A mechanical fault quantitative diagnosis method and system

Abstract

The application provides a mechanical fault quantitative diagnosis method and system, the method is that: obtaining frequency spectrum data by performing frequency spectrum analysis on original vibration data of each component; determining mechanical fault types of each component according to comparison results by comparing preset fault characteristic frequencies with the frequency spectrum data; extracting first fault frequency spectrum corresponding to each component in the frequency spectrum data according to a preset rule according to the determined mechanical fault types of each component; obtaining second fault frequency spectrum by performing reconstruction processing on the first fault frequency spectrum; extracting time domain data by performing IFFT calculation on the second fault frequency spectrum data; calculating corresponding time domain characteristic values after further simplifying the time domain data; and taking the time domain characteristic values as quantitative indexes to quantitatively diagnose mechanical faults. The system has the same beneficial effects.

Description

Technical Field

[0001] This invention relates to the field of mechanical fault diagnosis technology, and in particular to a quantitative diagnostic method and system for mechanical faults. Background Technology

[0002] As a crucial supporting component, the operating status of machinery directly affects the safety and reliability of industrial equipment. As a component prone to failure, it is essential to monitor its health status and quantitatively assess its faults. By employing certain monitoring methods, fault location and quantification of fault severity can be achieved, thereby enabling more accurate maintenance and life prediction of mechanical faults. This is a key issue that needs to be addressed at present.

[0003] Current mechanical fault diagnosis methods are mainly qualitative, with few quantitative methods available, and are mostly based on vibration signals.

[0004] In terms of time-domain analysis, vibration time-domain characteristic values ​​are calculated, such as root mean square, kurtosis, impulse index, peak value and skewness index. The most suitable parameter is selected as a quantitative identification index by entropy weight. However, in addition to fault signals, vibration signals also contain strong noise and interference components, which directly affect the calculation results.

[0005] In terms of frequency domain analysis, spectral analysis is performed on the vibration signal, and the complexity of the periodic components in the spectrum is evaluated based on the Lempe l_Ziv algorithm. The more periodic signals, the more severe the fault. However, no quantitative indicator is ultimately found. Furthermore, envelope demodulation analysis is performed on the original vibration signal to extract impact features, and the amplitude of the impact frequency is selected as the quantitative standard. However, the frequency of mechanical faults often exhibits multiple orders, and using only the first-order amplitude is inaccurate, resulting in a low fault accuracy rate for this method.

[0006] In addition, there are machine learning methods such as neural networks and fuzzy recognition used to monitor the operating status of equipment. However, these methods require a large number of fault samples to accumulate and are heavily dependent on prior data analysis experience, making them unsuitable for online monitoring.

[0007] Therefore, providing a quantitative diagnostic method and system for mechanical faults that can effectively solve the above-mentioned technical problems is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0008] The purpose of this invention is to provide a quantitative diagnosis method and system for mechanical faults. This method is logically clear, safe, effective, reliable and easy to operate, and can effectively improve the accuracy of rolling bearing fault diagnosis without the need for accumulating a large number of fault samples.

[0009] Based on the above objectives, the technical solution provided by the present invention is as follows:

[0010] A quantitative diagnostic method for mechanical faults includes the following steps:

[0011] The original vibration data of each component is analyzed by spectrum analysis to obtain spectrum data;

[0012] Based on the spectrum data and the preset mechanical fault characteristic frequencies, the mechanical fault type of each component is determined;

[0013] Based on the mechanical fault type of each component, the first fault spectrum corresponding to each component is extracted from the spectrum data according to a preset rule;

[0014] The first fault spectrum is reconstructed to obtain the second fault spectrum;

[0015] Perform an I FFT on the spectrum of the second fault to extract time-domain data;

[0016] After processing the time-domain data, time-domain feature values ​​are calculated;

[0017] Based on the time-domain feature values, mechanical faults are quantitatively diagnosed.

[0018] Preferably, determining the fault type of each component based on the spectrum data and preset mechanical fault characteristic frequencies includes the following steps:

[0019] The mechanical fault characteristic frequency of the i-th component is preset;

[0020] Determine whether there exists an actual characteristic frequency in the spectrum data that is the same as the preset mechanical fault characteristic frequency of the i-th component;

[0021] If so, then it is determined that there is a mechanical fault type in the i-th component;

[0022] Repeat the above steps until the type of mechanical failure of each component is determined.

[0023] Preferably, the step of extracting the first fault spectrum corresponding to each component from the spectrum data according to a preset rule based on the fault type of each component includes the following steps:

[0024] Based on the mechanical fault type of the i-th component, extract the main spectrum of the i-th component from the spectrum data;

[0025] Repeat the above steps until the main spectrum corresponding to each component is obtained, and use the main spectrum corresponding to each component as the first fault spectrum.

[0026] Specifically, the main spectrum of the i-th component is: the actual characteristic frequency that is the same as the preset mechanical fault characteristic frequency of the i-th component and all its higher orders.

[0027] Preferably, the process of reconstructing the first fault spectrum to obtain the second fault spectrum includes the following steps:

[0028] Construct a modulo function based on the initial sampling frequency and the actual characteristic frequency in the first fault spectrum;

[0029] Determine whether the result of the remainder function is equal to 0;

[0030] If not, the first sampling frequency is reconstructed over an entire period according to the first preset rule;

[0031] The feature samples are resampled according to the first sampling frequency and interpolation method to obtain the second fault spectrum;

[0032] Specifically, the initial sampling frequency is the sampling frequency set for acquiring the original vibration data of each component.

[0033] Preferably, the first preset rule is specifically:

[0034] f scg1 =N*f0*2

[0035] Among them, f scg1 Let f0 be the first sampling frequency, f0 be the actual characteristic frequency, and N be the order of the actual characteristic frequency sequence, with the highest order being the highest order.

[0036] Preferably, the step of extracting the first fault spectrum corresponding to each component from the spectrum data according to a preset rule based on the fault type of each component further includes the following steps:

[0037] Based on the mechanical fault type of the i-th component, extract the main spectrum and modulation spectrum of the i-th component from the spectrum data;

[0038] Repeat the above steps until the main spectrum corresponding to each component is obtained, and use the main spectrum corresponding to each component and the modulation spectrum as the first fault spectrum;

[0039] Specifically, the main spectrum of the i-th component is: the actual characteristic frequency that is the same as the preset mechanical fault characteristic frequency of the i-th component and all its higher orders;

[0040] The modulation spectrum specifically refers to the preset modulation frequency and all its higher orders.

[0041] Preferably, the process of reconstructing the first fault spectrum to obtain the second fault spectrum includes the following steps:

[0042] Obtain the least common multiple frequency of the modulation frequency and the actual characteristic frequency in the first fault spectrum;

[0043] Based on the least common multiple frequency, the initial sampling frequency, and the second preset rule, the second sampling frequency is reconstructed over an entire cycle;

[0044] The feature samples are resampled according to the second sampling frequency and interpolation method to obtain the second fault spectrum;

[0045] Specifically, the initial sampling frequency is the sampling frequency set for acquiring the original vibration data of each component.

[0046] Preferably, the second preset rule is as follows:

[0047] L = int(f s / 2 / f lcm )

[0048] Where L is the integer number of cycles, int is the downward fetching function, and f s f is the initial sampling frequency; lcm The least common multiple frequency;

[0049] f scg2 =L*f lcm *2

[0050] Among them, f scg2 This is the second sampling frequency.

[0051] Preferably, the step of processing the time-domain data and calculating the time-domain feature values ​​includes the following steps:

[0052] Based on the IFFT principle, the time-domain data is simplified.

[0053] The time-domain feature values ​​are calculated and obtained based on the simplified time-domain data.

[0054] Preferably, the step of quantitatively diagnosing mechanical faults based on the time-domain feature values ​​further includes the following steps:

[0055] The time-domain feature values ​​are normalized.

[0056] Based on the processed time-domain feature values ​​and the preset parameters of each component, the fault values ​​of each component are obtained.

[0057] Based on the fault value and the preset fault threshold, the mechanical faults of each component are quantitatively diagnosed.

[0058] Preferably, the step of performing spectral analysis on the original vibration data of each component to obtain spectral data further includes the following steps:

[0059] Wavelet analysis and envelope demodulation / resonance demodulation are performed sequentially on the original vibration data of each component to obtain original vibration feature samples;

[0060] The original vibration feature samples are subjected to spectral analysis to obtain the spectral data.

[0061] A quantitative diagnostic system for mechanical faults, comprising:

[0062] The spectrum analysis module is used to perform spectrum analysis on the raw vibration data of each component to obtain spectrum data;

[0063] The fault type module is used to determine the mechanical fault type of each component based on the spectrum data and the preset mechanical fault characteristic frequency.

[0064] The first fault spectrum module is used to extract the first fault spectrum corresponding to each component from the spectrum data according to a preset rule based on the mechanical fault type of each component.

[0065] The reconstruction module is used to reconstruct the first fault spectrum to obtain the second fault spectrum;

[0066] The extraction module is used to perform I FFT calculation on the spectrum of the second fault to extract time-domain data;

[0067] The processing module is used to process the time-domain data and calculate the time-domain feature values;

[0068] The fault diagnosis module is used to quantitatively diagnose mechanical faults based on the time-domain feature values.

[0069] The mechanical fault quantitative diagnosis method provided by this invention involves performing spectral analysis on the original vibration data of each component to obtain spectral data; comparing the spectral data with preset fault characteristic frequencies to determine the mechanical fault type of each component based on the comparison results; extracting the first fault spectrum corresponding to each component from the spectral data according to preset rules based on the determined mechanical fault type; reconstructing the first fault spectrum to obtain the second fault spectrum; performing I FFT calculation on the second fault spectrum data to extract time-domain data; further simplifying the time-domain data and calculating the corresponding time-domain feature values; and using the time-domain feature values ​​as quantitative indicators for quantitative diagnosis of mechanical faults.

[0070] This method does not require the accumulation of a large number of fault samples, nor does it rely on prior data analysis experience. Furthermore, the method employed here first qualitatively classifies the faults of each mechanical component. Based on the identified mechanical faults, according to different preset rules, a first fault spectrum corresponding to the fault type is extracted from the spectral data. Due to interference in the first fault spectrum, it is reconstructed to obtain more accurate second spectrum data. Then, after performing I-FFT calculations on the second fault spectrum, time-domain analysis is used to quantitatively diagnose the mechanical fault.

[0071] Compared to existing technologies, this case combines the advantages of frequency domain analysis and time domain analysis. It uses frequency domain analysis to characterize the bearing fault type, and then uses spectrum reconstruction to simplify the time domain feature values, removing interference components and effectively improving the accuracy of rolling bearing fault diagnosis.

[0072] The present invention also provides a quantitative diagnostic system for mechanical faults. Since the system and the method solve the same technical problem and belong to the same technical concept, they should have the same beneficial effects, and will not be described in detail here. Attached Figure Description

[0073] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0074] Figure 1 A flowchart of a quantitative diagnosis method for mechanical faults provided by the present invention;

[0075] Figure 2 A flowchart of step S2 provided in an embodiment of the present invention;

[0076] Figure 3 A flowchart illustrating one embodiment of step S3 provided in this invention;

[0077] Figure 4 A flowchart illustrating one embodiment of step S4 provided in this invention;

[0078] Figure 5 A flowchart illustrating another implementation of step S3 provided in an embodiment of the present invention;

[0079] Figure 6 A flowchart illustrating another implementation of step S4 provided in an embodiment of the present invention;

[0080] Figure 7 A flowchart of step S6 provided in an embodiment of the present invention;

[0081] Figure 8 A flowchart illustrating another implementation of step S7 provided in this embodiment of the invention;

[0082] Figure 9 This is a schematic diagram of the structure of a quantitative diagnostic system for mechanical faults provided by the present invention. Detailed Implementation

[0083] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0084] The embodiments of this invention are written in a progressive manner.

[0085] This invention provides a method and system for quantitative diagnosis of mechanical faults. It primarily addresses the technical problems in existing fault diagnosis methods, such as inaccurate fault diagnosis results due to signal interference, and the need for a large accumulation of fault samples to improve accuracy.

[0086] like Figure 1 As shown, a quantitative diagnosis method for mechanical faults includes the following steps:

[0087] S1. Perform spectral analysis on the original vibration data of each component to obtain spectral data;

[0088] S2. Determine the mechanical fault type of each component based on the spectrum data and the preset mechanical fault characteristic frequencies;

[0089] S3. Based on the mechanical fault type of each component, extract the first fault spectrum corresponding to each component from the spectrum data according to preset rules;

[0090] S4. Reconstruct the spectrum of the first fault to obtain the spectrum of the second fault;

[0091] S5. Perform IFFT calculation on the spectrum of the second fault to extract time-domain data;

[0092] S6. After processing the time-domain data, calculate the time-domain characteristic values;

[0093] S7. Quantitatively diagnose mechanical faults based on time-domain characteristic values.

[0094] Before step S1, the original vibration signal X(n) of the bearing is acquired by a sensor, with an initial sampling frequency of f. s The total sampling time is T, the number of sample points is L, and the total spectrum bandwidth is The spectral resolution is df;

[0095] In step S1, the original vibration data of each component are subjected to spectral analysis to obtain spectral data H(f);

[0096] It should be noted that spectrum analysis refers to the method of transforming a time-domain signal to the frequency domain for analysis. The purpose of spectrum analysis is to decompose a complex time-history waveform into several individual harmonic components through Fourier transform in order to obtain the frequency structure of the signal and the information of each harmonic and phase.

[0097] In step S2, the preset typical mechanical fault characteristic frequency f0 has a period of As shown in the table below:

[0098] Fault type Fault characteristic frequency Cage contact with outer ring failure <![CDATA[f bw =(D0-dcosA)×fn / (2D0)]]> Cage contact with inner ring failure <![CDATA[f bn =(D0+dcosA)×fn / (2D0)]]> Outer ring fault <![CDATA[f w =(D0-dcosA)×Z×fn / (2D0)]]> Inner loop fault <![CDATA[f n =(D0+dcosA)×Z×fn / (2D0)]]> Roller end face failure <![CDATA[f D =(D0 2 -d 2 cos 2 A)×fn / (2D0×d)]]> Roller circumferential failure <![CDATA[f S =(D0 2 -d 2 cos 2 A)×fn / (D0×d)]]>

[0099] In the table, D0 (mm) is the bearing's mean diameter, d (mm) is the rolling element diameter, and f n (Hz) represents the relative rotational speed frequency between the inner and outer rings, Z represents the number of rolling elements, and A represents the contact angle of the bearing rolling elements;

[0100] Based on the spectrum data and the mechanical fault characteristics in the table above, determine the mechanical fault types of each component as shown in the table above;

[0101] In step S3, based on the mechanical fault type of each component, the first fault spectrum corresponding to each component is extracted from the spectrum data;

[0102] In this embodiment, the fault frequency extracted from the fault spectrum is determined to correspond to one or more fault types in the table above;

[0103] In step S4, considering that for the discrete signal being processed, due to the truncation of the time domain and insufficient spectral resolution leading to energy leakage, the true frequency often cannot be aligned with the spectral line, further integer period reconstruction processing is required to restore the true spectral line. Therefore, the first fault spectrum is reconstructed to obtain a more accurate second fault spectrum.

[0104] In step S5, the second fault spectrum is subjected to I FFT calculation to extract the corresponding time-domain data;

[0105] It should be noted that I FFT refers to Inverse Fourier Transform. One of its calculation principles is to take the conjugate complex number (invert the imaginary part) of the frequency domain data (note that the frequency domain is a complex number) and then perform the FFT transformation, thus converting the frequency domain signal to the time domain.

[0106] In step S6, after simplifying the acquired time-domain data, the time-domain feature values ​​are calculated and obtained.

[0107] In step S7, based on the acquired time-domain feature values, the mechanical fault is quantitatively diagnosed according to preset rules;

[0108] In this embodiment, threshold values ​​for characteristic indicators are preset according to the operating conditions of different machines, and quantitative hierarchical diagnosis is achieved based on the calculated fault time-domain characteristic values. For example:

[0109] Impact samples generated by bearing failures, after feature extraction and calculation as described above, are assessed for fault severity using time-domain peak-to-peak value (SV), with a 3-level threshold M1 set. SV M2 SV M3 SV Quantitative grading diagnosis is performed based on the SV value, i.e.:

[0110] SV≤M1 SV It is a minor fault;

[0111] M1 SV <SV≤M2 SV This is a moderate fault.

[0112] M2 SV <SV≤M3 SV This is a serious malfunction.

[0113] like Figure 2 As shown, preferably, step S2 includes the following steps:

[0114] A1. Preset the mechanical fault characteristic frequency of the i-th component;

[0115] A2. Determine whether there exists an actual characteristic frequency in the spectrum data that is the same as the preset mechanical fault characteristic frequency of the i-th component;

[0116] A3. If so, then determine that there is a mechanical fault type in the i-th component;

[0117] A4. Repeat the above steps until the type of mechanical failure of each component is determined.

[0118] In step A1, the mechanical fault characteristic frequency f0 of the i-th component is preset;

[0119] In step A2, it is determined whether there is an actual fault frequency in the spectrum data H(f) that is the same as the preset mechanical fault characteristic frequency f0 of the i-th component, also denoted by f0;

[0120] In step A3, if it exists, then it is determined that there is a mechanical fault type of the i-th component;

[0121] In this embodiment, it is assumed that the mechanical fault characteristic frequency f0 of the i-th component is: fw=(D0-dcosA)×Z×fn / (2D0). It is determined whether there is a mechanical fault frequency with the same frequency in the spectrum data H(f). If it exists in the spectrum data H(f), it is determined that there is an outer loop fault.

[0122] In step A4, repeat steps A1 to A3 until several mechanical fault frequencies corresponding to each component are found in the spectrum data H(f), and determine the mechanical fault type of each component according to the table above based on the several mechanical fault frequencies corresponding to each component.

[0123] like Figure 3 As shown, preferably, step S3 includes the following steps:

[0124] B1. Based on the mechanical fault type of the i-th component, extract the main spectrum of the i-th component from the spectrum data;

[0125] B2. Repeat the above steps until the main spectrum corresponding to each component is obtained, and use the main spectrum corresponding to each component as the first fault spectrum.

[0126] Specifically, the main spectrum of the i-th component is: the actual characteristic frequency that is the same as the preset mechanical fault characteristic frequency of the i-th component and all its higher orders.

[0127] In step B1, based on the obtained mechanical fault type of the i-th component, the main spectrum of the i-th component is extracted from the spectrum data, that is, the actual characteristic frequency f0 and all its higher orders that are the same as the preset mechanical fault characteristic frequency of the i-th component.

[0128] In this embodiment, since the characteristic frequency f0 of mechanical faults generally exhibits multi-order characteristics, the highest order of the order sequence is N = int(f s / 2 / f0)(int: round down), extract the characteristic frequencies and all their higher orders to obtain the fault master spectrum F. z0 =[H(f0),H(2f0),H(3f0),…,H(N*f0)]. F z0 This can represent a fault equivalent to a fixed component of the sensor (positional fixed fault), in which the amplitude of the signal is almost uniform.

[0129] In step B2, the above steps are repeated until the main spectrum corresponding to each component is obtained, and the main spectrum corresponding to each component is used as the first fault spectrum, that is, only f0 and its higher order exist in the first fault spectrum.

[0130] like Figure 4 As shown, preferably, step S4 includes the following steps:

[0131] C1. Construct a remainder function based on the initial sampling frequency and the actual characteristic frequencies in the spectrum of the first fault;

[0132] C2. Determine whether the result of the remainder function is equal to 0;

[0133] C3. If not, then reconstruct the first sampling frequency for the entire cycle according to the first preset rule;

[0134] C4. Resample the feature samples according to the first sampling frequency and interpolation method to obtain the second fault spectrum;

[0135] The initial sampling frequency is specifically the sampling frequency set to acquire the original vibration data of each component.

[0136] In step C1, based on the initial sampling frequency f s Construct a modulo function with the actual characteristic frequency f0 in the spectrum of the first fault;

[0137] In this embodiment, a modulo function A = mod(f) is constructed. s / 0);

[0138] In step C2, it is determined whether the result of the above formula is equal to 0, that is, whether it is an integer cycle sampling;

[0139] In step C3, if not, then f s For sampling of f0 as a non-integer period, the first sampling frequency is reconstructed as an integer period according to the first preset rule;

[0140] In step C4, the first sampling frequency is used as the new sampling frequency, and the feature samples are resampled using interpolation to obtain the second fault spectrum.

[0141] It should be noted that resampling of discrete signals can be achieved using interpolation methods. Common interpolation methods include Lagrange interpolation, Newton interpolation, cubic spline interpolation, and linear interpolation.

[0142] After the feature samples are reconstructed over an integer period, the true frequency is corrected, resulting in a more accurate second fault spectrum F = [F z ,F T ].

[0143] Preferably, the first preset rule is as follows:

[0144] f scg1 =N*f0*2

[0145] Among them, f scg1 f0 is the first sampling frequency, f0 is the actual characteristic frequency, and N is the highest order of the order sequence of the actual characteristic frequencies.

[0146] In this embodiment, according to 1 / f scg1 The sampling interval forms a new time series T0 = 0:1 / scg1 :*T f0 .

[0147] like Figure 5 As shown, preferably, step S3 further includes the following steps:

[0148] D1. Based on the mechanical fault type of the i-th component, extract the main spectrum and modulation spectrum of the i-th component from the spectrum data;

[0149] D2. Repeat the above steps until the main spectrum corresponding to each component is obtained, and use the main spectrum and modulation spectrum corresponding to each component as the first fault spectrum;

[0150] Specifically, the main spectrum of the i-th component is: the actual characteristic frequency that is the same as the preset mechanical fault characteristic frequency of the i-th component and all its higher orders;

[0151] The modulation spectrum specifically refers to the preset modulation frequency and all its higher orders.

[0152] In step D1, based on the acquired mechanical fault type of the i-th component, the main spectrum and modulation spectrum are extracted from the spectrum data, namely the actual characteristic frequency f0 that is the same as the preset mechanical fault characteristic frequency of the i-th component, and all its higher-order and preset modulation frequencies f0. t0 and all its higher orders;

[0153] In this embodiment, there is a component fault that operates systematically relative to the sensor's mounting position (a cyclical fault). The distance between the fault signal point and the detection point changes, or due to unilateral mechanical load bearing, the fault signal is only generated in the load-bearing area, resulting in regular amplitude fluctuations. For example, in most machines, the outer ring does not rotate, while the inner ring rotates. If the sensor is mounted on a stationary mechanical base, the inner ring signal will exhibit amplitude modulation with one revolution. The modulation frequency f is further extracted to address this. t0 and its higher order, the highest order of the order sequence N1 = int(f s / 2 / f t0 (int: round down) to obtain the modulation spectrum F T0 =[H(t) t0 ),H(2f t0 ),H(3f t0 ),…,H(N1*f t0 )] and the fault main spectrum F z0 =[G(f0),H(2f0),H(3f0),…,H(N*f0)];

[0154] In step D2, the above steps are repeated until the main spectrum corresponding to each component is obtained. The main spectrum and modulation spectrum corresponding to each component are used as the first fault spectrum F0 = [F z0 ,F T0 ].

[0155] like Figure 6 As shown, preferably, step S4 includes the following steps:

[0156] E1. Obtain the least common multiple of the modulation frequency and the actual characteristic frequencies in the spectrum of the first fault;

[0157] E2. Based on the least common multiple frequency, the initial sampling frequency, and the second preset rule, reconstruct the second sampling frequency over the entire cycle;

[0158] E3. Resample the feature samples according to the second sampling frequency and interpolation method to obtain the second fault spectrum;

[0159] The initial sampling frequency is specifically the sampling frequency set to acquire the original vibration data of each component.

[0160] In step E1, according to the modulation frequency f t0 The least common multiple frequency between the actual characteristic frequency f0 in the spectrum of the first fault is obtained;

[0161] In this embodiment, the least common multiple is f. lcm ;

[0162] In step E2, based on the least common multiple f lcm Modulation frequency f t0 And the second preset rule, reconstructing the second sampling frequency over the entire cycle;

[0163] In step E3, the second sampling frequency is used as the new sampling frequency, and the feature samples are resampled using interpolation to obtain the second fault spectrum.

[0164] It should be noted that after the feature samples are reconstructed over an integer period, the true frequency is corrected, resulting in a more accurate second fault spectrum F = [F]. z ,F T ].

[0165] Preferably, the second preset rule is as follows:

[0166] L = int(f s / 2 / f lcm )

[0167] Where L is the integer number of cycles, int is the downward fetching function, and f s f is the initial sampling frequency; lcm The least common multiple frequency;

[0168] f scg2 =L*f lcm *2

[0169] Among them, f scg2 This is the second sampling frequency.

[0170] In this embodiment, the least common multiple f of the actual characteristic frequency and the modulation frequency is assumed to be... lcm=lcm(f0,f t0 Corresponding period T flcm According to 1 / f scg2 The sampling interval forms a new time series T0 = 0:1 / f scg2 :L*T lcm .

[0171] like Figure 7 As shown, preferably, step S6 includes the following steps:

[0172] F1. Based on the IFFT principle, simplify the time-domain data;

[0173] F2. Calculate and obtain time-domain feature values ​​based on the simplified time-domain data.

[0174] It should be noted that the working principle of IFFT is that since the result of FFT transformation is a complex number, the result of FFT transformation from the frequency domain is also a complex number. At this time, only the real part of the complex number needs to be taken, which is the original time domain signal (at the same time, the imaginary part of this complex number is already very small, so it is omitted).

[0175] In step F1, due to the unavoidable sidelobe spectral leakage, abnormal waveforms are generated at the beginning and end. The signals of the first and Nth periods T are removed, and the time domain data is simplified to obtain the simplified time domain data (i.e., the original time domain signal).

[0176] It should be noted that feature extraction mainly involves extracting statistical indicators; other features require the integration of domain knowledge for extraction. Statistical indicators include dimensional and dimensionless indicators. Dimensional indicators include: maximum value (max_y), minimum value (min_y), median (median_y), mean (mean_y), variance (var_y), peak value (peak), peak-to-peak ratio (peak2peak), and RMS (rms), etc. Dimensionless indicators include: crest factor (crestf), margin factor (margin), impulse factor (pulse), waveform factor (waveform), kurtosis factor (kur), and skewness factor (sk), etc.

[0177] In signal characterization, dimensional indices, while sensitive to signal characteristics, also vary with operating conditions (such as load) and are highly susceptible to environmental interference, exhibiting inherent instability. In contrast, dimensionless indices can eliminate the influence of these perturbation factors and are therefore widely used in feature extraction.

[0178] In step F2, in this embodiment, the time-domain feature value S can be selected as a fault feature indicator, such as peak value, effective value, or average value.

[0179] like Figure 8As shown, preferably, step S7 further includes the following steps:

[0180] G1. Normalizes time-domain eigenvalues;

[0181] G2. Based on the processed time-domain feature values ​​and the preset parameters of each component, obtain the fault value of each component;

[0182] G3. Based on the fault value and the preset fault threshold, quantitatively diagnose the mechanical faults of each component.

[0183] In steps G1 to G2, further, based on eliminating the influence of operating conditions on time-domain feature values, normalization processing can be performed according to the formula to calculate and obtain the fault quantity value;

[0184] In this embodiment, the impact time-domain characteristic value SV (obtained based on peak value, peak-to-peak value, effective value, and average value) is used as an example. This time-domain characteristic value is affected by the rotational speed and diameter of the shaft. The fault value is obtained through a normalization formula.

[0185] The specific formula is as follows:

[0186]

[0187] Among them, A dB SV is the fault value (indicating the severity of the fault); N is the rotational speed (r / min) of the shaft where the faulty component is located; and D is the shaft diameter (mm) of the shaft where the faulty component is located.

[0188] In step G3, after obtaining the fault value, a fault threshold is preset to quantitatively diagnose the mechanical faults of each component.

[0189] In this embodiment, M1 is set. dB M2 dB and M3 dB As a fault threshold, by comparing the fault value with the fault threshold, the degree of fault in each component of the bearing can be effectively diagnosed, as follows:

[0190] When A dB ≤M1 dB At that time, it was a minor malfunction;

[0191] When M1 dB <A dB ≤M2 dB At that time, it was a moderate fault;

[0192] When M2 dB <A dB ≤M3 dB At that time, it was a serious malfunction.

[0193] Preferably, step S1 further includes the following steps:

[0194] Wavelet analysis and envelope demodulation / resonance demodulation were performed sequentially on the original vibration data of each component to obtain the original vibration feature samples;

[0195] Spectral analysis was performed on the original vibration characteristic samples to obtain spectral data.

[0196] In practical applications, the original vibration data of each component can be processed in advance to obtain original vibration characteristic samples, and then spectrum analysis can be performed to obtain spectrum data.

[0197] It should be noted that wavelet analysis, or wavelet transform, refers to representing a signal using an oscillating waveform of finite length or rapidly decaying length, called the mother wavelet. This waveform is scaled and translated to match the input signal.

[0198] Envelope demodulation and resonance demodulation are very important tools in diagnostic analysis, used to demodulate the modulated signal. In other words, modulation is one process and demodulation is another.

[0199] The basic principle of envelope demodulation technology is that the low-frequency impulse pulse caused by the fault excites a high-frequency resonant waveform. By enveloping, detecting, and low-pass filtering (i.e. demodulating) it, a resonant demodulated waveform corresponding to the low-frequency impulse but amplified and broadened will be obtained.

[0200] The basic principle of resonant demodulation technology is as follows: Localized damage to the bearing causes impact spike signals. Because these impact spikes have a wide frequency band, they resonate after being received by the sensor, outputting a reduced-amplitude oscillation signal with a frequency close to the sensor's natural frequency. Compared to the original spike signal, this oscillation signal is amplified in amplitude and broadened in time, thus highlighting the fault information. A high-center-frequency bandpass filter effectively isolates the impact fault information from low-frequency interference signals, improving the signal-to-noise ratio. Finally, the signal output from the filter is demodulated, such as by taking the absolute value and extracting the envelope, to obtain the low-frequency resonant demodulated signal.

[0201] like Figure 9 As shown, a quantitative diagnostic system for mechanical faults includes:

[0202] The spectrum analysis module is used to perform spectrum analysis on the raw vibration data of each component to obtain spectrum data;

[0203] The fault type module is used to determine the mechanical fault type of each component based on spectrum data and preset mechanical fault characteristic frequencies.

[0204] The first fault spectrum module is used to extract the first fault spectrum corresponding to each component from the spectrum data according to preset rules based on the mechanical fault type of each component.

[0205] The reconstruction module is used to reconstruct the first fault spectrum to obtain the second fault spectrum;

[0206] The extraction module is used to perform I FFT calculation on the spectrum of the second fault to extract time-domain data;

[0207] The processing module is used to process the time-domain data and calculate the time-domain feature values.

[0208] The fault diagnosis module is used to quantitatively diagnose mechanical faults based on time-domain feature values.

[0209] This invention also discloses a quantitative mechanical fault diagnosis system. The system uses a spectrum analysis module to perform spectrum analysis on the raw vibration data of each component, obtaining spectrum data, which is then sent to a fault type module. The fault type module determines the mechanical fault type of each component based on the spectrum data and preset mechanical fault characteristic frequencies, and sends the mechanical fault type and spectrum data of each component to a first fault spectrum module. The first fault spectrum module extracts the corresponding first fault spectrum for each component from the spectrum data according to preset rules based on the mechanical fault type of each component, and sends the corresponding first fault spectrum to a reconstruction module. The reconstruction module reconstructs the first fault spectrum to obtain a second fault spectrum and sends the second fault spectrum to an extraction module. The extraction module performs I-FFT calculation on the second fault spectrum to extract time-domain data and sends the time-domain data to a processing module. The processing module processes the time-domain data, calculates time-domain feature values, and sends the time-domain feature values ​​to a normalization module in the fault diagnosis module. After processing the time-domain feature values, the normalization module performs quantitative diagnosis of the mechanical fault through the fault diagnosis module.

[0210] In the embodiments provided in this application, it should be understood that the disclosed methods and systems can be implemented in other ways. The system embodiments described above are merely illustrative. For example, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple modules or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or modules, and can be electrical, mechanical, or other forms.

[0211] Furthermore, in the various embodiments of the present invention, each functional module can be fully integrated into a processor, or each module can be a separate device, or two or more modules can be integrated into a device; each functional module in the various embodiments of the present invention can be implemented in hardware or in the form of hardware plus software functional units.

[0212] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by program instructions and related hardware. The aforementioned program instructions can be stored in a computer-readable storage medium. When the program instructions are executed, they perform the steps of the above method embodiments. The aforementioned storage medium includes various media that can store program code, such as mobile storage devices, read-only memory (ROM), magnetic disks, or optical disks.

[0213] It should be understood that the use of terms such as "system," "device," "unit," and / or "module" in this application is merely one method of distinguishing different components, elements, parts, sections, or assemblies at different levels. However, if other terms can achieve the same purpose, they may be replaced by other expressions.

[0214] As indicated in this application and claims, unless the context clearly indicates otherwise, the words "a," "an," "a," and / or "the" are not specifically singular and may include the plural. Generally, the terms "comprising" and "including" only indicate the inclusion of expressly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements. An element defined by the phrase "comprising an..." does not exclude the presence of other identical elements in the process, method, product, or apparatus that includes the element.

[0215] In the description of the embodiments of this application, unless otherwise stated, " / " means "or", for example, A / B can mean A or B; "and / or" in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Furthermore, in the description of the embodiments of this application, "multiple" refers to two or more.

[0216] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature.

[0217] If a flowchart is used in this application, it is used to illustrate the operations performed by the system according to embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed in exact order. Instead, the steps can be processed in reverse order or simultaneously. Furthermore, other operations can be added to these processes, or one or more steps can be removed from them.

[0218] The foregoing has provided a detailed description of a method and system for quantitative diagnosis of mechanical faults provided by the present invention. The above description of the disclosed embodiments enables those skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for quantitative diagnosis of mechanical faults, characterized in that, Includes the following steps: The original vibration data of each component is analyzed by spectrum analysis to obtain spectrum data; Based on the spectrum data and the preset mechanical fault characteristic frequencies, the mechanical fault type of each component is determined, wherein the preset mechanical fault characteristic frequencies are preset fault characteristic frequencies corresponding to the mechanical fault types of each component. Based on the mechanical fault type of each component, the first fault spectrum corresponding to each component is extracted from the spectrum data according to a preset rule; The first fault spectrum is reconstructed to obtain the second fault spectrum; IFFT calculation is performed on the spectrum of the second fault to extract time-domain data; After processing the time-domain data, time-domain feature values ​​are calculated; Based on the time-domain feature values, mechanical faults are quantitatively diagnosed.

2. The quantitative diagnosis method for mechanical faults as described in claim 1, characterized in that, The step of determining the fault type of each component based on the spectrum data and preset mechanical fault characteristic frequencies includes the following steps: Preset the first corresponding to the mechanical fault type i The characteristic frequencies of mechanical failures of components, among which, i This is the component number, used to iterate through each component; Determine whether there exists a value in the spectrum data that matches the preset first value. i The actual characteristic frequency is the same as the mechanical failure characteristic frequency of the component; If so, then determine the first i The component exists in the first i The mechanical failure type corresponding to the characteristic frequency of mechanical failure of a component; Repeat the above steps until the type of mechanical failure of each component is determined.

3. The quantitative diagnosis method for mechanical faults as described in claim 2, characterized in that, The step of extracting the first fault spectrum corresponding to each component from the spectrum data according to a preset rule based on the fault type of each component includes the following steps: According to the first i The mechanical failure type of the component is extracted from the spectrum data. i The main spectrum of the component; Repeat the above steps until the main spectrum corresponding to each component is obtained, and use the main spectrum corresponding to each component as the first fault spectrum. Among them, the first i The main spectrum of the component is specifically: compared with the preset first... i The mechanical failure characteristic frequency of the component is the same as its actual characteristic frequency and all its higher orders.

4. The quantitative diagnosis method for mechanical faults as described in claim 3, characterized in that, The process of reconstructing the first fault spectrum to obtain the second fault spectrum includes the following steps: Construct a modulo function based on the initial sampling frequency and the actual characteristic frequency in the first fault spectrum; Determine whether the result of the remainder function is equal to 0; If not, then the first sampling frequency is reconstructed for the entire cycle according to the first preset rule; The feature samples are resampled according to the first sampling frequency and interpolation method to obtain the second fault spectrum; Specifically, the initial sampling frequency is the sampling frequency set for acquiring the original vibration data of each component.

5. The quantitative diagnosis method for mechanical faults as described in claim 4, characterized in that, The first preset rule is as follows: ； in, The first sampling frequency, For actual characteristic frequencies, It is the highest order of the order sequence of actual characteristic frequencies.

6. The quantitative diagnosis method for mechanical faults as described in claim 2, characterized in that, The step of extracting the first fault spectrum corresponding to each component from the spectrum data according to a preset rule based on the fault type of each component further includes the following steps: According to the first i The mechanical failure type of the component is extracted from the spectrum data. i The main spectrum and modulation spectrum of the component; Repeat the above steps until the main spectrum corresponding to each component is obtained, and use the main spectrum corresponding to each component and the modulation spectrum as the first fault spectrum; Among them, the first i The main spectrum of the component is specifically: compared with the preset first... i The actual characteristic frequency and all its higher orders are the same as the mechanical failure characteristic frequency of the component; The modulation spectrum specifically refers to the preset modulation frequency and all its higher orders.

7. The quantitative diagnosis method for mechanical faults as described in claim 6, characterized in that, The process of reconstructing the first fault spectrum to obtain the second fault spectrum includes the following steps: Obtain the least common multiple frequency of the modulation frequency and the actual characteristic frequency in the first fault spectrum; Based on the least common multiple frequency, the initial sampling frequency, and the second preset rule, the second sampling frequency is reconstructed over an entire cycle; The feature samples are resampled according to the second sampling frequency and interpolation method to obtain the second fault spectrum; Specifically, the initial sampling frequency is the sampling frequency set for acquiring the original vibration data of each component.

8. The quantitative diagnosis method for mechanical faults as described in claim 7, characterized in that, The second preset rule is as follows: ； in, For integer period numbers, This is the floor function. This is the initial sampling frequency; The least common multiple frequency; ； in, This is the second sampling frequency.

9. The quantitative diagnosis method for mechanical faults as described in claim 1, characterized in that, The step of processing the time-domain data and calculating the time-domain feature values ​​includes the following steps: Based on the IFFT principle, the time-domain data is simplified. The time-domain feature values ​​are calculated and obtained based on the simplified time-domain data.

10. The method for quantitative diagnosis of mechanical faults as described in claim 1, characterized in that, The quantitative diagnosis of mechanical faults based on the time-domain feature values ​​further includes the following steps: The time-domain feature values ​​are normalized. Based on the processed time-domain feature values ​​and the preset parameters of each component, the fault values ​​of each component are obtained. Based on the fault value and the preset fault threshold, the mechanical faults of each component are quantitatively diagnosed.

11. The quantitative diagnosis method for mechanical faults as described in claim 1, characterized in that, The step of performing spectral analysis on the raw vibration data of each component to obtain spectral data also includes the following steps: Wavelet analysis and envelope demodulation / resonance demodulation are performed sequentially on the original vibration data of each component to obtain original vibration feature samples; The original vibration feature samples are subjected to spectral analysis to obtain the spectral data.

12. A quantitative diagnostic system for mechanical faults, characterized in that, include: The spectrum analysis module is used to perform spectrum analysis on the raw vibration data of each component to obtain spectrum data; The fault type module is used to determine the mechanical fault type of each component based on the spectrum data and the preset mechanical fault characteristic frequency, wherein the preset mechanical fault characteristic frequency is a preset fault characteristic frequency corresponding to the mechanical fault type of each component. The first fault spectrum module is used to extract the first fault spectrum corresponding to each component from the spectrum data according to a preset rule based on the mechanical fault type of each component. The reconstruction module is used to reconstruct the first fault spectrum to obtain the second fault spectrum; The extraction module is used to perform IFFT calculation on the spectrum of the second fault to extract time-domain data; The processing module is used to process the time-domain data and calculate the time-domain feature values; The fault diagnosis module is used to quantitatively diagnose mechanical faults based on the time-domain feature values.

Images