CN117473288A - Fault analysis method, analysis device and storage medium based on l1 trend filtering - Google Patents
Fault analysis method, analysis device and storage medium based on l1 trend filtering Download PDFInfo
- Publication number
- CN117473288A CN117473288A CN202210843196.5A CN202210843196A CN117473288A CN 117473288 A CN117473288 A CN 117473288A CN 202210843196 A CN202210843196 A CN 202210843196A CN 117473288 A CN117473288 A CN 117473288A
- Authority
- CN
- China
- Prior art keywords
- trend
- value
- determining
- fault
- spectral distance
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000001914 filtration Methods 0.000 title claims abstract description 56
- 238000004458 analytical method Methods 0.000 title claims abstract description 47
- 238000003860 storage Methods 0.000 title claims abstract description 9
- 230000003595 spectral effect Effects 0.000 claims abstract description 82
- 238000001228 spectrum Methods 0.000 claims abstract description 64
- 238000012512 characterization method Methods 0.000 claims abstract description 44
- 230000015556 catabolic process Effects 0.000 claims abstract description 32
- 238000006731 degradation reaction Methods 0.000 claims abstract description 32
- 238000000034 method Methods 0.000 claims abstract description 30
- 238000012545 processing Methods 0.000 claims abstract description 20
- 230000035945 sensitivity Effects 0.000 claims abstract description 13
- 238000003491 array Methods 0.000 claims description 26
- 238000004364 calculation method Methods 0.000 claims description 13
- 238000000354 decomposition reaction Methods 0.000 claims description 11
- 230000009191 jumping Effects 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 3
- 230000002159 abnormal effect Effects 0.000 abstract description 7
- 238000010586 diagram Methods 0.000 description 6
- 238000012544 monitoring process Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 229940060587 alpha e Drugs 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G07—CHECKING-DEVICES
- G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
- G07C3/00—Registering or indicating the condition or the working of machines or other apparatus, other than vehicles
- G07C3/005—Registering or indicating the condition or the working of machines or other apparatus, other than vehicles during manufacturing process
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F04—POSITIVE - DISPLACEMENT MACHINES FOR LIQUIDS; PUMPS FOR LIQUIDS OR ELASTIC FLUIDS
- F04D—NON-POSITIVE-DISPLACEMENT PUMPS
- F04D27/00—Control, e.g. regulation, of pumps, pumping installations or pumping systems specially adapted for elastic fluids
- F04D27/001—Testing thereof; Determination or simulation of flow characteristics; Stall or surge detection, e.g. condition monitoring
Landscapes
- Engineering & Computer Science (AREA)
- Manufacturing & Machinery (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Mechanical Engineering (AREA)
- General Engineering & Computer Science (AREA)
- Structures Of Non-Positive Displacement Pumps (AREA)
Abstract
The invention discloses a fault analysis method, an analysis device and a storage medium based on l1 trend filtering, wherein the method comprises the following steps: determining a first spectral distance dataset of the centrifugal compressor in a normal operating state; acquiring a second spectrum distance data set of the centrifugal compressor in a real-time running state; creating a sliding combination window based on the first spectral distance dataset and the second spectral distance dataset; determining regularization parameters based on the sliding combination window; and processing the sliding combination window based on the l1 trend filtering rule and the regularization parameter to obtain a fault trend characterization factor for the centrifugal compressor, wherein the fault trend characterization factor is used for predicting and analyzing faults of the centrifugal compressor. The spectral distance index of the centrifugal compressor is extracted to serve as an early fault sensitivity characteristic value, and a fault trend factor is further determined, so that the influence of sporadic abnormal points is reduced, fluctuation of data is reduced, the degradation trend of the data is increased, and the accuracy and the analysis efficiency of fault analysis are improved.
Description
Technical Field
The invention relates to the technical field of fault analysis, in particular to a fault analysis method based on l1 trend filtering, a fault analysis device based on l1 trend filtering and a computer readable storage medium.
Background
The centrifugal compressor is widely applied to various application scenes in daily life and industrial production of people, and safe and reliable operation of the centrifugal compressor can ensure normal operation of living equipment and industrial equipment and bring good economic benefit and social benefit, so that monitoring work for ensuring normal operation of the centrifugal compressor is very important.
In order to improve the accuracy of monitoring the centrifugal compressor, a technician desires to monitor the operation state of the centrifugal compressor and predict faults or anomalies thereof in advance, for example, technologies such as early warning, health assessment, life prediction and the like for the centrifugal compressor are important support technologies for predictive maintenance of equipment of the centrifugal compressor.
The technical staff monitors the centrifugal compressor by adopting a method based on l1 trend filtering, however, in the practical application process, the technical staff find that the existing monitoring method at least has the following technical problems:
because the performance of the l1 trend filtering depends on whether the selected regularization parameters are proper or not, the existing conventional parameter determination method is insufficient in accuracy and cannot meet the actual requirements; meanwhile, the l1 trend filtering can not meet the requirement of filtering processing of full life cycle data in the engineering practical application process, so that the accuracy of monitoring data is further reduced.
Disclosure of Invention
In order to overcome the technical problems in the prior art, the embodiment of the invention provides a fault analysis method based on l1 trend filtering, which is used for extracting a spectrum distance index of a centrifugal compressor as an early fault sensitivity characteristic value and further determining a fault trend factor, so that the influence of sporadic abnormal points is reduced, the fluctuation of data is reduced, the degradation trend of the data is increased, and the accuracy and analysis efficiency of fault analysis are improved.
In order to achieve the above object, an embodiment of the present invention provides a fault analysis method based on l1 trend filtering, the method including: determining a first spectral distance dataset of the centrifugal compressor in a normal operating state; acquiring a second spectral distance dataset of the centrifugal compressor in a real-time state; creating a sliding combination window based on the first spectral distance dataset and the second spectral distance dataset; determining a regularization parameter based on the sliding combination window; and processing the sliding combination window based on the l1 trend filtering rule and the regularization parameter to obtain a fault trend characterization factor for the centrifugal compressor, wherein the fault trend characterization factor is used for carrying out predictive analysis on faults of the centrifugal compressor.
Preferably, said determining a first spectral distance dataset of the centrifugal compressor in a normal operating state comprises: acquiring historical operation data of a full life cycle of the centrifugal compressor; determining a historical spectral distance dataset based on the historical operating data calculation; performing time continuous decomposition operation on the historical spectrum distance data set according to a preset length to obtain a plurality of corresponding spectrum distance arrays; determining a degradation trend factor based on the plurality of spectral distance arrays; a first spectral distance dataset is determined in the plurality of spectral distance arrays based on the degradation trend factor.
Preferably, said determining a historical spectral distance dataset based on said historical operating data calculation comprises:calculating corresponding spectral distance data R (J) xy ) The spectral distance data R (J xy ) Characterized by:wherein R (J) xy )∈(0～1)，J xy Characterized by the J-divergence between the operating normal state signal x (t) and the state signal to be evaluated y (t), αe (0-1) characterized by a sensitivity coefficient, wherein:wherein S is x (k) And S is y (k) The self-power spectrums are respectively characterized as signals x (t) and y (t), and P is the number of power spectrum lines; at the spectral distance data R (J xy ) And extracting a historical spectrum distance data set of the centrifugal compressor in a normal running state.
Preferably, the determining a degradation trend factor based on the plurality of spectral distance arrays includes: determining an initial regularization parameter value based on a preset data range; performing preliminary l1 trend filtering operation on each spectrum distance array in sequence based on the initial regularization parameter values to obtain corresponding filtered curves; calculating and determining the average value of the upper derivative of each filtered curve; and determining the upper derivative average value with the smallest absolute value as the minimum value, and taking the minimum value as the degradation trend factor.
Preferably, the determining regularization parameters based on the sliding combination window includes: s41) determining an initial value of the regularization parameter; s42) performing derivative operation on the sliding combination window based on the initial value to obtain a corresponding first derivative value; s43) performing self-increment operation on the initial value based on a preset self-increment rule to obtain corresponding self-increment value; s44) performing derivative operation on the sliding combination window based on the self-increment value to obtain a corresponding second derivative value; s45) judging whether the deviation between the first derivative value and the second derivative value is smaller than a preset deviation threshold value or not; s461) if yes, determining the self-increment value as the regularization parameter; s462) if not, jumping to step S43) taking the self-increment value as a new initial value.
Preferably, the processing the sliding combination window based on the l1 trend filtering rule and the regularization parameter to obtain a fault trend characterization factor for the centrifugal compressor includes: processing the sliding combination window based on the l1 trend filtering rule and the regularization parameter to obtain a corresponding processed line segment; acquiring slope information of the processed line segment; and taking the absolute value of the slope information as the fault trend characterization factor.
Preferably, the method further comprises: after the fault trend characterization factor is obtained, judging whether the fault trend characterization factor is larger than a preset slope threshold value or not; and outputting corresponding early-stage fault early-warning information under the condition that the fault trend characterization factor is larger than the preset slope threshold value.
Correspondingly, the embodiment of the invention also provides a fault analysis device based on the l1 trend filtering, which comprises: a determining unit for determining a first spectral distance dataset of the centrifugal compressor in a normal operating state; a data acquisition unit for acquiring a second spectral distance dataset in a real-time state of the centrifugal compressor; a sliding data creation unit for creating a sliding combination window based on the first spectral distance data set and the second spectral distance data set; a parameter calculation unit for determining regularization parameters based on the sliding combination window; the fault analysis unit is used for processing the sliding combination window based on the l1 trend filtering rule and the regularization parameter to obtain a fault trend characterization factor for the centrifugal compressor, and the fault trend characterization factor is used for carrying out predictive analysis on faults of the centrifugal compressor.
Preferably, the determining unit includes: the historical data acquisition module is used for acquiring historical operation data of the full life cycle of the centrifugal compressor; a first dataset determination module for computationally determining a historical spectral distance dataset based on the historical operating data; the decomposition module is used for performing time continuous decomposition operation on the historical spectrum distance data set according to a preset length to obtain a plurality of corresponding spectrum distance arrays; a trend factor determining module for determining a degradation trend factor based on the plurality of spectral distance arrays; a second data set determining module for determining a first spectral distance data set in the plurality of spectral distance arrays based on the degradation trend factor.
Preferably, the first data set determining module is specifically configured to: calculating corresponding spectral distance data R (J) xy ) The spectral distance data R (J xy ) Characterized by: wherein R (J) xy )∈(0～1)，J xy Characterized by the J-divergence between the operating normal state signal x (t) and the state signal to be evaluated y (t), αe (0-1) characterized by a sensitivity coefficient, wherein: />Wherein S is x (k) And S is y (k) The self-power spectrums are respectively characterized as signals x (t) and y (t), and P is the number of power spectrum lines; at the spectral distance data R (J xy ) And extracting a historical spectrum distance data set of the centrifugal compressor in a normal running state.
Preferably, the trend factor determining module is specifically configured to: determining an initial regularization parameter value based on a preset data range; performing preliminary l1 trend filtering operation on each spectrum distance array in sequence based on the initial regularization parameter values to obtain corresponding filtered curves; calculating and determining the average value of the upper derivative of each filtered curve; and determining the upper derivative average value with the smallest absolute value as the minimum value, and taking the minimum value as the degradation trend factor.
Preferably, the parameter calculation unit is specifically configured to: s41) determining an initial value of the regularization parameter; s42) performing derivative operation on the sliding combination window based on the initial value to obtain a corresponding first derivative value; s43) performing self-increment operation on the initial value based on a preset self-increment rule to obtain corresponding self-increment value; s44) performing derivative operation on the sliding combination window based on the self-increment value to obtain a corresponding second derivative value; s45) judging whether the deviation between the first derivative value and the second derivative value is smaller than a preset deviation threshold value or not; s461) if yes, determining the self-increment value as the regularization parameter; s462) if not, jumping to step S43) taking the self-increment value as a new initial value.
Preferably, the fault analysis unit includes: the data processing module is used for processing the sliding combination window based on the l1 trend filtering rule and the regularization parameter to obtain a corresponding processed line segment; the slope obtaining module is used for obtaining slope information of the processed line segment; and the fault factor determining module is used for taking the absolute value of the slope information as the fault trend characterization factor.
Preferably, the device further comprises an early warning unit, and the early warning unit is specifically configured to: after the fault trend characterization factor is obtained, judging whether the fault trend characterization factor is larger than a preset slope threshold value or not; and outputting corresponding early-stage fault early-warning information under the condition that the fault trend characterization factor is larger than the preset slope threshold value.
In another aspect, the embodiment of the present invention further provides a computer readable storage medium, where a computer program is stored, where the program when executed by a processor implements the method provided by the embodiment of the present invention.
Through the technical scheme provided by the invention, the invention has at least the following technical effects:
by extracting the spectrum distance index of the centrifugal compressor as the early fault sensitivity characteristic value and further calculating and determining the trend factor, compared with the early fault sensitivity characteristic value in the prior art, the influence of accidental abnormal points can be effectively reduced, fluctuation of data is reduced, and the degradation trend of the data is increased.
In the second aspect, the actual trend of the time sequence is extracted by adopting a method based on l1 trend filtering, so that interference of abnormal points is reduced, and meanwhile, the extraction accuracy of the actual trend of the time sequence can be further improved by combining with a method for adaptively determining regularization parameters, so that the calculation efficiency is improved, and the actual fault analysis and prediction requirements are met.
Additional features and advantages of embodiments of the invention will be set forth in the detailed description which follows.
Drawings
The accompanying drawings are included to provide a further understanding of embodiments of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain, without limitation, the embodiments of the invention. In the drawings:
FIG. 1 is a flowchart of a specific implementation of a fault analysis method based on l1 trend filtering according to an embodiment of the present invention;
FIG. 2 is a flowchart of a specific implementation of determining a spectral distance dataset in a fault analysis method based on l1 trend filtering according to an embodiment of the present invention;
FIG. 3 is a flowchart of a specific implementation of determining regularization parameters in a fault analysis method based on l1 trend filtering according to an embodiment of the present invention;
FIG. 4 is a flowchart of a specific implementation of obtaining a fault trend characterization factor in the fault analysis method based on l1 trend filtering according to an embodiment of the present invention;
FIG. 5 is a schematic diagram of a fault analysis method based on l1 trend filtering according to a first embodiment of the present invention;
FIG. 6 is a schematic diagram of a fault analysis method based on l1 trend filtering according to a second embodiment of the present invention;
fig. 7 is a schematic structural diagram of a fault analysis device based on l1 trend filtering according to an embodiment of the present invention.
Detailed Description
The following describes the detailed implementation of the embodiments of the present invention with reference to the drawings. It should be understood that the detailed description and specific examples, while indicating and illustrating the invention, are not intended to limit the invention.
The terms "system" and "network" in embodiments of the invention may be used interchangeably. "plurality" means two or more, and "plurality" may also be understood as "at least two" in this embodiment of the present invention. "and/or", describes an association relationship of an association object, and indicates that there may be three relationships, for example, a and/or B, and may indicate: a exists alone, A and B exist together, and B exists alone. The character "/", unless otherwise specified, generally indicates that the associated object is an "or" relationship. In addition, it should be understood that in the description of embodiments of the present invention, the words "first," "second," and the like are used merely for distinguishing between the descriptions and not be construed as indicating or implying a relative importance or order.
Referring to fig. 1, an embodiment of the present invention provides a fault analysis method based on l1 trend filtering, where the method includes:
s10) determining a first spectrum distance data set of the centrifugal compressor in a normal running state;
s20) acquiring a second spectrum distance data set of the centrifugal compressor in a real-time state;
s30) creating a sliding combination window based on the first spectral distance dataset and the second spectral distance dataset;
s40) determining regularization parameters based on the sliding combination window;
s50) processing the sliding combination window based on the l1 trend filtering rule and the regularization parameter to obtain a fault trend characterization factor for the centrifugal compressor, wherein the fault trend characterization factor is used for carrying out predictive analysis on faults of the centrifugal compressor.
In one possible embodiment, first, index reference data of the centrifugal compressor in a normal operation state is obtained, and specifically, the index reference data is a spectral distance index reference data set of the centrifugal compressor. Referring to fig. 2, in an embodiment of the present invention, determining a first spectral distance dataset of a centrifugal compressor in a normal operation state includes:
s11) acquiring historical operation data of a full life cycle of the centrifugal compressor;
s12) determining a historical spectral distance dataset based on the historical operating data calculation;
s13) performing time continuous decomposition operation on the historical spectrum distance data set according to a preset length to obtain a plurality of corresponding spectrum distance arrays;
s14) determining a degradation trend factor based on the plurality of spectral distance arrays;
s15) determining a first spectral distance dataset in the plurality of spectral distance arrays based on the degradation trend factor.
In one possible implementation manner, first, the full life cycle historical operation data of the centrifugal compressor is obtained, for example, the operation data of the whole life cycle of a certain centrifugal compressor from the beginning of use to the occurrence of a fault may be collected in advance to obtain the historical operation data of the full life cycle of the centrifugal compressor, and a corresponding historical spectrum distance data set is calculated and determined based on the historical operation data, specifically, in the embodiment of the present invention, the step of calculating and determining the historical spectrum distance data set based on the historical operation data includes: calculating corresponding spectral distance data R (J) xy ) The spectral distance data R (J xy ) Characterized by:wherein R (J) xy )∈(0～1)，J xy Characterized by the J-divergence between the operating normal state signal x (t) and the state signal to be evaluated y (t), αe (0-1) characterized by a sensitivity coefficient, wherein:wherein S is x (k) And S is y (k) The self-power spectrums are respectively characterized as signals x (t) and y (t), and P is the number of power spectrum lines; at the spectral distance data R (J xy ) And extracting a historical spectrum distance data set of the centrifugal compressor in a normal running state.
For example, the historical operation data is split firstly, the historical operation data is split into an operation normal state signal x (t) and a state signal y (t) to be evaluated, and the corresponding spectrum distance data R (J) is obtained by calculating the x (t) and the y (t) xy ) Wherein the spectral distance data R (J xy ) Can be achieved by, for exampleThe following formula is calculated:wherein R (J) xy )∈(0～1)，J xy For operating the J divergence between the normal state signal x (t) and the state signal to be evaluated y (t), alpha E (0-1) is the sensitivity coefficient, where J xy The method can be calculated by the following formula: /> Wherein S is x (k) And S is y (k) The self-power spectrum of the signals x (t) and y (t), P is the number of power spectrum lines, and then the spectrum distance data R (J) xy ) And extracting a historical spectrum distance data set of the centrifugal compressor in a normal running state.
At this time, the historical spectrum distance data set is subjected to a time-continuous decomposition operation according to a preset length to obtain a plurality of corresponding spectrum distance arrays, and in one embodiment, the historical spectrum distance data set is marked as X 0 (t s ) X is taken as 0 (t s ) Decomposition into time-continuous arrays [ x ] of length m j ，x j+1 ，..，L，..，x j+m-1 ](j=1, 2,..l,..s-m+1) at this time according to the above-mentioned plurality of arrays [ x j ，x j+1 ，..，L，..，x j+m-1 ]A degradation trend factor is determined.
In an embodiment of the present invention, the determining a degradation trend factor based on the plurality of spectral distance arrays includes: determining an initial regularization parameter value based on a preset data range; performing preliminary l1 trend filtering operation on each spectrum distance array in sequence based on the initial regularization parameter values to obtain corresponding filtered curves; calculating and determining the average value of the upper derivative of each filtered curve; and determining the upper derivative average value with the smallest absolute value as the minimum value, and taking the minimum value as the degradation trend factor.
In one possible embodiment, in determiningIn the process of the degradation trend factor, firstly, determining an initial regularization parameter value based on a preset data range, for example, the initial regularization parameter value is expressed as lambda, the preset data range is 1-100, namely, any value in the lambda value range is determined to be 1-100, then, carrying out preliminary l1 trend filtering operation on each spectrum distance array through randomly determined lambda, thereby obtaining a corresponding filtered curve, specifically, j=1 can be firstly caused, and the spectrum distance array is determined to be [ x ] 1 ，x 2 ，..，L，..，x m ]Placing the spectrum distance array into a fixed window w 1 Performing primary l1 trend filtering operation on the spectrum distance array to obtain a corresponding first filtered curve; and then j=2, repeating the operation to obtain a corresponding second filtered curve until all the filtered curves are obtained after the primary l1 trend filtering operation on all the spectrum distance arrays is completed.
Further calculation at this time determines the upper derivatives of each filtered curve, for example, performs an upper derivative operation for each point on each filtered curve, and obtains a corresponding plurality of upper derivatives, at this time, the average value of all the upper derivatives of the current filtered curve is obtained, that is, based on the same principle, the upper derivative average value of each filtered curve can be obtained, at this time, the minimum value with the smallest absolute value can be determined from the above obtained plurality of upper derivative average values, for example, the minimum value is marked as k 0 And marks the minimum value as k 0 Is determined as a degradation trend factor. At this time, a fault sensitive data set is determined from the plurality of spectrum distance arrays according to the degradation trend factor, and then the spectrum distance array corresponding to the degradation trend factor is used as a first spectrum distance data set.
In the embodiment of the invention, the spectrum distance index of the centrifugal compressor is extracted to serve as the spectrum distance reference data set and the early fault sensitivity characteristic value of the centrifugal compressor, so that the fluctuation of data can be effectively reduced, the degradation trend of the data is increased, and the accuracy of the subsequent fault analysis is improved. After the spectral distance index reference data set of the centrifugal compressor is determined, the faults of the centrifugal compressor can be analyzed and predicted in real time.
Referring to fig. 3, in an embodiment of the present invention, the determining a regularization parameter based on the sliding combination window includes:
s41) determining an initial value of the regularization parameter;
s42) performing derivative operation on the sliding combination window based on the initial value to obtain a corresponding first derivative value;
s43) performing self-increment operation on the initial value based on a preset self-increment rule to obtain corresponding self-increment value;
s44) performing derivative operation on the sliding combination window based on the self-increment value to obtain a corresponding second derivative value;
s45) judging whether the deviation between the first derivative value and the second derivative value is smaller than a preset deviation threshold value or not;
s461) if yes, determining the self-increment value as the regularization parameter;
s462) if not, jumping to step S43) taking the self-increment value as a new initial value.
In one possible implementation, the real-time operation data of the current centrifugal compressor is firstly obtained, and a sliding combination window is created according to the spectrum distance data set and the real-time operation data, for example, the sliding combination window can be a sliding array with the length of m+n, wherein m represents the number of data of a fixed array, and the data in the fixed array consists of data in the spectrum distance data set; n represents the number of data of the active array, and the data are composed of spectrum distance indexes of real-time running data. At this time, planning parameters are determined in a self-adaptive manner according to the sliding combination window.
For example, first determining an initial value of the regularization parameter, for example, a minimum settable value (for example, 1) of the regularization parameter may be set as its initial value, and then performing a derivative operation on the sliding combination window based on the initial value to obtain a corresponding first derivative value; and then gradually increasing the value of the regularization parameter and obtaining a corresponding self-increment value, at the moment, further executing derivative operation on the sliding combination window based on the self-increment value, obtaining a corresponding second derivative value, judging whether the deviation of the first derivative value and the second derivative value is smaller than a preset deviation threshold value, and determining that the derivative value of the sliding combination window is not changed on the basis of the self-increment value when the deviation is smaller than the preset deviation threshold value, wherein the self-increment value is determined to be the regularization parameter at the moment, otherwise, repeatedly executing the self-increment operation and continuously judging whether the derivative value of the derivative on the sliding combination window is changed or not so as to adaptively determine the regularization parameter.
In the embodiment of the invention, the regularization parameters are determined in a self-adaptive mode, and the technician is not required to determine or manually set according to experience, so that the analysis efficiency of fault prediction or fault analysis is improved, and the working efficiency is improved.
Referring to fig. 4, in an embodiment of the present invention, the processing the sliding combination window based on the l1 trend filtering rule and the regularization parameter to obtain a fault trend characterization factor for the centrifugal compressor includes:
s51) processing the sliding combination window based on the l1 trend filtering rule and the regularization parameter to obtain a corresponding processed line segment;
s52) acquiring slope information of the processed line segment;
s53) taking the absolute value of the slope information as the fault trend characterization factor.
Further, in an embodiment of the present invention, the method further includes: after the fault trend characterization factor is obtained, judging whether the fault trend characterization factor is larger than a preset slope threshold value or not; and outputting corresponding early-stage fault early-warning information under the condition that the fault trend characterization factor is larger than the preset slope threshold value.
In one possible implementation manner, after determining the regularization parameter, the sliding combination window is processed based on the l1 trend filtering rule and the regularization parameter to obtain a processed line segment, and at this time, slope information of the processed line segment is obtained, and an absolute value of the slope information is used as a fault trend characterization factor.
At this time, whether the fault trend characterization factor is larger than a preset slope threshold value can be judged, if so, the risk of faults of the current centrifugal compressor can be determined, so that corresponding early fault early warning information is immediately output to prompt corresponding technicians to process in time.
In the embodiment of the invention, the method is carried out by carrying out the steps according to the time series data with the limited length 1 During trend filtering, the fitted curve becomes a line segment characteristic along with the increase of the regularization parameter lambda, and the time sequence data is subjected to l 1 The trend filtering, the adaptive calculation and the determination of the regularization parameter value, and the further determination of the corresponding fault trend factor, compared with the early fault sensitivity characteristic value in the prior art, the influence of occasional abnormal points can be effectively reduced, the fluctuation of data is reduced, the degradation trend of the data is increased, and the actual fault prediction and analysis requirements are met.
Specifically, referring to fig. 5, a schematic diagram of a fault analysis method based on l1 trend filtering according to a first embodiment of the present invention is shown, in the operation process of the centrifugal compressor, the centrifugal compressor is sporadically abnormal, a line diagram before spectral distance index filtering is formed by a plurality of corresponding sample points according to sampling operation data of the centrifugal compressor, and then a corresponding line segment can be obtained after the operation data of the centrifugal compressor is filtered based on the spectral distance index.
Referring to fig. 6, a schematic diagram of a fault analysis method based on l1 trend filtering according to a second embodiment of the present invention is shown, in which there are 5 consecutive points deviating from normal data points, and even if the deviation amplitude of a single abnormal point is not particularly large, the overall trend is still affected, that is, the absolute value of the slope of the calculated line segment is larger, so that the degradation trend near the early fault point can be increased.
The fault analysis device based on the l1 trend filtering provided by the embodiment of the invention is described below with reference to the accompanying drawings.
Referring to fig. 7, based on the same inventive concept, an embodiment of the present invention provides a fault analysis device based on l1 trend filtering, the device includes: a determining unit for determining a first spectral distance dataset of the centrifugal compressor in a normal operating state; a data acquisition unit for acquiring a second spectral distance dataset in a real-time state of the centrifugal compressor; a sliding data creation unit for creating a sliding combination window based on the first spectral distance data set and the second spectral distance data set; a parameter calculation unit for determining regularization parameters based on the sliding combination window; the fault analysis unit is used for processing the sliding combination window based on the l1 trend filtering rule and the regularization parameter to obtain a fault trend characterization factor for the centrifugal compressor, and the fault trend characterization factor is used for carrying out predictive analysis on faults of the centrifugal compressor.
In an embodiment of the present invention, the determining unit includes: the historical data acquisition module is used for acquiring historical operation data of the full life cycle of the centrifugal compressor; a first dataset determination module for computationally determining a historical spectral distance dataset based on the historical operating data; the decomposition module is used for performing time continuous decomposition operation on the historical spectrum distance data set according to a preset length to obtain a plurality of corresponding spectrum distance arrays; a trend factor determining module for determining a degradation trend factor based on the plurality of spectral distance arrays; a second data set determining module for determining a first spectral distance data set in the plurality of spectral distance arrays based on the degradation trend factor.
In an embodiment of the present invention, the first data set determining module is specifically configured to: calculating corresponding spectral distance data R (J) xy ) The spectral distance data R (J xy ) Characterized by:wherein R (J) xy )∈(0～1)，J xy Characterized by the J-divergence between the operating normal state signal x (t) and the state signal to be evaluated y (t), αe (0-1) characterized by a sensitivity coefficient, wherein: />Wherein S is x (k) And S is y (k) The self-power spectrums are respectively characterized as signals x (t) and y (t), and P is the number of power spectrum lines; at the spectral distance data R (J xy ) And extracting a historical spectrum distance data set of the centrifugal compressor in a normal running state.
In an embodiment of the present invention, the trend factor determining module is specifically configured to: determining an initial regularization parameter value based on a preset data range; performing preliminary l1 trend filtering operation on each spectrum distance array in sequence based on the initial regularization parameter values to obtain corresponding filtered curves; calculating and determining the average value of the upper derivative of each filtered curve; and determining the minimum value with the minimum absolute value in the upper derivative average value, and taking the minimum value as the degradation trend factor.
In an embodiment of the present invention, the parameter calculation unit is specifically configured to: s41) determining an initial value of the regularization parameter; s42) performing derivative operation on the sliding combination window based on the initial value to obtain a corresponding first derivative value; s43) performing self-increment operation on the initial value based on a preset self-increment rule to obtain corresponding self-increment value; s44) performing derivative operation on the sliding combination window based on the self-increment value to obtain a corresponding second derivative value; s45) judging whether the deviation between the first derivative value and the second derivative value is smaller than a preset deviation threshold value or not; s461) if yes, determining the self-increment value as the regularization parameter; s462) if not, jumping to step S43) taking the self-increment value as a new initial value.
In an embodiment of the present invention, the fault analysis unit includes: the data processing module is used for processing the sliding combination window based on the l1 trend filtering rule and the regularization parameter to obtain a corresponding processed line segment; the slope obtaining module is used for obtaining slope information of the processed line segment; and the fault factor determining module is used for taking the absolute value of the slope information as the fault trend characterization factor.
In an embodiment of the present invention, the apparatus further includes an early warning unit, where the early warning unit is specifically configured to: after the fault trend characterization factor is obtained, judging whether the fault trend characterization factor is larger than a preset slope threshold value or not; and outputting corresponding early-stage fault early-warning information under the condition that the fault trend characterization factor is larger than the preset slope threshold value.
Further, the embodiment of the present invention also provides a computer readable storage medium, on which a computer program is stored, which when executed by a processor, implements the method described in the embodiment of the present invention.
The foregoing details of the optional implementation of the embodiment of the present invention have been described in detail with reference to the accompanying drawings, but the embodiment of the present invention is not limited to the specific details of the foregoing implementation, and various simple modifications may be made to the technical solution of the embodiment of the present invention within the scope of the technical concept of the embodiment of the present invention, and these simple modifications all fall within the protection scope of the embodiment of the present invention.
In addition, the specific features described in the above embodiments may be combined in any suitable manner without contradiction. In order to avoid unnecessary repetition, various possible combinations of embodiments of the present invention are not described in detail.
Those skilled in the art will appreciate that all or part of the steps in implementing the methods of the embodiments described above may be implemented by a program stored in a storage medium, including instructions for causing a single-chip microcomputer, chip or processor (processor) to perform all or part of the steps of the methods described in the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
In addition, any combination of various embodiments of the present invention may be performed, so long as the concept of the embodiments of the present invention is not violated, and the disclosure of the embodiments of the present invention should also be considered.
Claims (15)
1. A fault analysis method based on l1 trend filtering, the method comprising:
determining a first spectral distance dataset of the centrifugal compressor in a normal operating state;
acquiring a second spectral distance dataset of the centrifugal compressor in a real-time state;
creating a sliding combination window based on the first spectral distance dataset and the second spectral distance dataset;
determining a regularization parameter based on the sliding combination window;
and processing the sliding combination window based on the l1 trend filtering rule and the regularization parameter to obtain a fault trend characterization factor for the centrifugal compressor, wherein the fault trend characterization factor is used for carrying out predictive analysis on faults of the centrifugal compressor.
2. The method of claim 1, wherein said determining a first spectral distance dataset for a centrifugal compressor under normal operating conditions comprises:
acquiring historical operation data of a full life cycle of the centrifugal compressor;
determining a historical spectral distance dataset based on the historical operating data calculation;
performing time continuous decomposition operation on the historical spectrum distance data set according to a preset length to obtain a plurality of corresponding spectrum distance arrays;
determining a degradation trend factor based on the plurality of spectral distance arrays;
a first spectral distance dataset is determined in the plurality of spectral distance arrays based on the degradation trend factor.
3. The method of claim 2, wherein said computing a determined historical spectral distance dataset based on said historical operating data comprises:
calculating corresponding spectral distance data R (J) xy ) The spectral distance data R (J xy ) Characterized by:
wherein R (J) xy )∈(0～1)，J xy Characterized by the J-divergence between the operating normal state signal x (t) and the state signal to be evaluated y (t), αe (0-1) characterized by a sensitivity coefficient, wherein:
wherein S is x (k) And S is y (k) The self-power spectrums are respectively characterized as signals x (t) and y (t), and P is the number of power spectrum lines;
at the spectral distance data R (J xy ) And extracting a historical spectrum distance data set of the centrifugal compressor in a normal running state.
4. The method of claim 2, wherein the determining a degradation trend factor based on the plurality of spectral distance arrays comprises:
determining an initial regularization parameter value based on a preset data range;
performing preliminary l1 trend filtering operation on each spectrum distance array in sequence based on the initial regularization parameter values to obtain corresponding filtered curves;
calculating and determining the average value of the upper derivative of each filtered curve;
and determining the upper derivative average value with the smallest absolute value as the minimum value, and taking the minimum value as the degradation trend factor.
5. The method of claim 1, wherein the determining a regularization parameter based on the sliding combination window comprises:
s41) determining an initial value of the regularization parameter;
s42) performing derivative operation on the sliding combination window based on the initial value to obtain a corresponding first derivative value;
s43) performing self-increment operation on the initial value based on a preset self-increment rule to obtain corresponding self-increment value;
s44) performing derivative operation on the sliding combination window based on the self-increment value to obtain a corresponding second derivative value;
s45) judging whether the deviation between the first derivative value and the second derivative value is smaller than a preset deviation threshold value or not;
s461) if yes, determining the self-increment value as the regularization parameter;
s462) if not, jumping to step S43) taking the self-increment value as a new initial value.
6. The method of claim 1, wherein the processing the sliding combination window based on the l1 trend filtering rule and the regularization parameter to obtain a fault trend characterization factor for the centrifugal compressor comprises:
processing the sliding combination window based on the l1 trend filtering rule and the regularization parameter to obtain a corresponding processed line segment;
acquiring slope information of the processed line segment;
and taking the absolute value of the slope information as the fault trend characterization factor.
7. The method of claim 6, wherein the method further comprises:
after the fault trend characterization factor is obtained, judging whether the fault trend characterization factor is larger than a preset slope threshold value or not;
and outputting corresponding early-stage fault early-warning information under the condition that the fault trend characterization factor is larger than the preset slope threshold value.
8. A fault analysis device based on l1 trend filtering, the device comprising:
a determining unit for determining a first spectral distance dataset of the centrifugal compressor in a normal operating state;
a data acquisition unit for acquiring a second spectral distance dataset in a real-time state of the centrifugal compressor;
a sliding data creation unit for creating a sliding combination window based on the first spectral distance data set and the second spectral distance data set;
a parameter calculation unit for determining regularization parameters based on the sliding combination window;
the fault analysis unit is used for processing the sliding combination window based on the l1 trend filtering rule and the regularization parameter to obtain a fault trend characterization factor for the centrifugal compressor, and the fault trend characterization factor is used for carrying out predictive analysis on faults of the centrifugal compressor.
9. The apparatus according to claim 8, wherein the determining unit comprises:
the historical data acquisition module is used for acquiring historical operation data of the full life cycle of the centrifugal compressor;
a first dataset determination module for computationally determining a historical spectral distance dataset based on the historical operating data;
the decomposition module is used for performing time continuous decomposition operation on the historical spectrum distance data set according to a preset length to obtain a plurality of corresponding spectrum distance arrays;
a trend factor determining module for determining a degradation trend factor based on the plurality of spectral distance arrays;
a second data set determining module for determining a first spectral distance data set in the plurality of spectral distance arrays based on the degradation trend factor.
10. The apparatus of claim 9, wherein the first data set determining module is specifically configured to:
calculating corresponding spectral distance data R (J) xy ) The spectral distance data R (J xy ) Characterized by:
wherein R (J) xy )∈(0～1)，J xy Characterized by the J-divergence between the operating normal state signal x (t) and the state signal to be evaluated y (t), αe (0-1) characterized by a sensitivity coefficient, wherein:
wherein S is x (k) And S is y (k) The self-power spectrums are respectively characterized as signals x (t) and y (t), and P is the number of power spectrum lines;
at the spectral distance data R (J xy ) And extracting a historical spectrum distance data set of the centrifugal compressor in a normal running state.
11. The apparatus of claim 9, wherein the trend factor determination module is specifically configured to:
determining an initial regularization parameter value based on a preset data range;
performing preliminary l1 trend filtering operation on each spectrum distance array in sequence based on the initial regularization parameter values to obtain corresponding filtered curves;
calculating and determining the average value of the upper derivative of each filtered curve;
and determining the upper derivative average value with the smallest absolute value as the minimum value, and taking the minimum value as the degradation trend factor.
12. The apparatus according to claim 8, wherein the parameter calculation unit is specifically configured to:
s41) determining an initial value of the regularization parameter;
s42) performing derivative operation on the sliding combination window based on the initial value to obtain a corresponding first derivative value;
s43) performing self-increment operation on the initial value based on a preset self-increment rule to obtain corresponding self-increment value;
s44) performing derivative operation on the sliding combination window based on the self-increment value to obtain a corresponding second derivative value;
s45) judging whether the deviation between the first derivative value and the second derivative value is smaller than a preset deviation threshold value or not;
s461) if yes, determining the self-increment value as the regularization parameter;
s462) if not, jumping to step S43) taking the self-increment value as a new initial value.
13. The apparatus of claim 8, wherein the fault analysis unit comprises:
the data processing module is used for processing the sliding combination window based on the l1 trend filtering rule and the regularization parameter to obtain a corresponding processed line segment;
the slope obtaining module is used for obtaining slope information of the processed line segment;
and the fault factor determining module is used for taking the absolute value of the slope information as the fault trend characterization factor.
14. The apparatus of claim 13, further comprising an early warning unit, the early warning unit being specifically configured to:
after the fault trend characterization factor is obtained, judging whether the fault trend characterization factor is larger than a preset slope threshold value or not;
and outputting corresponding early-stage fault early-warning information under the condition that the fault trend characterization factor is larger than the preset slope threshold value.
15. A computer readable storage medium, on which a computer program is stored, characterized in that the program, when being executed by a processor, implements the method of any of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210843196.5A CN117473288A (en) | 2022-07-18 | 2022-07-18 | Fault analysis method, analysis device and storage medium based on l1 trend filtering |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210843196.5A CN117473288A (en) | 2022-07-18 | 2022-07-18 | Fault analysis method, analysis device and storage medium based on l1 trend filtering |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117473288A true CN117473288A (en) | 2024-01-30 |
Family
ID=89633467
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210843196.5A Pending CN117473288A (en) | 2022-07-18 | 2022-07-18 | Fault analysis method, analysis device and storage medium based on l1 trend filtering |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117473288A (en) |
-
2022
- 2022-07-18 CN CN202210843196.5A patent/CN117473288A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111262750B (en) | Method and system for evaluating baseline model | |
| KR102067344B1 (en) | Apparatus and Method for Detecting Abnormal Vibration Data | |
| CN113579851B (en) | Non-stationary drilling process monitoring method based on adaptive segmented PCA | |
| CN116308300B (en) | Power equipment state monitoring evaluation and command method and system | |
| CN116739829B (en) | Big data-based power data analysis method, system and medium | |
| CN116086537A (en) | Equipment state monitoring method, device, equipment and storage medium | |
| CN111176226A (en) | Automatic analysis method for alarm threshold of equipment characteristic parameter based on operation condition | |
| CN113987027A (en) | Equipment state determination method, device and medium based on fault prediction | |
| CN111145895A (en) | Abnormal data detection method and terminal equipment | |
| CN117318297A (en) | Alarm threshold setting method, system, equipment and medium based on state monitoring | |
| CN116401137B (en) | Core particle health state prediction method and device, electronic equipment and storage medium | |
| CN117473288A (en) | Fault analysis method, analysis device and storage medium based on l1 trend filtering | |
| CN112308391A (en) | Real-time monitoring and anomaly detection method for equipment state based on neural network | |
| CN116361695A (en) | Abnormal data detection method and device | |
| CN115221769B (en) | Fault prediction method, system, electronic equipment and storage medium | |
| CN112816898B (en) | Battery failure prediction method and device, electronic equipment and storage medium | |
| CN113283157A (en) | System, method, terminal and medium for predicting life cycle of intelligent stamping press part | |
| CN114322446A (en) | Cooling system fault early warning method and device, cooling system and operation machine | |
| CN117172431B (en) | Food processing apparatus and equipment management method and system | |
| CN116705271B (en) | Medical flushing equipment operation monitoring system based on big data | |
| CN117577137B (en) | Cutter health assessment method, device, equipment and storage medium | |
| CN117038048B (en) | Remote fault processing method and system for medical instrument | |
| CN114548177A (en) | Equipment health assessment method, system and equipment | |
| CN118037090A (en) | Equipment situation analysis method, device, equipment and medium based on index evaluation | |
| CN116067688A (en) | Health degree prediction method and system and engineering machinery |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |