain waveform signal data into spectrum data and envelope waveform data, and to convert envelope waveform data into envelope spectrum data. The timing calculation module is used to calculate the phase detection time marker by combining time-domain waveform signal data with vibration signal data. The variable speed signal acquisition module is used to synchronously acquire vibration signal data and equal angle sampling signal data under variable speed conditions according to the phase detection time scale. The variable speed data conversion module is used to convert equal-angle sampled signal data into order spectrum data and order envelope waveform data, and to rotate the order envelope waveform data to obtain order envelope spectrum data. The transformation analysis module is used to perform fault analysis and diagnosis on the order spectrum data and order envelope spectrum data under variable speed conditions, and to obtain the signal fault analysis results.

Images