Equipment fault detection method based on baseline data space

Abstract

The present invention provides a method for detecting equipment faults based on a baseline data space, comprising: determining an operating condition parameter having a predetermined correlation with the operating condition of the equipment; discretizing the original data set of the operating condition parameters of the equipment according to a fixed-length time period of the operating condition parameters of the equipment, and generating a discrete operating condition parameter set containing data of multiple fixed-length time periods; constructing a baseline frequency domain data space of the equipment in a normal state according to the discrete operating condition parameter set and the acquired signal data of the equipment in a normal state, and setting a fault judgment threshold; and calculating an over-limit difference evaluation index, comparing the over-limit difference evaluation index with the fault judgment threshold, and judging whether the equipment has a fault. In the present application, by disassembling and discretizing complex operating conditions, the influence of multiple operating conditions on fault feature extraction, fault identification, etc. is reduced, and the fault is intuitively responded to through the over-limit difference space, making it easier for relevant business personnel to understand and grasp the fault information and the degree of fault degradation.

Description

Technical Field

[0001] The present application relates to the field of mechanical equipment fault detection, and more specifically, to an equipment fault detection method based on baseline data space. Background Art

[0002] Currently, research on mechanical equipment fault detection is extensive, with universities, research institutions, and major industrial entities showing strong interest in this area, resulting in a proliferation of related studies. However, the practical application of intelligent methods in industry is relatively limited and limited, indicating that intelligent equipment fault detection is still in the theoretical research stage and still a long way from widespread industrial implementation. Intelligent equipment fault detection and prediction are currently hampered by numerous factors, including a lack of effective data from actual production, a lack of accumulated fault data, complex industrial production site environments with overlapping multi-source signals, and the poor interpretability of widely used methods such as neural networks, which make the results difficult for field engineers to understand and accept.

[0003] The development of intelligent equipment fault detection is currently limited by numerous factors. For example, complex on-site working conditions make fault feature extraction difficult; the poor quality of effective data accumulation affects model prediction results; and the poor interpretability and low accuracy of model detection prediction results make it difficult for relevant business engineers to directly understand the cause and basis of the alarm. Summary of the Invention

[0004] In order to solve the above problems, the present invention provides an equipment fault detection method based on baseline data space. The detection method can disassemble and discretize complex working conditions, reduce the impact of multiple working conditions on fault feature extraction, fault identification, etc., and intuitively respond to faults through the out-of-limit difference space, making it easier for relevant business personnel to understand and grasp fault information and the degree of fault degradation.

[0005] In order to achieve the above-mentioned purpose, the equipment fault detection method based on baseline data space according to the present application includes: determining the operating condition parameters that have a predetermined correlation with the operating condition of the equipment; discretizing the original data set of the operating condition parameters of the equipment according to the fixed-length time period of the operating condition parameters of the equipment, and generating a discrete operating condition parameter set containing multiple fixed-length time period data; constructing the baseline frequency domain data space of the equipment in the normal state according to the discrete operating condition parameter set and the signal data of the equipment in the normal state, and setting a fault judgment threshold; and calculating the out-of-limit difference evaluation index, and comparing the out-of-limit difference evaluation index with the fault judgment threshold to determine whether the equipment has a fault.

[0006] Furthermore, the original data set is discretized according to the fixed-length time period of the operating parameters of the equipment to generate a discrete operating parameter set containing data of multiple fixed-length time periods, including: based on the original data set, lagging one time point to obtain a new data set of the lagging time point, wherein the new data set is the same length as the original data set; calculating the difference sequence between the new data set and the original data set to generate a difference sequence set, setting a fault tolerance threshold σ, and setting the data value of the difference sequence <σ to zero; and setting a time length threshold t_lim, truncating the data of the difference sequence that is continuously zero for a time length > t_lim to generate the discrete operating parameter set.

[0007] Furthermore, based on the discrete operating condition parameter set and the acquired signal data, a baseline frequency domain data space of the equipment in a normal state is constructed, including: obtaining a baseline time domain data set and a baseline frequency domain data set based on the discrete operating condition parameter set; band-pass filtering the baseline frequency domain data set to generate a second baseline frequency domain data set corresponding to the operating condition parameters in the frequency band of interest; and constructing the baseline frequency domain data space based on the second baseline frequency domain data set.

[0008] Furthermore, based on the discrete operating condition parameter set, a baseline time domain data set and a baseline frequency domain data set are obtained, including: discretizing and segmenting the signal data corresponding to the time nodes in the signal data under the normal state, and marking the parameter values ​​of the operating condition parameters corresponding to the time period of the segmented data to generate the baseline time domain data set; and performing fast Fourier transform on the baseline time domain data set to obtain the baseline frequency domain data set.

[0009] Furthermore, constructing the baseline frequency domain data space based on the second baseline frequency domain data set includes: expanding each segment of spectral data in the second baseline frequency domain data set one by one according to the frequency dimension to obtain a frequency matrix; calculating statistical index values ​​of the frequency matrix column by column to obtain a baseline frequency domain upper limit space; performing a Hilbert transform on the baseline frequency domain upper limit space and calculating the upper envelope to obtain the baseline frequency domain data space.

[0010] Furthermore, the statistical indicator values ​​include: mean, median, maximum and extreme values.

[0011] Furthermore, the equipment fault detection method further includes: based on the baseline frequency domain data space, according to the parameter value of the target unknown operating parameter of the equipment, using an interpolation algorithm to obtain an interpolation baseline frequency domain data space under the target unknown operating parameter.

[0012] Furthermore, the equipment fault detection method also includes: setting an evaluation index threshold of the interpolation baseline frequency domain data space; calculating the separation degree between the baseline frequency domain data space and the interpolation baseline frequency domain data space; and comparing the separation degree with the evaluation index threshold to determine whether the interpolation baseline frequency domain data space is appropriate.

[0013] Furthermore, the equipment fault detection method also includes: based on the comparison result of the separation degree and the evaluation index threshold, determining that the interpolation baseline frequency domain data space is appropriate, adding the interpolation baseline frequency domain data space to the baseline frequency domain data space and as a part of the baseline frequency domain data space; or based on the comparison result of the separation degree and the evaluation index threshold, determining that the interpolation baseline frequency domain data space is inappropriate, supplementing the signal data under the parameter value adjacent to the parameter value of the target unknown operating condition parameter; and based on the baseline frequency domain data space and the supplemented signal data, using an interpolation algorithm to obtain the interpolation baseline frequency domain data space under the target unknown operating condition parameter.

[0014] Furthermore, the operating condition parameters are parameters that are not highly correlated with the spectral structure distribution. Based on the baseline frequency domain data space, according to the parameter values ​​of the target unknown operating condition parameters of the equipment, an interpolation algorithm is used to obtain the interpolation baseline frequency domain data space under the target unknown operating condition parameters, including: superimposing the baseline frequency domain data spaces in sequence to construct a joint distribution matrix; according to the parameter values ​​of the target unknown operating condition parameters of the equipment, an interpolation algorithm is used to interpolate the joint distribution matrix to obtain the interpolation baseline frequency domain data space under the target unknown operating condition parameters.

[0015] Furthermore, the operating condition parameters are parameters that are highly correlated with the spectral structure distribution. Based on the baseline frequency domain data space, according to the parameter values ​​of the target unknown operating condition parameters of the equipment, an interpolation algorithm is used to obtain an interpolation baseline frequency domain data space under the target unknown operating condition parameters, including: superimposing the baseline frequency domain data spaces in sequence to construct a joint distribution matrix; transforming the frequency values ​​in the joint distribution matrix into order values ​​to obtain a joint distribution matrix under the order dimension; according to the parameter values ​​of the target unknown operating condition parameters of the equipment, an interpolation algorithm is used to interpolate the joint distribution matrix under the order dimension to obtain an interpolation baseline data order space under the target unknown operating condition parameters; inversely transforming the order values ​​of the interpolation baseline data order space under the target unknown operating condition parameters to restore them back to frequency values ​​to obtain an interpolation baseline frequency domain data space under the target unknown operating condition parameters.

[0016] Furthermore, the separation degree is the sum of the absolute values ​​of the amplitude differences between the baseline frequency domain data space and the interpolation baseline frequency domain space.

[0017] Furthermore, the device fault detection method further includes: performing a cleaning process on the acquired original data set to supplement missing values ​​and remove zero values, and using the cleaned data set as the original data set.

[0018] Furthermore, the missing value is supplemented by averaging the nearest neighbor data before and after the position of the missing value.

[0019] Furthermore, the device fault detection method further includes: acquiring signal data of the normal state of the device, wherein the signal data includes vibration signal data.

[0020] Furthermore,

[0021] Out-of-limit difference evaluation index = sum(abs(baseline upper limit data of baseline frequency domain data space - data to be detected)),

[0022] Among them, abs(baseline upper limit data of baseline frequency domain data space - data to be detected) means finding the absolute value of the difference between the baseline upper limit data of baseline frequency domain data space and the data to be detected, and sum(abs(baseline upper limit data of baseline frequency domain data space - data to be detected)) means finding the sum of the absolute values ​​of the difference between the baseline upper limit data of baseline frequency domain data space and the data to be detected.

[0023] According to another aspect of the present application, a computer device is provided, comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and when the processor executes the computer program, the steps of the above-mentioned device fault detection method based on the baseline data space are implemented.

[0024] According to another aspect of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the device fault detection method based on the baseline data space is implemented.

[0025] The equipment fault detection method of this application discretizes, processes, and stores historical operating conditions based on strongly correlated parameters. This helps refine fault detection scenarios, limits the impact of different operating conditions on the effectiveness of fault detection results, and simplifies the conditions and steps for fault detection in each operating condition. Furthermore, by setting an out-of-limit difference evaluation indicator, fault responses can be intuitively displayed, making it easier for relevant business personnel to understand and grasp fault information and the degree of fault degradation.

[0026] Furthermore, according to the equipment fault detection method of the present application, the original data of historical operating parameters are truncated using a difference sequence and other step sizes, which can quickly and accurately obtain discrete operating parameter data and corresponding time periods.

[0027] Moreover, according to the equipment fault detection method of the present application, based on the historical discrete operating parameters, a baseline data vibration original data set (baseline time domain data set), a baseline data vibration frequency domain data set (baseline frequency domain data set), and a baseline data space of the equipment's normal operation under different discrete operating parameters are constructed, providing a reliable evaluation standard and basis method for equipment fault detection.

[0028] In addition, based on the baseline data space under historical finite discrete operating condition parameters, the interpolation fitting of the baseline data space of unknown operating conditions is automatically generated, which enriches and improves the discrete operating conditions and baseline data space, and avoids the limitation of fault detection in the baseline data space caused by the lack of historical operating conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The accompanying drawings, which constitute part of this application, are intended to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are intended to explain the present invention and do not constitute an undue limitation of the present invention. In the accompanying drawings:

[0030] Figure 1 A flowchart of a device fault detection method based on a baseline data space according to an embodiment of the present application is shown;

[0031] Figure 2 A flowchart of a device fault detection method based on baseline data space according to a preferred embodiment of the present application is shown. DETAILED DESCRIPTION

[0032] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0033] The present application provides an equipment fault detection method based on baseline data space, which constructs a baseline data space based on strongly correlated parameter data of the equipment's normal operating conditions and time domain and frequency domain data of vibration signals.

[0034] According to this application, if Figure 1As shown, a device fault detection method based on a baseline data space is provided, the method comprising: determining an operating condition parameter having a predetermined correlation with the operating condition of the device (S101); discretizing an original data set of the acquired operating condition parameters of the device according to a fixed-length time period of the operating condition parameters of the device, and generating a discrete operating condition parameter set containing data of multiple fixed-length time periods (S102); constructing a baseline frequency domain data space of the device in a normal state based on the discrete operating condition parameter set and the acquired signal data of the device in a normal state, and setting a fault judgment threshold (S103); and calculating an out-of-limit difference evaluation index, and comparing the out-of-limit difference evaluation index with the fault judgment threshold to determine whether the device has a fault (S104).

[0035] According to the equipment fault detection method of the present application, by disassembling and discretizing complex working conditions, the impact of multiple working conditions on fault feature extraction, fault identification, etc. is reduced, and the fault is intuitively responded to through the out-of-limit difference space, which makes it easier for relevant business personnel to understand and grasp the fault information and the degree of fault degradation.

[0036] According to this application, if Figure 2 The following description primarily covers discrete operating condition parameter processing and standardized mining (discretizing the operating conditions and constructing a baseline data space based on those conditions), baseline data space construction, automatic interpolation and fitting of baseline data spaces for unknown operating conditions, and validation of baseline data space fault detection effectiveness. Each of these modules includes specific construction steps, and the combination of these modules generates a baseline data space database for each discrete operating condition.

[0037] The baseline data space data under complex discrete working conditions can directly calculate the out-of-limit difference evaluation index for the equipment's operating data under the same working conditions, and based on the level of the out-of-limit difference evaluation index, determine whether the current state of the equipment is in a fault state. According to the present application, the baseline data space, on the one hand, collects basic vibration data information of the equipment's normal operation under various historical discrete working conditions, and on the other hand, can be interpolated and fitted to obtain the corresponding baseline data space for the current new and unknown working conditions. Therefore, the method of the present application is not limited by the number of historical working conditions, and reflects the equipment's fault state through the out-of-limit difference evaluation index, which is relatively intuitive and easier for on-site operators or fault diagnosis engineers to understand the basis and fault state of the equipment failure.

[0038] The following is a detailed description of the device fault detection method based on the baseline data space.

[0039] 1. Discrete working condition parameter processing and standardized mining

[0040] Determine the operating parameters that are highly correlated with the complex operating conditions of the equipment.

[0041] The working conditions are quantified and discretized according to the determined parameters, that is, the numerical values ​​of the equipment operation-related parameters that are strongly related to the working conditions are discretized.

[0042] Taking the fracturing truck-mounted plunger pump at a fracturing well site as an example, parameters that influence the operating fluctuations of the fracturing pump include pressure, speed, sand ratio, etc. Of these parameters, speed is the primary influencing factor. According to this application, a baseline data space is constructed for fracturing pump gearbox fault detection, using the scenario of fracturing pump gearbox fault detection as the basis for gearbox equipment fault detection. The discrete operating condition parameter (i.e., the operating condition parameter to be discretized, hereinafter referred to as the discrete operating condition parameter) is the drive motor or engine speed, that is, the operating condition parameter is determined to be speed.

[0043] Get the original data set of the equipment's operating parameters.

[0044] According to an embodiment of the present application, the raw data set of main relevant operating parameters of the equipment working condition is obtained JDS={(time i ,speed i )}, where i = 0, 1, 2, ..., n. time i The original data time data column of the equipment working condition affecting parameters, speed i This is the original data column of the parameters affecting the equipment working condition. Taking the fracturing pump as an example, the main parameters affecting the equipment working condition are the driving motor or engine speed, speed i is the speed of the fracturing pump motor or engine.

[0045] According to an embodiment of the present application, the original data set JDS is cleaned, that is, (time i ,speed i ) data to supplement missing values ​​and remove zero values. According to one embodiment of the present application, missing values ​​are mainly supplemented by averaging the nearest neighbor data before and after the missing value position, and zero value data are directly removed. Taking the fracturing pump as an example, the speed is mainly i Missing values ​​are added and zero values ​​are removed. The cleaned data set is used as the original data set.

[0046] According to an embodiment of the present application, based on the original data set JDS, time i Column, lag one time point granularity, obtain the new JDS data of the lag time point JDS_new = {(time_new i+1 ,speed_new i+1)} and delete the row containing the last time point data in the original data set JDS, so that JDS and JDS_new are equal in length, and the data in the same row differ by one time point. The new JDS data JDS_new is the data of the original data set JDS that is one time point behind. For example, if the original data set JDS starts at 1 second, the new JDS data JDS_new starts at a time point after the original data set JDS, such as 2 seconds.

[0047] For JDS_new and JDS speed_new i+1 Column and speed i Column calculation difference sequence speed_diff i+1 =speed_new i+1 -speed i , generate difference sequence data set JDS_diff={(time_new i+1 ,speed_diffi i+1 )}, and set the fault tolerance threshold σ (generally <5, can be set according to the specific situation), and set speed_diffi i+1 Data values ​​< σ are set to zero.

[0048] Start speed_diffi from time_new1 of the difference sequence data set JDS_diff i+1 Continuous data traversal, set the discrete working condition parameter stable time length threshold t_lim (generally >10min, can be set according to specific circumstances), intercept speed_diffi i+1 The start time of the truncated data when the continuous zero duration is greater than t_lim is time_s k and end time time_e k And the working condition parameter speed corresponding to the current time period k , generate discrete working condition parameter set JDS resut ={(time_s k ,time_e k ,speed k )}, the original working condition is discretized into a fixed-length time period according to the working condition parameters, where And k is the number of fixed-length time periods, len(JDS_diff) represents the total length of the JDS_diff data set, that is, the total time length of the data set, time_s k Indicates the current speed start time, time_e k , indicates the end time of the current speed, speed k Indicates the current speed. And JDS resutCreate structured database tables and store data according to the table structure and field naming. There is no limit on the database type.

[0049] According to the present application, the original data set is discretized according to the fixed-length time periods of the operating parameters of the equipment to generate a discrete operating parameter set containing data of multiple fixed-length time periods.

[0050] 2. Build a baseline data space

[0051] Acquire signal data of the device in a normal state, where the signal data includes vibration signal data.

[0052] According to one embodiment of the present application, the vibration signal source data V collected by the multi-channel vibration sensor at the target position of the equipment under normal condition is obtained. signal ={AI1,AI2,…,AIi},AIi=(AI time ,AI signal ). Where i=1,2,3,…, AI time Indicates the vibration signal acquisition time of this channel, AI signal Indicates the original collected data of the vibration signal of this channel.

[0053] Taking the fracturing pump reducer as an example, five channels of vibration sensors are installed: on the reducer input side (H), on the opposite side of the reducer input (V), on the opposite axial side of the parallel stage input (H), on the large-end input side (H) of the parallel stage, and on the planetary stage housing (H). The following describes the construction, verification, and interpolation process for the entire baseline data space using a single channel. The calculation process for other channels is the same and can be used as a reference.

[0054] Of course, the existing vibration signal data set of the device in normal state can also be used.

[0055] According to the obtained discrete operating condition parameter set, a baseline time domain data set and a baseline frequency domain data set are obtained.

[0056] According to an embodiment of the present application, according to each time period in the discrete working condition parameter set, it is determined whether there is original vibration signal acquisition data in the time period. If not, the working condition parameter time period is skipped; if so, according to JDS resut The start and end time of each discrete working condition parameter detailed record information data are signal The discrete time domain data set AIi is obtained by discretizing the vibration signal of a channel AIi into discrete time periods corresponding to the time nodes under the discrete working condition parameters, and marking the parameter value of the discrete working condition parameter corresponding to the time period for each data segment. k =(AI time_k ,AI signal_k ,AI speed_k), among which AI time_k Indicates the vibration signal acquisition time, AI signal_k Indicates the baseline frequency domain amplitude, AI speed_k Indicates the parameter value of the discrete working condition parameter corresponding to the baseline time domain data set, and AIi k According to the table structure and field naming, the structured database table is created and the data is stored. The database type is not limited. Taking the fracturing pump as an example, the AIi k The data set is the original data of the vibration signal in the corresponding time period of each discrete constant speed.

[0057] For the vibration data of each discrete working condition parameter time period of a certain channel, fast Fourier transform is performed in sequence according to the fixed number of spectral lines fr and the fixed number of sampling points fd (the specific value can be customized) to obtain the baseline frequency domain data set Fre_AIi k =(AI fre_k ,AI signal_k ,AI speed_k ). Among them, AI fre_k Indicates the baseline frequency domain value, AI signal_k Indicates the baseline frequency domain amplitude, AI speed_k Indicates the parameter value of the discrete operating condition parameter corresponding to the baseline frequency domain data set. k According to the table structure and field naming, create a structured database table and store data. The database type is not limited. Taking the fracturing pump as an example, set the fixed number of spectrum lines fr = 12800, the fixed number of sampling points fd = 51200, and intercept the obtained Fre_AIi k The data set is the baseline frequency domain data set for the time period corresponding to each discrete constant speed.

[0058] Band-pass filtering is performed on the obtained baseline frequency domain data set to generate a second baseline frequency domain data set corresponding to the operating condition parameters of the frequency band of interest.

[0059] According to an embodiment of the present application, the upper and lower frequency limits are set to [fre start ,fre end ] bandpass filter, customize the upper and lower limit data according to the analysis requirements, and perform the baseline frequency domain data set Fre_AIi k Perform bandpass filtering to generate the baseline frequency domain data set Fre′_AIi corresponding to the discrete parameters of the frequency band of interest k , that is, the second baseline frequency domain dataset.

[0060] A baseline frequency domain data space is constructed based on the second baseline frequency domain data set.

[0061] According to an embodiment of the present application, traverse k=1, 2, 3, ..., and convert the second baseline frequency domain data set Fre′_AIik Each segment of spectrum data in the two-dimensional list Fre_AIi′ k =(AI′ fre_k ,AI′ signal_k ) Expand each spectrum one by one according to the frequency dimension to obtain the frequency matrix, as shown below:

[0062]

[0063] Among them, each row in the above matrix is ​​Fre_AIi′ k A complete spectrum in the frequency range is 0~MHz, and the number of rows is the number of sampling time periods of the signal vibration data corresponding to the fixed number of sampling points fd.

[0064] According to one embodiment of the present application, the above matrix is ​​calculated by column by column for a certain statistical index (which may be one of the following indexes but not limited to: mean, median, maximum value, extreme value, etc.), and Fre_AIi′ can be obtained. k The upper limit waveform of the baseline data (the upper limit space of the baseline frequency domain), that is in, Represents the upper frequency domain amplitude of baseline data under different discrete parameters, AI′ fre_k is the upper frequency domain value of the baseline data under different discrete parameters.

[0065] For example, according to one embodiment of the present application, the median is used as a statistical indicator. For Fre_AIi′ k According to another embodiment of the present application, the maximum value is used as the statistical indicator. For Fre_AIi′ k Find the maximum value of a matrix column by column.

[0066] right Baseline upper limit data Perform Hilbert transform and find its upper envelope to obtain the baseline frequency domain data space (i.e., baseline data space) Fre_hil_AIi′ k .

[0067] Set the fault judgment threshold γ, calculate the over-limit difference evaluation index, and compare the over-limit difference evaluation index with the fault judgment threshold to determine whether the equipment has a fault.

[0068] Out-of-limit difference evaluation index = sum(abs(baseline upper limit data of baseline frequency domain data space - data to be detected)),

[0069] Among them, abs(baseline upper limit data of baseline frequency domain data space - data to be detected) means finding the absolute value of the difference between the baseline upper limit data of baseline frequency domain data space and the data to be detected, and sum(abs(baseline upper limit data of baseline frequency domain data space - data to be detected)) means finding the sum of the absolute values ​​of the difference between the baseline upper limit data of baseline frequency domain data space and the data to be detected.

[0070] When the over-limit difference evaluation index is less than or equal to the fault judgment threshold, the data to be tested is in a normal operating state. When the over-limit difference evaluation index is greater than the fault judgment threshold, the data to be tested is in a fault state.

[0071] According to one embodiment of the present application, based on the baseline frequency domain data space Fre_hil_AIi′ obtained as above k , construct the out-of-limit difference evaluation index DIFF between the data to be tested and the vibration signal baseline data space under the same discrete working condition parameters to evaluate whether the data to be tested is in a fault state, that is:

[0072]

[0073] in, Indicates the frequency domain data amplitude of the data to be tested. When DIFF≤γ, the data to be tested is in normal operation status; when DIFF>γ, the data to be tested is in fault status.

[0074] Therefore, the above method can be relatively intuitive and easier for on-site operators or fault diagnosis engineers to understand the basis and fault status of equipment failures.

[0075] Furthermore, according to the present application, historical operating conditions are discretized, decomposed, processed and stored based on strongly correlated parameters, which is beneficial to refining fault detection scenarios, limiting the impact of different operating condition factors on the effectiveness of fault detection results, and simplifying the fault detection conditions and steps for each operating condition; and, by using a difference sequence and other step sizes to truncate the original data of historical operating condition parameters, it is possible to quickly and accurately obtain discrete operating condition parameter data and corresponding time periods; and based on each discrete operating condition parameter in history, a baseline data vibration time domain data set, a baseline data vibration frequency domain data set, and a baseline frequency domain data space for normal operation of equipment under different discrete operating condition parameters are constructed, providing a reliable evaluation standard and basis method for equipment fault detection.

[0076] When the number of historical operating conditions is limited and the obtained signal data is insufficient, the baseline data space that has been obtained can be used to generate more data to improve or supplement the historical operating condition signal data.

[0077] According to a preferred embodiment of the present application, based on the obtained baseline frequency domain data space and according to the parameter values ​​of the target unknown operating parameters of the equipment, an interpolation algorithm is used to obtain the interpolated baseline frequency domain data space under the target unknown operating parameters.

[0078] The following describes a method for generating an unknown operating condition baseline data space based on the obtained baseline frequency domain data space by way of an embodiment.

[0079] 3. Automatic generation of spatial interpolation fitting of unknown working condition baseline data

[0080] Traverse k=1,2,3,…, and change Fre_hil_AIi′ k Single channel baseline frequency domain data space, that is, a two-dimensional list Superimpose them one by one to construct a 2*k-dimensional joint distribution data frame (joint distribution matrix), as shown below:

[0081]

[0082] Because the parameters that are highly correlated with the spectral structure distribution, such as speed, affect the frequency and cannot be directly interpolated and fitted on the baseline frequency domain data space, while the operating parameters are other parameters that are not highly correlated with the spectral structure distribution, they can be directly interpolated and fitted on the baseline frequency domain data space. If the discrete operating parameter in the above steps is speed, it is necessary to transform the frequency value to the order value. If it is other parameters that are not highly correlated with the spectral structure distribution, it is not necessary to transform the frequency value to the order value.

[0083] The transformation step from frequency value to order value is mainly based on the joint distribution data frame (joint distribution matrix) Fre_hil2_AIi′ k , for (AI′ fre_0 ,AI′ fre_1 ,…,AI′ fre_k ) in AI′ fre_i (where i = 0, 1, 2, ..., k), perform the following calculation to convert the frequency value into the order value. The calculation formula is as follows:

[0084]

[0085] Among them, AI′ JC_i Represents the values ​​of each order of the baseline data space under a single constant speed. Therefore, the joint distribution data frame (joint distribution matrix) Fre_hil2_AIi′ k , can be converted into a joint distribution data frame (joint distribution matrix) under the order dimension, that is,

[0086]

[0087] For Fre_hil2_AIi′k Or JC_mean_AIi′ (when the discrete operating parameter is speed), according to the target unknown discrete operating parameter speed x (where x does not belong to [0, k]) Set the interpolation resolution to H (the resolution can be customized according to the situation), and use the relevant interpolation algorithm (the interpolation algorithm may include but is not limited to the multi-strip interpolation algorithm, the Lagrange interpolation algorithm, etc.) according to each frequency dimension AI′ fre_i Or AI′ of various orders JC_i Speed x The frequency amplitudes or orders under the target are interpolated to obtain the baseline data frequency domain space Fre_hil_AIi′ under the unknown working condition parameters. X Or the baseline data order space JC_mean_AIx′.

[0088] If the discrete operating parameter is the speed, the order value to frequency value conversion operation needs to be performed. If it is other operating parameters that are not highly correlated with the spectrum structure distribution, the order value to frequency value conversion operation can be skipped.

[0089] The transformation operation from order value to frequency value is mainly to inversely transform the order value of the baseline data under the interpolation target working condition parameters through the following formula, restore it back to the frequency value, and obtain the baseline frequency domain data space Fre_hil_AIi′ under the interpolation target working condition parameters X , the specific calculation formula is as follows:

[0090]

[0091] Therefore, according to the above method, when the discrete operating parameter is the speed, it is necessary to obtain Fre_hil2_AIi′ k AI in fre_i Perform the transformation step from frequency value to order value, interpolate the joint distribution data frame (joint distribution matrix) under the order dimension to obtain the baseline data order space JC_mean_AIx′ under the target unknown working condition parameters, and perform the order value to frequency value transformation operation on the order value of the baseline data order space JC_mean_AIx′ obtained after interpolation to obtain the interpolated baseline data frequency domain space Fre_hil_AIi′ X .

[0092] When the discrete operating parameters are other parameters that are not highly correlated with the spectrum structure distribution, there is no need to obtain Fre_hil2_AIi′ k AI in fre_iPerform the frequency value to order value conversion step and perform the order value to frequency value conversion operation on the order value of the interpolated baseline data order space JC_mean_AIx′, and perform the order value to frequency value conversion operation on Fre_hil2_AIi′ k Interpolation is performed to obtain the interpolation baseline data frequency domain space Fre_hil_AIi′ X .

[0093] The interpolation effect is judged by calculating the separation index (sum_diff, the sum of the absolute values ​​of the amplitude differences) of the corresponding frequencies in the original baseline data space and the interpolated baseline frequency domain data space.

[0094] Specifically, set the evaluation index threshold β, when sum diff ≤β (β is customized based on experience), the interpolation effect is considered good; otherwise, the interpolation effect is poor.

[0095] If the interpolation effect is appropriate or good, the interpolated baseline frequency domain data space is added to the existing baseline frequency domain data space and is taken as a part of the existing baseline frequency domain data space.

[0096] If the interpolation effect is not appropriate or poor, continue to supplement the parameter values ​​of the equipment's operating parameters, such as speed i The historical vibration source signal data under the adjacent parameter values ​​are supplemented, and based on the baseline frequency domain data space and the supplemented signal data, the above-mentioned interpolation algorithm is used to obtain the interpolation baseline frequency domain data space under the target unknown working condition parameters, and the interpolation effect is judged again according to the evaluation index threshold β until the interpolation baseline frequency domain data space with good interpolation effect is obtained, and the interpolation baseline frequency domain data space is added to the existing baseline frequency domain data space and used as a part of the existing baseline frequency domain data space.

[0097] According to the present application, based on the baseline frequency domain data space under the historical finite discrete operating condition parameters, the interpolation fitting generation of the unknown new operating condition baseline data space is performed, which enriches and improves the discrete operating condition and baseline data space, and avoids the limitation of the baseline data space for fault detection caused by the lack of historical operating conditions.

[0098] According to the above method, a baseline frequency domain data space is obtained, and the obtained baseline frequency domain data space can be verified by the method described below to confirm whether the obtained baseline frequency domain data space is effective in detecting equipment failures.

[0099] 4. Verification of the effectiveness of baseline data space fault detection

[0100] Obtain the original data of normal vibration signals corresponding to the target channel and target operating parameters and the vibration frequency domain data under the equipment fault state.

[0101] The sample data is divided according to a certain ratio to generate the baseline frequency domain data space and the frequency domain data to be detected (including normal state baseline frequency domain data and fault state frequency domain data).

[0102] Calculate the frequency amplitude out-of-limit difference between the target frequency domain data to be detected and the baseline frequency domain data space, and generate the frequency domain out-of-limit difference distribution of the target frequency domain data to be detected.

[0103] The distribution of out-of-limit differences is compared with the fault judgment threshold, and the spatial distribution of the data to be tested and the baseline data can be analyzed and compared through visual charts.

[0104] Based on the obtained comparison results (ie, data labels), the accuracy of the detection results is verified.

[0105] The following describes several specific embodiments to illustrate the effectiveness of the obtained baseline frequency domain data space and the interpolated baseline frequency domain data space under target unknown operating parameters.

[0106] 5. Example of verification of the effectiveness of baseline data spatial fault detection and the effectiveness of spatial interpolation fitting of unknown working condition baseline data

[0107] Using vibration signal data collected from fracturing pump wellsite operations as sample source data, three comparative examples (Examples 1 to 3) were set up to verify the effectiveness of spatial fault detection using baseline frequency domain data: verification of normal state data from the same unit, verification of normal state data from different units, and verification of equipment fault states from different units. In addition, Example 4 was set up to verify the effectiveness of spatial interpolation fitting of baseline frequency domain data under unknown operating conditions.

[0108] The three groups of comparative vibration signal data are all collected by the vibration sensor of the fracturing pump reducer, and the operating parameter is the rotational speed. The specific sensor installation positions are: Example 1 and Example 2 use the reducer to input the opposite side V channel data, and Example 3 and Example 4 use the reducer to input the H channel data. And the above data acquisition frequency is all 51.2k HZ, and the fixed number of spectral lines fr=12800 and the fixed number of sampling points fd=51200 are set in the fast Fourier transform. And based on the accumulated calculation experience of historical data, the threshold value γ of the evaluation index of the over-limit difference is set to 1.5. The following constructs the above-mentioned baseline frequency domain data space for the three groups of data, and calculates the over-limit difference evaluation index. The results are shown as follows:

[0109] Example 1:

[0110] From the historical operation data of a certain device on a certain platform, the period from November 4 to November 14, 2021, with a discrete constant speed of 1200 rpm, was selected as the control group. The accumulated normal operation time of 8 hours was used as the control group to generate the baseline frequency domain data space under this constant speed condition. From October 23 to October 30, 2021, the accumulated normal operation time of 8 hours was divided into four test groups. Each test group generated 7200 sets of segmented frequency domain data under this constant speed condition. Next, the out-of-limit difference evaluation index of the test group data was calculated based on the baseline frequency domain data space generated by the control group. The results are shown below:

[0111]

[0112] Table 1

[0113] Table 1 shows that among the four test data sets, the second group has the most samples with maximum values ​​of out-of-limit differences greater than zero. However, the maximum value of the out-of-limit difference evaluation index across all its sample data is around 0.3, indicating a very low out-of-limit difference level. This is also true for the other groups, where the maximum values ​​of the out-of-limit difference evaluation index are all relatively low and below the threshold γ. This indicates that this test data is completely covered by its baseline frequency domain data space and is in normal operating condition, which is consistent with the actual data. Therefore, equipment fault detection based on the baseline data space is effective for normal operating data from the same unit.

[0114] Example 2:

[0115] From the historical operating data of two different devices on a certain platform, the vibration signal data of the discrete constant speed of 1100rpm from October 7 to October 17, 2021, with a cumulative normal operating time of 8 hours, was selected as the control group to generate the baseline frequency domain data space under the constant speed condition; from October 23 to October 30, 2021, the vibration signal data of the cumulative normal operating time of 8 hours was divided into four test groups, and each test group generated 7200 groups of segmented frequency domain data under the constant speed. Next, the out-of-limit difference evaluation index of the test group data was calculated based on the baseline frequency domain data space generated by the control group. The experimental results are shown as follows:

[0116]

[0117] Table 2

[0118] Table 2 shows that among the four test data sets, the fourth group has the largest number of samples with a maximum value greater than zero for the out-of-limit difference evaluation index. However, the maximum value of the out-of-limit difference evaluation index across all its sample data is approximately 0.2, indicating a very low level. This is also true for the other groups, where the maximum values ​​of the out-of-limit difference evaluation index are all relatively low and below the threshold γ. This indicates that this test data is completely covered by its baseline data space and is in normal operating condition, consistent with the actual data. This demonstrates that equipment fault detection based on the baseline data space is effective for normal state data from different units.

[0119] Example 3:

[0120] From the historical operation data of two different devices on a certain platform, the vibration signal data with a discrete constant speed working condition of 1080rpm from November 6 to November 16, 2021, with a cumulative normal operation time of 8 hours was selected as the control group to generate the baseline frequency domain data space under the constant speed; from December 20 to December 29, 2021, the vibration signal data with a cumulative operation time of 4 hours was averaged into 2 test groups. Among them, the equipment status corresponding to the test group is damage to the reducer bearing roller, that is, the equipment failure state. Each test group generates a total of 7200 groups of segmented frequency domain data under the constant speed. Below, the out-of-limit difference evaluation index of the test group data is calculated based on the baseline frequency domain data space generated by the control group. The experimental results are shown as follows:

[0121]

[0122] Table 3

[0123] Table 3 shows that the spectral structure of the equipment fault state at the same speed completely exceeds the limits of the baseline frequency domain data space. The maximum value of the out-of-limit difference evaluation index reaches over 10, significantly exceeding the threshold γ, and all sample data are basically out of limit. The baseline frequency domain data space effectively reflects the damage to the equipment gearbox bearing roller, and the response is consistent with the actual data fault state. This shows that equipment fault detection based on the baseline data space is effective for fault state data of different units.

[0124] From the above three experiments, it can be seen that, regardless of whether it is the same unit or different units, under the same operating conditions, the baseline data space can intuitively and effectively detect the equipment failure state and has no relevant response when the equipment is in normal state.

[0125] Example 4:

[0126] The sample data of different discrete constant speed working conditions are screened from the historical operation data of a certain equipment on a certain platform, and the above baseline data space construction steps are performed to generate Fre_hil_AIi′ under different speed working conditions. k. The discrete operating speeds are [1000, 1100, 1200, 1300, 1400, 1500, 1600, 1700, 1800, 1900, 2000, 2100], with a total of twelve discrete speed conditions corresponding to 12 baseline data spaces. The speeds in the above list are traversed and eliminated one by one, and the interpolation resolution H is set to 12. The amplitudes of each order of the baseline data space under the remaining discrete speed conditions are used to interpolate the amplitudes of each order of the eliminated speed conditions (the interpolation method uses the cubic spline interpolation algorithm). Then, the order inverse transform is performed to obtain the spectrum interpolation data, and the sum of the absolute values ​​of the amplitude difference between the interpolated data and the original data is calculated as the sum_diff evaluation index (i.e., the separation index). The threshold β is set to 5.5 based on the statistical results of historical accumulated experimental data and expert experience. The sum_diff calculation results of each group of interpolation sample data are as follows:

[0127]

[0128] Table 4

[0129] As can be seen from Table 4 above, in the 12 groups of interpolation experiments, the sum_diff evaluation index is all less than β, that is, the interpolation effect is relatively good, the interpolation data accuracy is relatively high, and the baseline data space under the interpolation speed condition has relatively practical application significance and value.

[0130] According to this application, by discretizing, processing and storing historical working conditions based on strongly correlated parameters, it is beneficial to refine the fault detection scenarios, limit the impact of different working condition factors on the effectiveness of fault detection results, and simplify the fault detection conditions and steps for each working condition.

[0131] Moreover, the use of difference sequence and other step sizes to truncate the original data of historical operating condition parameters can quickly and accurately obtain discrete operating condition parameter data and the corresponding time period.

[0132] Furthermore, based on the historical discrete operating parameters, the baseline data vibration raw data set, baseline data vibration frequency domain data set, and baseline data space of the equipment's normal operation under different discrete operating parameters are constructed, providing a reliable evaluation standard and basis method for equipment fault detection.

[0133] In addition, based on the baseline data space under historical finite discrete operating condition parameters, the interpolation fitting of the baseline data space of unknown operating conditions is automatically generated, which enriches and improves the discrete operating conditions and baseline data space, and avoids the limitation of fault detection in the baseline data space caused by the lack of historical operating conditions.

[0134] Furthermore, by setting an out-of-limit difference evaluation index, faults can be responded to intuitively, making it easier for relevant business personnel to understand and grasp fault information and the degree of fault degradation.

[0135] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0136] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

Claims

1. A device fault detection method based on baseline data space, characterized in that: The device fault detection method comprises: determining an operating condition parameter having a predetermined correlation with an operating condition of the device; Discretizing the original data set of the operating parameters of the equipment according to the fixed-length time periods of the operating parameters of the equipment to generate a discrete operating parameter set containing data of multiple fixed-length time periods; Constructing a baseline frequency domain data space of the device in a normal state according to the discrete operating condition parameter set and the acquired signal data of the device in a normal state, and setting a fault judgment threshold; Based on the baseline frequency domain data space, according to the parameter value of the target unknown operating parameter of the equipment, using an interpolation algorithm, an interpolation baseline frequency domain data space under the target unknown operating parameter is obtained; and Calculating an over-limit difference evaluation index, and comparing the over-limit difference evaluation index with the fault judgment threshold to determine whether the device has a fault, The operating condition parameter is a parameter that is highly correlated with the spectral structure distribution. Based on the baseline frequency domain data space and according to the parameter value of the target unknown operating condition parameter of the equipment, an interpolation algorithm is used to obtain an interpolation baseline frequency domain data space under the target unknown operating condition parameter, including: The baseline frequency domain data spaces are sequentially superimposed to construct a joint distribution matrix; Convert the frequency values ​​in the joint distribution matrix into order values ​​to obtain a joint distribution matrix in the order dimension; According to the parameter value of the target unknown operating parameter of the equipment, using an interpolation algorithm, interpolation calculation is performed on the joint distribution matrix under the order dimension to obtain the interpolation baseline data order space under the target unknown operating parameter; The order value of the interpolation baseline data order space under the target unknown operating parameters is inversely transformed and restored to the frequency value to obtain the interpolation baseline frequency domain data space under the target unknown operating parameters.

2. The device fault detection method according to claim 1, characterized in that: The original data set is discretized according to the fixed-length time periods of the operating parameters of the equipment to generate a discrete operating parameter set containing data of multiple fixed-length time periods, including: Based on the original data set, a new data set at the delayed time point is obtained by lagging the original data set by one time point, wherein the new data set is of the same length as the original data set; Calculating a difference sequence between the new data set and the original data set to generate a difference sequence set, setting a fault tolerance threshold σ, and setting data values ​​of the difference sequence <σ to zero; and A time length threshold t_lim is set, and truncation data of the difference sequence whose continuous zero time length is greater than t_lim is truncated to generate the discrete operating condition parameter set.

3. The device fault detection method according to claim 1 or 2, characterized in that: Constructing a baseline frequency domain data space of the device in a normal state according to the discrete operating condition parameter set and the acquired signal data, including: Obtaining a baseline time domain data set and a baseline frequency domain data set according to the discrete operating condition parameter set; performing bandpass filtering on the baseline frequency domain data set to generate a second baseline frequency domain data set corresponding to the operating condition parameter in the frequency band of interest; and The baseline frequency domain data space is constructed based on the second baseline frequency domain data set.

4. The device fault detection method according to claim 3, characterized in that: According to the discrete operating condition parameter set, a baseline time domain data set and a baseline frequency domain data set are obtained, including: Discretizing and segmenting the signal data corresponding to the time nodes in the signal data under the normal state, and marking the parameter values ​​of the operating condition parameters corresponding to the time periods of the segmented data to generate the baseline time domain data set; and Performing a fast Fourier transform on the baseline time domain data set to obtain the baseline frequency domain data set.

5. The device fault detection method according to claim 3, characterized in that: Constructing the baseline frequency domain data space based on the second baseline frequency domain data set includes: Expand each segment of spectrum data in the second baseline frequency domain data set one by one according to the frequency dimension to obtain a frequency matrix; Calculating statistical index values ​​for the frequency matrix column by column to obtain the upper limit space of the baseline frequency domain; Performing Hilbert transform on the upper limit space of the baseline frequency domain and calculating the upper envelope to obtain the baseline frequency domain data space.

6. The device fault detection method according to claim 5, characterized in that: The statistical indicator values ​​include: mean, median, maximum and extreme values.

7. The device fault detection method according to claim 1, characterized in that: The device fault detection method further includes: Setting an evaluation index threshold of the interpolation baseline frequency domain data space; Calculating the separation between the baseline frequency domain data space and the interpolated baseline frequency domain data space; The separation degree is compared with the evaluation index threshold to determine whether the interpolation baseline frequency domain data space is appropriate.

8. The device fault detection method according to claim 7, characterized in that: The device fault detection method further includes: Based on the comparison result of the separation degree and the evaluation index threshold, it is determined that the interpolated baseline frequency domain data space is appropriate, and the interpolated baseline frequency domain data space is added to the baseline frequency domain data space as a part of the baseline frequency domain data space; or Based on the comparison result of the separation degree and the evaluation index threshold, determining that the interpolation baseline frequency domain data space is inappropriate, and supplementing the signal data under the parameter value of the target unknown operating condition parameter that is adjacent to the parameter value; and Based on the baseline frequency domain data space and the supplemented signal data, an interpolation algorithm is used to obtain an interpolation baseline frequency domain data space under the target unknown operating parameters.

9. The device fault detection method according to claim 1, characterized in that: The operating condition parameter is a parameter that is not highly correlated with the frequency spectrum structure distribution. Based on the baseline frequency domain data space and according to the parameter value of the target unknown operating condition parameter of the equipment, an interpolation algorithm is used to obtain an interpolation baseline frequency domain data space under the target unknown operating condition parameter, including: The baseline frequency domain data spaces are sequentially superimposed to construct a joint distribution matrix; According to the parameter value of the target unknown operating parameter of the equipment, an interpolation algorithm is used to perform interpolation calculation on the joint distribution matrix to obtain an interpolation baseline frequency domain data space under the target unknown operating parameter.

10. The device fault detection method according to claim 7, characterized in that: The separation degree is the sum of the absolute values ​​of the amplitude differences between the baseline frequency domain data space and the interpolation baseline frequency domain space.

11. The device fault detection method according to claim 1, characterized in that: The device fault detection method further includes: performing a cleaning process on the acquired original data set to supplement missing values ​​and remove zero values, and using the cleaned data set as the original data set.

12. The device failure detection method according to claim 11, characterized in that: The missing value is supplemented by averaging the nearest neighbor data before and after the missing value.

13. The device fault detection method according to claim 1, characterized in that: The device fault detection method further includes: Acquire signal data of the device in a normal state, wherein the signal data includes vibration signal data.

14. The device fault detection method according to claim 1, characterized in that: Out-of-limit difference evaluation index = sum(abs(baseline upper limit data of baseline frequency domain data space - data to be detected)) Among them, abs(baseline upper limit data of baseline frequency domain data space - data to be detected) means finding the absolute value of the difference between the baseline upper limit data of baseline frequency domain data space and the data to be detected, and sum(abs(baseline upper limit data of baseline frequency domain data space - data to be detected)) means finding the sum of the absolute values ​​of the difference between the baseline upper limit data of baseline frequency domain data space and the data to be detected.

15. A computer device comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 14 are implemented.

16. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 14 are implemented.

Images