Method, device, and medium for fault detection of a rotating device

Abstract

The present invention relates to a method, device and medium for fault detection of a rotating device. The method includes: obtaining acceleration waveform data of each detection position of the rotating device from acceleration sensing to generate steady-state vibration data and real-time vibration data; performing signal transformation on the steady-state vibration data and the real-time vibration data to generate time-domain signals; extracting fault features based on the time-domain signals via a mechanism model and a predetermined rule to generate a fault multi-dimensional form, the fault multi-dimensional form including a steady-state fault multi-dimensional form and a real-time fault multi-dimensional form; and predicting the fault development stage of the rotating device based on the steady-state fault multi-dimensional form and the real-time fault multi-dimensional form via a multi-dimensional form recognition model. Thus, the present invention can effectively detect changes in the fault development stage of the rotating device.

Description

Technical Field

[0001] The present invention generally relates to fault detection of rotating equipment, and specifically, to a method, a computing device, and a computer-readable storage medium for fault detection of rotating equipment. Background Art

[0002] Traditional methods for fault detection of rotating equipment usually include: judging the fault by manually judging the morphology of the spectrum. For example, detecting whether there are peaks at specific frequencies in the spectrum, and harmonics of specific frequencies, etc. In the above traditional methods, the influence of historical data is not considered, the features are not quantified, and the changes in the fault development stage and the degree of fault development cannot be identified.

[0003] In summary, the deficiency of the traditional method for fault detection of rotating equipment is that it cannot effectively detect the changes in the fault development stage of the rotating equipment. Summary of the Invention

[0004] In view of the above problems, the present invention provides a method, a computing device, and a computer-readable storage medium for fault detection of rotating equipment, which can effectively detect the changes in the fault development stage of the rotating equipment.

[0005] Furthermore, it can effectively detect the changes in the degree of fault development of the rotating equipment.

[0006] According to a first aspect of the present invention, there is provided a method for fault detection of rotating equipment, the method comprising: acquiring acceleration waveform data of each detection position of the rotating equipment from an acceleration sensor to generate steady-state vibration data and real-time vibration data; performing signal transformation on the steady-state vibration data and the real-time vibration data to generate a time-domain signal; extracting fault features based on the time-domain signal via a mechanism model and a predetermined rule to generate a fault multi-dimensional morphology, the fault multi-dimensional morphology including a steady-state fault multi-dimensional morphology and a real-time fault multi-dimensional morphology; and predicting the fault development stage of the rotating equipment based on the steady-state fault multi-dimensional morphology and the real-time fault multi-dimensional morphology via a multi-dimensional morphology recognition model, the fault development stage being used to indicate which development stage the fault of the rotating equipment is in.

[0007] In some embodiments, generating a multi-dimensional fault form includes: obtaining multi-dimensional fault features at multiple historical time points to generate multiple time fault sequences corresponding to the multiple time points, each time fault sequence including multi-dimensional fault features; calculating the mean and standard deviation of the fault features of each dimension; calculating a threshold for each dimension based on the mean and standard deviation of the fault features of each dimension to generate multiple thresholds corresponding to the fault features of multiple dimensions; and generating a steady-state fault multi-dimensional form based on the multiple thresholds corresponding to the fault features of multiple dimensions.

[0008] In some embodiments, the multi-dimensional form recognition model is constructed based on the similarity of the fault multi-dimensional forms. Predicting the fault development stage of a rotating device includes: constructing a multi-dimensional fault sequence based on the normalized multi-dimensional fault features to obtain a steady-state fault multi-dimensional form sequence and a real-time fault multi-dimensional form sequence; calculating the distance between the real-time fault multi-dimensional form sequence and the steady-state fault multi-dimensional form sequence to obtain a form change degree value of the real-time fault multi-dimensional form sequence relative to the steady-state fault multi-dimensional form sequence; and comparing the calculated form change degree value with a first-stage fault dimension threshold and a second-stage fault dimension threshold to predict the fault development stage of the rotating device based on the comparison result.

[0009] In some embodiments, extracting fault features via a mechanism model and predetermined rules based on a time-domain signal to generate a fault multi-dimensional form includes: extracting fault features based on the time-domain signal via the mechanism model and predetermined rules; quantifying the extracted fault features; and normalizing the quantified fault features.

[0010] In some embodiments, the time-domain signal is a spectrum. Normalizing the quantified fault features includes: calculating the total energy based on a predetermined frequency range of the input spectrum; and dividing the peak energy, bottom bulge energy, sideband energy, and harmonic energy by the total energy respectively to generate normalized peak features, bottom bulge features, sideband features, and harmonic features.

[0011] In some embodiments, the method for fault detection of a rotating device further includes: calculating the fault energy corresponding to each time point among multiple time points based on the spatial overlap area of the fault multi-dimensional forms at multiple historical time points; constructing a time fault energy sequence based on the fault energy corresponding to each time point; calculating the mean and standard deviation of each fault energy in the time fault energy sequence; and determining the energy development stage of the target time point based on the mean and standard deviation of the fault energy.

[0012] In some embodiments, determining the energy development stage at a target time point based on the mean and standard deviation of fault energy includes: calculating an energy development stage threshold based on the mean and standard deviation of fault energy; comparing the fault energy at the target time point with the energy development stage threshold; and determining the energy development stage at the target time point based on the comparison result.

[0013] In some embodiments, the method for fault detection of a rotating device further includes: calculating the slope of a feature prediction model of a predetermined dimension corresponding to different time points, so as to construct a slope sequence based on the slopes corresponding to different time points; calculating the mean slope and slope standard deviation of each slope in the slope sequence, so as to calculate a slope development stage threshold based on the calculated mean slope and slope standard deviation; and determining the development stage of the fault feature of the predetermined dimension based on the comparison result between the slope of the feature prediction model at the target time point and the slope development stage threshold.

[0014] In some embodiments, the method for fault detection of a rotating device further includes: constructing a real-time fault development degree matrix and a future fault development degree matrix based on the fault development stage of the rotating device at the target time point and the energy development stage, or the fault development stage and the development stage of the fault feature of the predetermined dimension; and determining the corresponding maintenance status of the rotating device based on the corresponding relationship between the fault development stage and the energy development stage, or the corresponding relationship between the fault development stage and the development stage of the fault feature of the predetermined dimension, so as to fill the real-time fault development degree matrix and the future fault development degree matrix, and the corresponding maintenance status includes attention, maintenance, overhaul, or major overhaul.

[0015] According to a second aspect of the present invention, there is provided a computing device, which includes: at least one processing unit; at least one memory, the at least one memory being coupled to the at least one processing unit and storing instructions for execution by the at least one processing unit, and when the instructions are executed by the at least one processing unit, the device is caused to execute the steps of the method according to the first aspect.

[0016] According to a third aspect of the present invention, there is provided a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a machine, it implements the method according to the first aspect.

[0017] The summary of the invention is provided to introduce a selection of concepts in a simplified form, which will be further described in the detailed implementation below. The summary of the invention is not intended to identify the key features or main features of the present invention, nor is it intended to limit the scope of the present invention. Brief Description of the Drawings

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on the provided drawings.

[0019] Figure 1 The schematic diagram shows a system for implementing a method for fault detection of a rotating device according to an embodiment of the present invention.

[0020] Figure 2 The flowchart shows a method for fault detection of a rotating device according to an embodiment of the present invention.

[0021] Figure 3 The flowchart illustrates a method for generating a multi-dimensional fault morphology according to an embodiment of the present invention.

[0022] Figure 4 The steady-state multi-dimensional fault morphology according to an embodiment of the present invention is shown.

[0023] Figure 5 The flowchart illustrates a method for generating a spatial overlap area prediction model for multi-dimensional fault morphology according to an embodiment of the present invention.

[0024] Figure 6 The schematic diagram illustrates a method for generating the spatial overlap area of multi-dimensional fault morphology according to an embodiment of the present invention.

[0025] Figure 7 The flowchart illustrates a method for determining the energy development stage of a target time point according to an embodiment of the present invention.

[0026] Figure 8 The flowchart illustrates a method for identifying the degree of development of fault characteristics according to an embodiment of the present invention.

[0027] Figure 9 The flowchart illustrates a method for determining the corresponding maintenance status of a rotating device according to an embodiment of the present invention.

[0028] Figure 10 The multi-dimensional fault morphology change diagram according to an embodiment of the present invention is shown.

[0029] Figure 11 The block diagram schematically shows an electronic device suitable for implementing the embodiments of the present invention. Detailed implementation manners

[0030] Preferred embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the preferred embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided so that the present disclosure will be more thorough and complete, and will fully convey the scope of the present disclosure to those skilled in the art.

[0031] As used herein, the term "comprising" and variations thereof mean open-ended inclusion, that is, "including but not limited to". Unless otherwise specified, the term "or" means "and / or". The term "based on" means "at least partially based on". The terms "an exemplary embodiment" and "an embodiment" mean "at least one exemplary embodiment". The term "another embodiment" means "at least one additional embodiment". The terms "first", "second", etc. may refer to different or the same objects.

[0032] As described above, the disadvantage of traditional methods for fault detection of rotating equipment is that they cannot effectively detect changes in the fault development stage of rotating equipment.

[0033] To at least partially solve one or more of the above problems and other potential problems, the present invention proposes a method for fault detection of rotating equipment. In the solution of the present invention, by acquiring acceleration waveform data of each detection position of the rotating equipment from acceleration sensing to generate steady-state vibration data and real-time vibration data; performing signal transformation on the steady-state vibration data and real-time vibration data to generate time-domain signals; based on the time-domain signals, extracting fault features via a mechanism model and predetermined rules to generate a fault multi-dimensional form, the present invention can realize the use of the features of faults in the mechanism model and introduce multi-dimensional forms to characterize the faults, achieving multi-dimensional expression of the faults and significantly enhancing the applicability of fault characterization. In addition, by predicting the fault development stage of the rotating equipment via a multi-dimensional form recognition model based on the steady-state fault multi-dimensional form and the real-time fault multi-dimensional form, the present invention can effectively characterize the fault development stage through changes in the fault multi-dimensional form. Therefore, the present invention can effectively detect changes in the fault development stage of rotating equipment.

[0034] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments, and should not be construed as limiting the scope of the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

[0035] Figure 1FIG. 1 shows a schematic diagram of a system 100 for implementing a method for fault detection of a rotating device according to an embodiment of the present invention. As Figure 1 shown, the system 100 includes: a computing device 110, a rotating device 130, and an acceleration sensor 140. In some embodiments, the computing device 110, the rotating device 130, and the acceleration sensor directly or via a network perform data interaction.

[0036] Regarding the rotating device 130, it is, for example but not limited to, a pump (e.g., a variable-frequency drive pump or a power-frequency drive pump to be detected), a motor, a bearing, or a generator, etc.

[0037] Regarding the acceleration sensor 140 (e.g., a triaxial acceleration sensor or a uniaxial sensor), it collects acceleration waveform data at various detection positions of the rotating device. Taking a pump as an example, the various detection positions of the rotating device include but are not limited to: the non-drive end detection position of the pump, the drive end detection position of the pump, the drive end detection position of the motor, and the non-drive end detection position of the motor. The acceleration waveform data for each detection position includes, for example: the acceleration waveform data in the vertical direction of the bearing, the acceleration waveform data in the horizontal direction of the bearing, and the acceleration waveform data in the axial direction of the bearing.

[0038] Regarding the computing device 110, it is used for fault detection of the rotating device. Specifically, the computing device 110 is used to obtain acceleration waveform data from the acceleration sensor at various detection positions of the rotating device to generate steady-state vibration data and real-time vibration data; and perform signal transformation on the steady-state vibration data and real-time vibration data to generate a time-domain signal. The computing device 110 is also used to extract fault features based on the time-domain signal via a mechanism model and a predetermined rule to generate a fault multi-dimensional form, and the fault multi-dimensional form includes a steady-state fault multi-dimensional form and a real-time fault multi-dimensional form; and based on the steady-state fault multi-dimensional form and the real-time fault multi-dimensional form, via a multi-dimensional form recognition model, predict the fault development stage of the rotating device.

[0039] In some embodiments, the computing device 110 may have one or more processing units, including dedicated processing units such as GPUs, FPGAs, and ASICs, and general-purpose processing units such as CPUs. In addition, one or more virtual machines may also be running on each computing device. The computing device 110 includes, for example: a steady-state vibration data and real-time vibration data generation unit 112, a time-domain signal generation unit 114, a fault multi-dimensional form generation unit 116, and a fault development stage prediction unit 118 of the rotating device. The above steady-state vibration data and real-time vibration data generation unit 112, time-domain signal generation unit 114, fault multi-dimensional form generation unit 116, and fault development stage prediction unit 118 of the rotating device may be configured on one or more computing devices 110.

[0040] Regarding the steady-state vibration data and real-time vibration data generation unit 112, it is used to obtain the acceleration waveform data of each detection position of the rotating device from acceleration sensing, so as to generate steady-state vibration data and real-time vibration data.

[0041] Regarding the time-domain signal generation unit 114, it is used to perform signal transformation on the steady-state vibration data and real-time vibration data, so as to generate time-domain signals.

[0042] Regarding the fault multi-dimensional form generation unit 116, it is used to extract fault features based on the time-domain signals through a mechanism model and predetermined rules, so as to generate fault multi-dimensional forms, where the fault multi-dimensional forms include steady-state fault multi-dimensional forms and real-time fault multi-dimensional forms.

[0043] Regarding the fault development stage prediction unit 118 of the rotating device, it is used to predict the fault development stage of the rotating device based on the steady-state fault multi-dimensional form and the real-time fault multi-dimensional form through a multi-dimensional form recognition model, and the fault development stage is used to indicate which development stage the fault of the rotating device is in.

[0044] Figure 2 The flowchart of the method 200 for fault detection of a rotating device according to an embodiment of the present invention is shown. It should be understood that the method 200 can be executed, for example, at Figure 11 the described electronic device 1100. It can also be executed at Figure 1 the described computing device 110. It should be understood that the method 200 may further include additional actions not shown and / or may omit the shown actions, and the scope of the present invention is not limited in this regard.

[0045] At step 202, the computing device 110 obtains the acceleration waveform data of each detection position of the rotating device from acceleration sensing, so as to generate steady-state vibration data and real-time vibration data.

[0046] Regarding the sampling period of the acceleration waveform data, for example but not limited to, it is not less than 30 minutes each time.

[0047] Regarding the steady-state vibration data, for example, it is obtained based on the acceleration waveform data cached for a period of time. Regarding the period of caching the acceleration waveform data, for example but not limited to, it is one week, greater than or equal to two weeks.

[0048] Regarding the real-time vibration data, for example, it is obtained based on the current acceleration waveform data.

[0049] At step 204, the computing device 110 performs signal transformation on the steady-state vibration data and real-time vibration data, so as to generate time-domain signals.

[0050] Regarding the time-domain signal, for example, it is a spectrum, such as a velocity spectrum or an acceleration spectrum.

[0051] Regarding the method for generating a time-domain signal, for example, it includes: the computing device 110 performs a Fourier transform on the steady-state vibration data to be processed; processes the result of the Fourier transform; and performs an inverse Fourier transform on the processing result to obtain the integrated or differentiated time-domain signal.

[0052] The following formulas (1) to (4) schematically show the algorithms for signal transformation. Among them, formula (1) schematically shows the calculation method of the acceleration signal at any frequency.

[0053] (1)

[0054] In the above formula (1), represents the Fourier component of the acceleration signal at frequency ; A represents the corresponding coefficient. j represents the imaginary number. represents the frequency. t represents the time.

[0055] The following formula (2) schematically shows the velocity signal component obtained by performing a time-domain integration on the acceleration signal component when the initial velocity component is 0.

[0056] (2)

[0057] In the above formula (2), represents the velocity signal component. V represents the corresponding coefficient. represents the acceleration signal component.

[0058] The following formula (3) schematically shows the displacement signal component obtained by performing two-time time-domain integrations on the acceleration signal component when both the initial velocity and initial displacement components are 0.

[0059] (3)

[0060] In the above formula (3), represents the Fourier component of the displacement signal at frequency ; X represents the corresponding coefficient. represents the acceleration signal component.

[0061] The signal transformation of the present invention for the spectrum mainly uses the fast Fourier transform, which is a fast algorithm for the discrete Fourier transform. The following formula (4) schematically shows the algorithm for the discrete Fourier transform.

[0062] (4)

[0063] In the above formula (4), represents the amplitude and phase information of the sine and cosine components with a frequency of , represents the nth point in the discrete time series. N represents the length of the series.

[0064] At step 206, the computing device 110 extracts fault features based on the time-domain signal via a mechanism model and a predetermined rule to generate a fault multi-dimensional form, which includes a steady-state fault multi-dimensional form and a real-time fault multi-dimensional form. The real-time fault multi-dimensional form is, for example, a fault multi-dimensional form calculated based on current vibration data.

[0065] Regarding the faults of rotating equipment, for example, they include: parallel misalignment, angular misalignment, cavitation, outer race fault, inner race fault, etc. It should be understood that for each fault of rotating equipment, it involves multi-dimensional features, for example. Table 1 below exemplarily shows the multi-dimensional features involved in different faults of rotating equipment.

[0066] Table 1

[0067]

[0068] It should be understood that among the different types of features extracted, some have a dimension of "phase", some have a dimension of "angle" or "amplitude", etc. The above different categories of dimensions need to be normalized in order to construct a fault multi-dimensional form based on the normalized fault features.

[0069] Regarding the method for generating a fault multi-dimensional form, for example, it includes: extracting fault features based on the time-domain signal via a mechanism model and a predetermined rule; quantifying the extracted fault features; normalizing the quantified fault features.

[0070] It should be understood that the main types of faults of rotating equipment include: peak recognition, bottom bulge recognition, sideband recognition, harmonic recognition. For example, the morphological similarity method is used to recognize the protruding morphology of frequency, the bulging morphology of bottom noise, and the morphology of sidebands. However, the energy of the protrusion and the energy of the bottom noise are generally difficult to quantify. Therefore, the present invention can adopt the following method to quantify the extracted faults. Regarding the calculation method of total energy, for example, it includes: the computing device 110 calculates the total energy based on a predetermined frequency range of the input spectrum (for example, but not limited to 10~1000HZ). The following formula (5) shows the algorithm for calculating the total energy.

[0071] (5)

[0072] In the above formula (5), Represents the total calculated energy. Represents the amplitude amp corresponding to the frequency freq. freq represents the frequency. amp represents the amplitude.

[0073] Regarding the calculation method of the peak energy, for example, it includes: the calculation device 110 obtains the input spectrum and a predetermined frequency; determines whether there is a peak in the input spectrum at the predetermined frequency.

[0074] The following expression (6) illustrates the method of peak identification.

[0075] (6)

[0076] In the above formula (6), Represents the peak energy. Represents the amplitude of the input spectrum at the predetermined frequency. fault_freq represents the predetermined frequency.

[0077] Regarding the calculation method of the bottom bulge energy, for example, it includes: the calculation device 110 obtains the input spectrum and a predetermined frequency; determines whether there is bottom noise at the predetermined frequency.

[0078] The following expression (7) illustrates the calculation method of the bottom bulge energy.

[0079] (7)

[0080] In the above formula (7), Represents the bottom bulge energy. Represents the frequency Of the amplitude of the input spectrum at. fault_freq represents the predetermined frequency. And The default range of the bottom noise is plus or minus 3HZ.

[0081] Regarding the calculation method of the sideband energy, for example, it includes: the calculation device 110 obtains the input spectrum, a predetermined frequency, and a sideband frequency; determines whether there is a sideband at the predetermined frequency. The following expression (8) illustrates the calculation method of the bottom bulge energy.

[0082] (8)

[0083] In the above formula (8), Represents the sideband energy. band_freq represents the sideband frequency. fault_freq represents the predetermined frequency. Represents the amplitude at the frequency after subtracting the sideband frequency from the predetermined frequency. Represents the amplitude at the frequency after adding the sideband frequency to the predetermined frequency.

[0084] Regarding the calculation method of harmonic energy, for example, it includes: the computing device 110 obtains the input spectrum and the predetermined frequency; and determines whether there are harmonics at the predetermined frequency. The following expression (9) illustrates the calculation method of harmonic energy.

[0085] (9)

[0086] In the above formula (9), represents the harmonic energy. fault_freq represents the predetermined frequency. represents the amplitude of the input spectrum at the predetermined frequency. n represents the multiple, for example, 1 times the frequency, 2 times the frequency.

[0087] It should be understood that since the dimensions of each fault feature are different, it may cause the area of the subsequent multi-dimensional fault morphology calculation to be affected by the amplitude change of a single dimension. Therefore, it is necessary to normalize all fault features.

[0088] Regarding the method for normalizing the quantized fault features, for example, it includes: the computing device 110, for example, includes: calculating the total energy based on the predetermined frequency range of the input spectrum; and dividing the peak energy, bottom bulge energy, sideband energy, and harmonic energy by the total energy to generate the normalized peak feature, bottom bulge feature, sideband feature, and harmonic feature.

[0089] The following expressions (10) to (13) illustrate the algorithm for quantifying the extracted fault features.

[0090] (10)

[0091] (11)

[0092] (12)

[0093] (13)

[0094] In the above formulas (10) to (13), represents the peak feature. represents the bottom bulge feature. represents the sideband feature. represents the harmonic feature. represents the total energy. represents the peak energy. represents the bottom bulge energy. represents the sideband energy. Represents harmonic energy. It should be understood that since normalization has been performed on each fault feature during the construction of multi-dimensional features, the present invention can improve the recognition effect of the model in subsequent morphology recognition and avoid the instability of the recognition model caused by changes in the dimension of a single feature.

[0095] Regarding the multi-dimensional morphology of faults, it includes, for example, the multi-dimensional morphology of steady-state faults and the multi-dimensional morphology of real-time faults.

[0096] Regarding the method for generating the multi-dimensional morphology of steady-state faults, it includes, for example: The computing device 110 obtains multi-dimensional fault features at multiple historical time points to generate multiple time fault sequences corresponding to multiple historical time points, and each time fault sequence includes multi-dimensional fault features; calculates the mean and standard deviation of the fault features of each dimension; based on the mean and standard deviation of the fault features of each dimension, calculates the threshold of each dimension to generate multiple thresholds corresponding to multiple feature dimensions; and generates the multi-dimensional morphology of steady-state faults based on the multiple thresholds corresponding to multiple feature dimensions. The following will be combined with Figure 3 The method 300 for generating the multi-dimensional morphology of steady-state faults will be described in detail, and will not be elaborated here.

[0097] At step 208, the computing device 110 predicts the fault development stage of the rotating device via the multi-dimensional morphology recognition model based on the multi-dimensional morphology of steady-state faults and the multi-dimensional morphology of real-time faults.

[0098] It should be understood that the change in the multi-dimensional morphology of faults may indicate the change in the fault development stage. For example, for a certain fault characterized by multi-dimensional features, if the multi-dimensional morphology of faults constructed based on the multi-dimensional features is scaled proportionally, it may indicate that the rotating device is in the same development stage of the fault, only the fault energy has changed. If the multi-dimensional morphology of faults has changed in shape, such as changing from an obtuse triangle to an acute triangle, it at least indicates that the change degree of the fault features in a certain dimension is relatively large, and the fault features with a relatively large change degree may indicate that the rotating device has entered a different development stage of the fault.

[0099] Regarding the multi-dimensional morphology recognition model, it is used to calculate the change magnitude of the real-time multi-dimensional morphology of faults relative to the multi-dimensional morphology of steady-state faults to identify the change degree of the multi-dimensional morphology. The input features of the multi-dimensional morphology recognition model are the real-time multi-dimensional morphology sequence of faults and the multi-dimensional morphology sequence of steady-state faults. The output result of the multi-dimensional morphology recognition model is the fault development stage of the rotating device. Regarding the multi-dimensional morphology recognition model, in some embodiments, it can be constructed based on existing pre-trained models such as ResNet, DenseNet, and Yolo.

[0100] In some other embodiments, the computing device 110 may adopt a method based on the dimensionality of the feature space to calculate the similarity between the multi-dimensional morphology of the real-time fault and the multi-dimensional morphology of the steady-state fault, so as to construct a multi-dimensional morphology recognition model. Thus, the present invention can build a multi-dimensional morphology recognition model based on the similarity of the multi-dimensional morphology of the fault, and use the change of the multi-dimensional morphology to characterize the development stage of the fault.

[0101] For example, the computing device 110 builds a multi-dimensional fault sequence based on the multi-dimensional fault features obtained through normalization, so as to obtain a multi-dimensional morphology sequence of the steady-state fault and a multi-dimensional morphology sequence of the real-time fault; calculates the distance between the multi-dimensional morphology sequence of the real-time fault and the multi-dimensional morphology sequence of the steady-state fault, so as to obtain a value of the degree of morphological change between the multi-dimensional morphology sequence of the real-time fault and the multi-dimensional morphology sequence of the steady-state fault; determines the fault development stage of the rotating device based on the value of the degree of morphological change, the first development stage threshold, and the second development stage threshold.

[0102] The multi-dimensional morphology sequence of the steady-state fault is, for example:

[0103] The multi-dimensional morphology sequence of the real-time fault is, for example:

[0104] The following expression (14) illustrates an algorithm for calculating the value of the degree of morphological change between the multi-dimensional morphology sequence of the real-time fault and the multi-dimensional morphology sequence of the steady-state fault.

[0105] (14)

[0106] In the above formula (14), represents the fault dimension threshold of the first stage. represents the fault dimension threshold of the second stage. represents the fault dimension threshold of the nth stage. represents the real-time fault feature of the fault dimension of the first stage. represents the real-time fault feature of the fault dimension of the second stage. represents the real-time fault feature of the fault dimension of the nth stage.

[0107] Regarding the fault development stage of a rotating device, it is used to indicate which development stage the fault of the rotating device is in. The fault development stage of the rotating device is divided into three different development stages according to the range of the multi-dimensional morphological change degree of the fault. For example, it includes: the first fault development stage, the second fault development stage, and the third fault development stage. The range of the multi-dimensional morphological change degree in the first fault development stage is, for example, greater than 0 and less than or equal to the threshold of the first development stage, that is, (0, Form_stage1). The range of the multi-dimensional morphological change degree in the second development stage is, for example, greater than the threshold of the first development stage (Form_stage1) and less than or equal to the threshold of the second development stage (Form_stage 2), that is, (Form_stage 1, Form_stage2). The range of the multi-dimensional morphological change degree in the third fault development stage is, for example, greater than the threshold of the second development stage (Form_stage2). Accordingly, the present invention can calculate in real time which development stage the current fault of the rotating device is in. The following formulas (15)-(16) illustrate the calculation methods of the threshold of the first development stage and the threshold of the second development stage.

[0108] (15)

[0109] (16)

[0110] In the above formulas (15)-(16), represents the threshold of the first development stage. represents the threshold of the second fault development stage. represents the mean value of the sequence of the morphological change degree values of the multi-dimensional morphology of the fault. represents the standard deviation of the sequence of the morphological change degree values of the multi-dimensional morphology of the fault. It should be understood that the multi-dimensional morphology formed by each historical data can calculate the morphological change degree value of the multi-dimensional morphology of the fault relative to the steady-state multi-dimensional morphology of the fault. If there are n groups of historical data, then there will be n morphological change degree values, and these n morphological change degree values can be used to construct a sequence of morphological change degree values (F1, F2,…,Fn). Among them, Fn represents the nth morphological change degree value. Based on the n morphological change degree values, the mean value and the standard deviation can be calculated.

[0111] In some embodiments, the computing device 110 can construct a time-series multi-dimensional morphology prediction graph based on a plurality of historical time points and a plurality of multi-dimensional morphologies respectively corresponding to the plurality of historical time points; and calculate the fault development stage information at a future time point based on the multi-dimensional morphology of the future time point predicted by the time-series multi-dimensional morphology prediction graph.

[0112] It should be understood that for each fault dimension, a dimension prediction model can be built according to the method described above, and once the dimension prediction model is constructed, it can predict the change of fault characteristics at future time points, and by synthesizing the change of fault characteristics of each fault dimension, the multi-dimensional morphology diagram at future time points can be predicted in advance. As Figure 10 shown, Figure 10 from left to right in turn are: for the first feature, the fault multi-dimensional morphology 1010 at the first time point t1, the fault multi-dimensional morphology 1012 at the second time point t2, the fault multi-dimensional morphology 1014 at the nth time point tn, and the fault multi-dimensional morphology 1016 at the (n + 1)th time point tn+1. Figure 10 Each fault multi-dimensional morphology in it includes, for example, three fault dimensions, and the fault multi-dimensional morphology at a future time point can be obtained through the dimension prediction model.

[0113] In the above solution, by obtaining the acceleration waveform data of each detection position of the rotating equipment from the acceleration sensor to generate steady-state vibration data and real-time vibration data; performing signal transformation on the steady-state vibration data and real-time vibration data to generate time-domain signals; based on the time-domain signals, extracting fault characteristics through the mechanism model and predetermined rules to generate fault multi-dimensional morphologies, the present invention can realize representing faults with multi-dimensional morphologies, expressing faults in multiple dimensions, and significantly enhancing the applicability of fault representation. In addition, by predicting the fault development stage of the rotating equipment based on the steady-state fault multi-dimensional morphology and real-time fault multi-dimensional morphology through the multi-dimensional morphology recognition model, the present invention can effectively characterize the fault development stage through the change of the fault multi-dimensional morphology. Therefore, the present invention can effectively detect the change of the fault development stage of the rotating equipment.

[0114] Figure 3 The figure shows a flowchart of a method 300 for generating a fault multi-dimensional morphology according to an embodiment of the present invention. It should be understood that the method 300 can be executed, for example, at Figure 11 the electronic device 1100 described. It can also be executed at Figure 1 the computing device 110 described. It should be understood that the method 300 may further include additional actions not shown and / or may omit the actions shown, and the scope of the present invention is not limited in this regard.

[0115] At step 302, the computing device 110 obtains multi-dimensional fault characteristics at multiple historical time points to generate multiple time fault sequences corresponding to multiple time points, and each time fault sequence includes multi-dimensional fault characteristics.

[0116] The following expressions (17)-(19) exemplify multiple time fault sequences.

[0117] (17)

[0118] (18) ...

[0119] (19)

[0120] In the above expressions (17)-(19), represents the nth time fault sequence corresponding to the nth time point. represents the fault feature of the mth dimension in the nth time fault sequence corresponding to the nth time point.

[0121] At step 304, the computing device 110 calculates the mean and standard deviation of the fault features of each dimension.

[0122] At step 306, the computing device 110 calculates the threshold of each dimension based on the mean and standard deviation of the fault features of each dimension, so as to generate multiple thresholds corresponding to multiple feature dimensions.

[0123] The computing device 110 calculates the mean and standard deviation of the fault features of each dimension in order to obtain the threshold of the fault features of each dimension. Taking the fault feature of the first dimension as an example, calculate the sequence of the mean u and standard deviation σ. The following expression (20) exemplifies the specific calculation method of the threshold of the fault feature of the first dimension.

[0124] (20)

[0125] In the above expression (20), represents the threshold of the fault feature of the first dimension. represents the mean of the fault feature of the first dimension. represents the standard deviation of the fault feature of the first dimension.

[0126] At step 308, the computing device 110 generates a steady-state fault multi-dimensional form based on the multiple thresholds corresponding to the fault features of multiple dimensions.

[0127] For example. The computing device 110 generates a steady-state fault multi-dimensional form in a polygonal manner based on multiple fault features and multiple thresholds corresponding to the fault features of multiple dimensions. Figure 4 shows the steady-state fault multi-dimensional form according to an embodiment of the present invention. For example, Figure 4 The first steady-state fault multi-dimensional form 410 shown in the left part is a triangle constructed based on 3 dimensions and 3 stage fault dimension thresholds corresponding to the fault features of 3 dimensions (the 3 stage fault dimension thresholds are respectively ), where Represents the first-stage fault dimension threshold 412. Represents the second-stage fault dimension threshold 414. Represents the third-stage fault dimension threshold 416.. Figure 4 The second steady-state fault multi-dimensional form 420 shown in the middle part is a quadrilateral constructed based on 4 dimensions and 4-stage fault dimension thresholds corresponding to the fault characteristics of the 4 dimensions (the 4-stage fault dimension thresholds are respectively ). Among them, Represents the fourth-stage fault dimension threshold 418. Figure 4 The third steady-state fault multi-dimensional form 430 shown in the right part is a pentagon constructed based on 5 dimensions and 5-stage fault dimension thresholds corresponding to the fault characteristics of the 5 dimensions (the 5-stage fault dimension thresholds are respectively ). Among them, Represents the fifth-stage fault dimension threshold 422.

[0128] By adopting the above means, the present invention introduces multi-dimensional forms to characterize the steady-state faults of rotating equipment, which is beneficial to providing a comparison basis for the real-time fault characterization of rotating equipment.

[0129] Figure 5 Illustrates a flowchart of a method 500 for generating a spatial overlap area prediction model of a fault multi-dimensional form according to an embodiment of the present invention. Figure 6 Illustrates a schematic diagram of a method for generating the spatial overlap area of a fault multi-dimensional form according to an embodiment of the present invention. It should be understood that the method 500 can be executed, for example, at Figure 11 the electronic device 1100 described. It can also be executed at Figure 1 the computing device 110 described. It should be understood that the method 500 may further include additional actions not shown and / or may omit the shown actions, and the scope of the present invention is not limited in this regard.

[0130] At step 502, the computing device 110 determines the vertex coordinates of the steady-state fault multi-dimensional form and the vertex coordinates of the real-time fault multi-dimensional form.

[0131] As Figure 6 shown, the vertex coordinates of the steady-state fault multi-dimensional form 610 are, for example: . The steady-state fault multi-dimensional form 610 is, for example, a triangle.

[0132] The vertex coordinates of the real-time fault multi-dimensional form 620 are, for example: . The real-time fault multi-dimensional form 620 is, for example, a triangle.

[0133] At step 504, the computing device 110 determines the intersection coordinates of each side in the steady-state fault multi-dimensional form with the corresponding side in the real-time fault multi-dimensional form.

[0134] For each side in the steady-state fault multi-dimensional form and the real-time fault multi-dimensional form, its straight-line equation can be calculated. For example, the following expressions (21) and (22) illustrate the calculation equation of the straight line from A1 to A2.

[0135] (21)

[0136] (22)

[0137] In the above expressions (21) and (22), , are the vertex coordinates of A1. , are the vertex coordinates of A2. M represents the slope of the straight line from A1 to A2. For example, the computing device 110 calculates the intersection point S2 between the side A1 A2 in the steady-state fault multi-dimensional form and the side B1 B2 in the real-time fault multi-dimensional form. And so on, the intersection points S1, S2, S3, S4, S5 are obtained.

[0138] At step 506, the computing device 110 calculates the spatial overlap area between the steady-state fault multi-dimensional form and the real-time fault multi-dimensional form based on the vertex coordinates and intersection coordinates of the steady-state fault multi-dimensional form and the real-time fault multi-dimensional form. By adopting the above means, the present invention can characterize the development degree of the fault by the spatial overlap area of the fault multi-dimensional form.

[0139] Regarding the spatial overlap area between the steady-state fault multi-dimensional form and the real-time fault multi-dimensional form, for example Figure 6 as indicated by the label 630 in

[0140] It should be understood that the size of the spatial overlap area represents the size of the energy transformation. The larger the spatial overlap area, the greater the energy and the deeper the fault degree. In some embodiments, the Shoelace formula can be used to calculate the polygon area of the steady-state fault multi-dimensional form and the real-time fault multi-dimensional form. The following formula (23) shows the algorithm for calculating the spatial overlap area between the steady-state fault multi-dimensional form and the real-time fault multi-dimensional form.

[0141] (23)

[0142] In the above expression (23), n represents the number of vertices of the fault multi-dimensional form. (xi, yi) represents the coordinates of a certain vertex. represents the spatial overlap area between the steady-state fault multi-dimensional form and the real-time fault multi-dimensional form.

[0143] At step 508, the computing device 110 constructs a time overlap area sequence based on the spatial overlap areas corresponding to each time point.

[0144] The following expression (24) illustrates the time overlap area sequence.

[0145] (24)

[0146] In the above expression (24), represents the corresponding spatial overlap area at time T1. represents the corresponding spatial overlap area at time T2. represents the corresponding spatial overlap area at time Tn.

[0147] At step 510, the computing device 110 constructs a model of time and feature change based on the time overlap area sequence, thereby generating a spatial overlap area prediction model for the multi-dimensional form of the fault.

[0148] By adopting the above means, the present invention can construct a spatial overlap area prediction model for the multi-dimensional form of the fault.

[0149] Figure 7 FIG. illustrates a flowchart of a method 700 for determining the energy development stage of a target time point according to an embodiment of the present invention. It should be understood that the method 700 can be executed, for example, at Figure 11 the electronic device 1100 described. It can also be executed at Figure 1 the computing device 110 described. It should be understood that the method 400 may further include additional actions not shown and / or may omit the actions shown, and the scope of the present invention is not limited in this regard.

[0150] At step 702, the computing device 110 calculates the fault energy corresponding to each time point among a plurality of historical time points of the multi-dimensional form of the fault based on the spatial overlap areas of the multi-dimensional form of the fault at the plurality of historical time points.

[0151] For example, the historical n time points are t1, t2... tn. The fault energies corresponding to the n time points are, for example, E1, E2... En.

[0152] At step 704, the computing device 110 constructs a time fault energy sequence based on the fault energy corresponding to each time point.

[0153] For example, the time fault energy sequence constructed by the computing device 110 is (E1, E2,..., En).

[0154] At step 706, the computing device 110 calculates the mean and standard deviation of each fault energy in the time fault energy sequence.

[0155] For example, the computing device 110 calculates the mean value of each fault energy in the time fault energy sequence (E1, E2, …, En). and the standard deviation .

[0156] At step 708, the computing device 110 determines the energy development stage at the target time point based on the mean value and the standard deviation of the fault energy.

[0157] For example, the computing device 110 calculates the energy development stage threshold based on the mean value and the standard deviation of the fault energy; compares the fault energy at the target time point with the energy development stage threshold; and determines the energy development stage at the target time point based on the comparison result.

[0158] Regarding the energy development stage threshold, it includes, for example, a first energy development stage threshold (Energy_stage1) and a second energy development stage threshold (Energy_stage2).

[0159] Regarding the energy development stage, it includes, for example: a first energy development stage, a second energy development stage, and a third energy development stage.

[0160] The energy change range corresponding to the first energy development stage is (0, Energy_stage1). The energy change range corresponding to the second energy development stage is (Energy_stage1, Energy_stage2). The energy change range corresponding to the third energy development stage is greater than Energy_stage2.

[0161] Regarding the target time point, it is, for example, the current time point or a future time point.

[0162] The following expressions (25) and (26) illustrate the calculation methods of the first energy development stage threshold and the second energy development stage threshold.

[0163] (25)

[0164] (26)

[0165] In the above expressions (25) and (26), Energy_stage1 represents the first energy development stage threshold. Energy_stage2 represents the second energy development stage threshold. represents the mean value of each fault energy in the time fault energy sequence. represents the standard deviation of each fault energy in the time fault energy sequence.

[0166] By adopting the above means, the present invention can effectively identify the fault energy of the rotating device at the target time point and its energy development stage.

[0167] In some embodiments, method 200 further includes method 800 for identifying the degree of development of fault characteristics. Figure 8 The flowchart of method 800 for identifying the degree of development of fault characteristics according to an embodiment of the present invention is illustrated. It should be understood that method 800 can be executed, for example, at Figure 11 the described electronic device 1100. It can also be executed at Figure 1 the described computing device 110. It should be understood that method 800 may further include additional actions not shown and / or may omit the shown actions, and the scope of the present invention is not limited in this regard.

[0168] At step 802, the computing device 110 calculates the slope of the feature prediction model of a predetermined dimension corresponding to different time points, so as to construct a slope sequence based on the slopes corresponding to different time points.

[0169] Regarding the feature prediction model of a predetermined dimension, it is constructed by using the time series data of a predetermined dimension at historical time points. Taking the first feature (i.e., the fault feature of the first dimension) as an example, first, the computing device 110 obtains the time feature sequence: ( ). Wherein, represents the nth time point. represents the first feature corresponding to the nth historical time point . Secondly, the computing device 110 constructs a model of the change of time and features through a linear regression model according to the time feature sequence, so as to obtain the feature prediction model of a predetermined dimension. Since the change of features and time may be a non-linear change, a polynomial is introduced, and the core is to solve the parameters by using the least square method. By adopting the above means, the present invention can build a single-dimensional feature prediction model based on the change of a single predetermined dimension feature to capture the change trend of a specific single-dimensional fault feature.

[0170] It should be understood that as time points pass, the fault features of different dimensions accumulate continuously, and the feature prediction model of a predetermined dimension will also be updated continuously. Therefore, the slopes of the feature prediction models of a predetermined dimension corresponding to different time points can be calculated.

[0171] For example, the slope sequence constructed by the computing device 110 is (SL1, SL2,..SLn), where SLn represents the slope of the feature prediction model of a predetermined dimension corresponding to the nth time point.

[0172] At step 804, the computing device 110 calculates the mean and standard deviation of the slopes in the slope sequence, so as to calculate the slope development stage threshold based on the calculated mean and standard deviation of the slopes.

[0173] At step 806, the computing device 110 determines the development stage of the fault feature of a predetermined dimension based on the comparison result between the slope of the feature prediction model at the target time point and the slope development stage threshold.

[0174] Regarding the slope development stage threshold, for example, it includes: the first slope development stage threshold and the second slope development stage threshold.

[0175] Regarding the development stage of the fault feature of a predetermined dimension, for example, it indicates the slope development stage where the slope of the feature prediction model at the target time point is located. The development stage of the fault feature of a predetermined dimension, for example, includes: the first slope development stage, the second slope development stage, and the third slope development stage. The slope change range corresponding to the first slope development stage is (0, Slope_stage1). The slope change range corresponding to the second slope development stage is (Slope_stage1, Slope_stage2). The slope change range corresponding to the third slope development stage is greater than Slope_stage2.

[0176] The following expressions (27) and (28) illustrate the calculation methods of the first slope development stage threshold and the second slope development stage threshold.

[0177] (27)

[0178] (28)

[0179] In the above expressions (27) and (28), represents the first slope development stage threshold. represents the second slope development stage threshold. represents the mean of the slopes in the slope sequence. represents the standard deviation of the slopes in the slope sequence.

[0180] By adopting the above means, the present invention can predict any dimension feature in the multi-dimensional form of the fault, so as to capture the change trend of the fault feature of a specific dimension.

[0181] Figure 9 The flowchart of a method 900 for determining the corresponding maintenance state of a rotating device according to an embodiment of the present invention is illustrated. It should be understood that the method 900 can be executed, for example, at Figure 8 the described electronic device 900. It can also be at Figure 1Execute at the described computing device 110. It should be understood that method 900 may also include additional actions not shown and / or the actions shown may be omitted, and the scope of the present invention is not limited in this regard.

[0182] At step 902, the computing device 110 constructs a real-time fault development degree matrix and a future fault development degree matrix based on the fault development stage and energy development stage of the rotating device at the target time point, or the fault development stage and the development stage of the fault characteristics of a predetermined dimension.

[0183] At step 904, the computing device 110 determines the corresponding maintenance status regarding the rotating device based on the correspondence between the fault development stage and the energy development stage, or the correspondence between the fault development stage and the development stage of the fault characteristics of a predetermined dimension, so as to fill the real-time fault development degree matrix and the future fault development degree matrix. The corresponding maintenance status includes attention, maintenance, overhaul, or major overhaul.

[0184] For example, Table 2 below shows a real-time fault development degree matrix constructed based on the fault development stage and the energy development stage. For example, in the real-time fault development degree matrix, if the rotating device at the current time point is in the first fault development stage and the first energy development stage, it indicates that the rotating device has a potential fault but does not affect the overall performance of the device. Therefore, the determined corresponding maintenance status is "attention".

[0185] Table 2

[0186]

[0187] For example, Table 3 below shows a future fault development degree matrix constructed based on the fault development stage and the energy development stage. For example, in the future fault development degree matrix, if the rotating device at a future time point is in the first fault development stage and the third energy development stage, it indicates that the rotating device may have a fault at the future time point. Therefore, the determined corresponding maintenance status is "maintenance".

[0188] Table 3

[0189]

[0190] In addition, it should be understood that the development stage of the fault characteristics of a predetermined dimension can not only indicate which feature is currently the core bringing about the overall development stage or the change in fault energy, but also, since the same fault characteristic in the mechanism model may also be an influencing factor for other faults, the development stage of the fault characteristics of a predetermined dimension can also be used as one of the triggering conditions for evaluating the development of other faults.

[0191] By adopting the above means, the present invention can effectively determine the corresponding maintenance status of the rotating device.

[0192] Figure 11 FIG. schematically shows a block diagram of an electronic device 1100 suitable for implementing the embodiments of the present invention. The electronic device 1100 can be used to implement the execution of Figure 2 , 3 , 5, 7, 8, 9 shown in the methods 200, 300, 500, 700, 800, 900. As Figure 11 shown, the electronic device 1100 includes a central processing unit (i.e., CPU 1101), which can execute various appropriate actions and processes according to computer program instructions stored in a read-only memory (i.e., ROM 1102) or computer program instructions loaded from a storage unit 1108 into a random access memory (i.e., RAM 1103). In the RAM 1103, various programs and data required for the operation of the electronic device 1100 can also be stored. The CPU 1101, ROM 1102, and RAM 1103 are connected to each other via a bus 1104. An input / output interface (i.e., I / O interface 1105) is also connected to the bus 1104.

[0193] Multiple components in the electronic device 1100 are connected to the I / O interface 1105, including: an input unit 1106, an output unit 1107, a storage unit 1108, and the CPU 1101 executes the various methods and processes described above, such as executing the methods 200, 300, 500, 700, 800, 900. For example, in some embodiments, the methods 200, 300, 500, 700, 800, 900 can be implemented as computer software programs, which are stored in a machine-readable medium, such as the storage unit 1108. In some embodiments, part or all of the computer program can be loaded and / or installed onto the electronic device 1100 via the ROM 1102 and / or the communication unit 1109. When the computer program is loaded into the RAM 1103 and executed by the CPU 1101, one or more operations of the methods 200, 300, 500, 700, 800, 900 described above can be executed. Alternatively, in other embodiments, the CPU 1101 can be configured to execute one or more actions of the methods 200, 300, 500, 700, 800, 900 in any other suitable manner (e.g., by means of firmware).

[0194] It should be further noted that the present invention can be a method, a device, a system, and / or a computer program product. The computer program product can include a computer-readable storage medium having computer-readable program instructions for performing various aspects of the present invention loaded thereon.

[0195] A computer-readable storage medium can be a tangible device that can hold and store instructions for use by an instruction execution device. A computer-readable storage medium may be, for example, but is not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of the computer-readable storage medium include: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a static random access memory (SRAM), a portable compact disc read-only memory (CD-ROM), a digital versatile disc (DVD), a memory stick, a floppy disk, a mechanically encoded device such as a punched card or raised structures in grooves having instructions stored thereon, and any suitable combination of the foregoing. The computer-readable storage medium as used herein is not construed to be a transient signal per se, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagating through a waveguide or other transmission medium (e.g., an optical pulse through an optical fiber cable), or an electrical signal transmitted through a wire.

[0196] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to various computing / processing devices, or downloaded to an external computer or external storage device through a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network may include a copper transmission cable, an optical fiber transmission, a wireless transmission, a router, a firewall, a switch, a gateway computer, and / or an edge server. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in a computer-readable storage medium in each computing / processing device.

[0197] The computer program instructions for performing the operations of the present invention may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine - related instructions, microcode, firmware instructions, state - setting data, or source code or object code written in any combination of one or more programming languages, including object - oriented programming languages such as Smalltalk, C++, etc., and conventional procedural programming languages such as the "C" language or similar programming languages. The computer - readable program instructions may be executed entirely on the user's computer, partially on the user's computer, executed as a stand - alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider). In some embodiments, by using the state information of the computer - readable program instructions to customize an electronic circuit, such as a programmable logic circuit, a field - programmable gate array (FPGA), or a programmable logic array (PLA), the electronic circuit can execute the computer - readable program instructions to implement various aspects of the present invention.

[0198] Aspects of the present invention are described herein with reference to the flowchart and / or block diagram of methods, apparatus (systems), and computer program products according to embodiments of the invention. It should be understood that each block of the flowchart and / or block diagram, and the combinations of blocks in the flowchart and / or block diagram, can be implemented by computer - readable program instructions.

[0199] These computer - readable program instructions can be provided to the processing unit of a processor in a voice interaction device, a general - purpose computer, a special - purpose computer, or other programmable data - processing device, thereby producing a machine such that when these instructions are executed by the processing unit of the computer or other programmable data - processing device, a means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram is produced. These computer - readable program instructions can also be stored in a computer - readable storage medium, and these instructions cause the computer, programmable data - processing device, and / or other devices to work in a specific manner. Thus, the computer - readable medium storing the instructions includes a manufacture, which includes instructions for implementing various aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0200] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device, causing a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process such that the instructions executed on the computer, other programmable data processing apparatus, or other device implement the functions / acts specified in one or more boxes of the flowchart and / or block diagram.

[0201] The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of devices, methods, and computer program products according to various embodiments of the present invention. In this regard, each box in the flowchart or block diagram may represent a module, a segment of a program, or a portion of an instruction, which contains one or more executable instructions for implementing the specified logical function. In some alternative implementations, the functions noted in the boxes may occur in a different order than noted in the figures. For example, two consecutive boxes may actually be executed substantially in parallel, or they may sometimes be executed in the reverse order, depending on the functionality involved. It should also be noted that each box in the block diagrams and / or flowcharts, and combinations of boxes in the block diagrams and / or flowcharts, can be implemented by a dedicated hardware-based system that performs the specified functions or acts, or by a combination of dedicated hardware and computer instructions.

[0202] The embodiments of the present invention have been described above. The above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The choice of terms used herein is intended to best explain the principles of the embodiments, the practical application, or improvements made to the technology in the market, or to enable other ordinary skilled persons in the art to understand the embodiments disclosed herein.

[0203] The above are only optional embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for fault detection of rotating equipment, characterized in that: include: Acquire acceleration waveform data about each detection position of the rotating device from the acceleration sensor to generate steady-state vibration data and real-time vibration data; Perform signal transformation on steady-state vibration data and real-time vibration data to generate time domain signals; Based on the time domain signal, the fault feature is extracted via the mechanism model and the predetermined rules to generate a polygonal multi-dimensional fault form, wherein each vertex coordinate of the polygonal multi-dimensional fault form corresponds to a fault feature of each dimension, and includes a polygonal steady-state multi-dimensional fault form and a real-time multi-dimensional fault form; as well as Based on the multi-dimensional morphology of steady-state faults and the multi-dimensional morphology of real-time faults, the fault development stage of the rotating equipment is predicted via a multi-dimensional morphology recognition model. The fault development stage is used to indicate which development stage the fault of the rotating equipment is in. The multi-dimensional morphology recognition model is constructed based on the spatial overlapping area of ​​the real-time fault multi-dimensional morphology and the steady-state fault multi-dimensional morphology.

2. The method according to claim 1, characterized in that Generate multi-dimensional fault patterns including: Acquire multidimensional fault features at multiple historical time points to generate multiple time fault sequences corresponding to the multiple time points, each time fault sequence including the multidimensional fault features; Calculate the mean and standard deviation of the fault characteristics of each dimension; Calculating a threshold value of each dimension based on a mean value and a standard deviation of the fault signature of each dimension, so as to generate a plurality of threshold values ​​corresponding to the fault signatures of the plurality of dimensions; and Based on multiple thresholds corresponding to the fault features in multiple dimensions, a steady-state fault multi-dimensional morphology is generated.

3. The method according to claim 1, characterized in that Predicting the development stages of failures in rotating equipment includes: A multidimensional fault sequence is constructed based on the normalized multidimensional fault features, thereby obtaining a steady-state fault multidimensional morphology sequence and a real-time fault multidimensional morphology sequence; Calculating the distance between the real-time fault multidimensional morphology sequence and the steady-state fault multidimensional morphology sequence, thereby obtaining a morphology change degree value of the real-time fault multidimensional morphology sequence relative to the steady-state fault multidimensional morphology sequence; and The calculated morphology change degree value is compared with a first development stage threshold value and a second development stage threshold value to predict a fault development stage with respect to the rotating equipment based on the comparison result.

4. The method according to claim 1, characterized in that: Based on the time domain signal, the fault features are extracted through the mechanism model and predetermined rules to generate the multi-dimensional fault morphology including: Based on the time domain signal, the fault characteristics are extracted through the mechanism model and the predetermined rules; quantifying the extracted fault features; and Normalization is performed on the quantified fault features.

5. The method according to claim 4, characterized in that The time domain signal is a frequency spectrum, and normalization is performed on the quantified fault feature including: Calculating total energy based on a predetermined frequency range of the input spectrum; and The peak energy, bottom ridge energy, sideband energy, and harmonic energy are respectively divided by the total energy to generate normalized peak characteristics, bottom ridge characteristics, sideband characteristics, and harmonic characteristics.

6. The method according to claim 1, characterized in that The method further comprises: Based on the spatial overlapping area of ​​the multi-dimensional fault morphology at multiple historical time points, the fault energy corresponding to each time point in the multiple time points is calculated; Based on the fault energy corresponding to each time point, a time fault energy sequence is constructed; calculating the mean and standard deviation of each fault energy in the temporal fault energy series; and Based on the mean and standard deviation of fault energy, the energy development stage at the target time point is determined.

7. The method according to claim 6, characterized in that Based on the mean and standard deviation of fault energy, the energy development stages at the target time point are determined as follows: Calculate the energy development stage threshold based on the mean and standard deviation of fault energy; comparing the fault energy at the target time point with the energy development stage threshold; and Based on the comparison results, the energy development stage at the target time point is determined.

8. The method according to claim 3, characterized in that The method further comprises: Calculating the slope of the feature prediction model of the predetermined dimension corresponding to different time points, so as to construct a slope sequence based on the slopes corresponding to the different time points; calculating a slope mean and a slope standard deviation of each slope in the slope sequence, so as to calculate a slope development stage threshold based on the calculated slope mean and slope standard deviation; and Based on the comparison result of the slope of the feature prediction model at the target time point and the slope development stage threshold, the development stage of the fault feature of the predetermined dimension is determined.

9. The method according to claim 3, characterized in that: The method further comprises: Based on the fault development stage and energy development stage of the rotating equipment at the target time point, or the fault development stage and the development stage of the fault characteristics of the predetermined dimension, a real-time fault development degree matrix and a future fault development degree matrix are constructed; Based on the correspondence between the fault development stage and the energy development stage, or the correspondence between the fault development stage and the development stage of the fault characteristics of a predetermined dimension, the corresponding maintenance status of the rotating equipment is determined to fill the real-time fault development degree matrix and the future fault development degree matrix. The corresponding maintenance status includes attention, maintenance, inspection, or overhaul.

10. The method according to claim 3, characterized in that: The method further comprises: Determine vertex coordinates of a steady-state fault multi-dimensional morphology and vertex coordinates of a real-time fault multi-dimensional morphology; Determine the coordinates of the intersection of each edge in the steady-state fault multidimensional form and the corresponding edge in the real-time fault multidimensional form; Based on the vertex coordinates and intersection coordinates of the steady-state fault multidimensional form and the real-time fault multidimensional form, the spatial overlapping area of ​​the steady-state fault multidimensional form and the real-time fault multidimensional form is calculated; Based on the spatial overlapping area corresponding to each time point, construct a temporal overlapping area sequence; and Based on the time overlapping area sequence, a model of time and feature changes is constructed to generate a spatial overlapping area prediction model for multi-dimensional fault morphology.

11. A computing device comprising: at least one processing unit; At least one memory, the at least one memory being coupled to the at least one processing unit and storing instructions for execution by the at least one processing unit, the instructions, when executed by the at least one processing unit, causing the computing device to perform the steps of the method according to any one of claims 1 to 10.

12. A computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the method according to any one of claims 1 to 10 when executed by a machine.

Images