CN112307619A - Construction method of early warning model, and equipment fault early warning method and device - Google Patents
Construction method of early warning model, and equipment fault early warning method and device Download PDFInfo
- Publication number
- CN112307619A CN112307619A CN202011182775.7A CN202011182775A CN112307619A CN 112307619 A CN112307619 A CN 112307619A CN 202011182775 A CN202011182775 A CN 202011182775A CN 112307619 A CN112307619 A CN 112307619A
- Authority
- CN
- China
- Prior art keywords
- correlation
- monitoring data
- early warning
- fault
- change
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 53
- 238000010276 construction Methods 0.000 title claims abstract description 14
- 238000012544 monitoring process Methods 0.000 claims abstract description 151
- 230000008859 change Effects 0.000 claims abstract description 128
- 238000001228 spectrum Methods 0.000 claims description 28
- 239000011159 matrix material Substances 0.000 claims description 24
- 230000002159 abnormal effect Effects 0.000 claims description 22
- 238000004364 calculation method Methods 0.000 claims description 9
- 238000001514 detection method Methods 0.000 claims description 3
- 238000011156 evaluation Methods 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 3
- 238000012216 screening Methods 0.000 claims description 2
- 230000001052 transient effect Effects 0.000 description 8
- 238000001914 filtration Methods 0.000 description 7
- 230000008569 process Effects 0.000 description 7
- 230000006870 function Effects 0.000 description 6
- 230000036541 health Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 5
- 238000005070 sampling Methods 0.000 description 5
- 238000004422 calculation algorithm Methods 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 230000008901 benefit Effects 0.000 description 3
- 238000007405 data analysis Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 230000015556 catabolic process Effects 0.000 description 2
- 238000004590 computer program Methods 0.000 description 2
- 230000003247 decreasing effect Effects 0.000 description 2
- 238000006731 degradation reaction Methods 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 208000024891 symptom Diseases 0.000 description 2
- 241000668842 Lepidosaphes gloverii Species 0.000 description 1
- 102100029469 WD repeat and HMG-box DNA-binding protein 1 Human genes 0.000 description 1
- 101710097421 WD repeat and HMG-box DNA-binding protein 1 Proteins 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000017105 transposition Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/10—Noise analysis or noise optimisation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The application relates to a construction method of an early warning model, and an equipment fault early warning method and device, wherein the construction method of the early warning model comprises the following steps: acquiring monitoring data samples of at least two linkage parts in equipment; calculating correlation change parameters among monitoring data samples corresponding to different linkage parts, wherein the correlation change parameters are used for representing time sequence-based fault change trends; and constructing an early warning model according to the monitoring data samples and the related change parameters. The correlation is the relative variation between the monitoring data and is irrelevant to the monitoring data of single equipment, and the correlation variation parameter can represent the equipment fault variation trend based on time sequence, so that the correlation variation parameter can be used as a parameter for judging equipment abnormity and abnormity degree, a fault early warning model is constructed based on the correlation variation parameter, the fault early warning precision can be improved, and the influence of insufficient knowledge of a single signal and external interference on fault early warning is reduced.
Description
Technical Field
The application relates to the field of equipment health management, in particular to an equipment fault early warning model construction method, an early warning method and an early warning device.
Background
The Health management algorithm model is a core technology for realizing fault prediction and Health management of complex equipment PHM (probabilistic and Health management), and the fault prediction is the most critical technology in the field of Health management. On the other hand, due to the fact that some equipment are complex in structure, multiple in working conditions and frequent in switching, and external interference is added, the fault prediction difficulty is large, and a single data analysis method is difficult to meet practical requirements in terms of fault prediction accuracy. The two aspects result in that the fault prediction for the complex equipment is important and challenging, and the research on the application of the fault prediction technology in the field of equipment health management has high potential and economic benefit.
Therefore, how to accurately predict the equipment failure becomes a technical problem to be solved urgently.
Disclosure of Invention
In order to solve the technical problem or at least partially solve the technical problem, the present application provides a method for constructing an early warning model, including: acquiring monitoring data samples of at least two linkage parts in equipment; calculating correlation change parameters among monitoring data samples corresponding to different linkage parts, wherein the correlation change parameters are used for representing time sequence-based fault change trends; and constructing an early warning model according to the monitoring data samples and the related change parameters.
Optionally, calculating a correlation variation parameter between monitoring data samples of different monitored components comprises: calculating a plurality of correlation coefficients between monitoring data samples of different components corresponding to a plurality of unit times; constructing a correlation coefficient matrix in a preset time period based on the plurality of correlation coefficients; and obtaining a correlation change parameter by using the correlation coefficient matrix.
Optionally, the obtaining of the correlation change parameter by using the correlation coefficient matrix includes: constructing a time-sequence-based correlation coefficient change curve of the correlation coefficient based on the correlation coefficient matrix; and calculating the envelope spectrum of the correlation coefficient change curve to obtain the correlation change parameter.
Optionally, calculating an envelope spectrum of the correlation coefficient variation curve to obtain the correlation variation parameter includes: constructing a Hilbert envelope spectrum of a correlation coefficient change curve; and screening the Hilbert envelope spectrum based on the preset fault variation trend to obtain an envelope line meeting the preset fault variation trend.
Optionally, before or after the hilbert envelope spectrum is screened based on the preset fault variation trend, the method for constructing the early warning model further includes: and adjusting the hyper-parameters of the envelope spectrum according to a preset target.
Optionally, constructing a computational mapping from the monitoring data samples to the correlation parameters to obtain the early warning model includes: extracting envelope characteristics; and determining the evaluation standard of the fault early warning model according to the envelope characteristics.
Optionally, after calculating the correlation variation parameter between the monitoring data samples corresponding to different monitored components, the method further includes: fitting the data in each data neighborhood in the monitoring data sample by using an n-order polynomial to obtain a denoised monitoring data sample, wherein the coefficient of the n-order polynomial is determined by a least square method criterion under the condition of the minimum fitting error, and n is an integer greater than or equal to 1.
In a second aspect, the present application provides an equipment fault early warning method, including: acquiring monitoring data of at least two linkage parts in equipment; the monitoring data are input into a fault early warning model to obtain an abnormal monitoring result, the fault early warning model is obtained by building based on monitored related change parameters between monitoring data samples corresponding to different linkage parts, the related change parameters are used for representing a time sequence-based fault change trend, and the monitoring data samples are obtained by monitoring at least two linkage parts of equipment.
Optionally, before inputting the monitoring data into the fault pre-warning model, the method further includes: fitting the data in each data neighborhood in the monitoring data by utilizing an n-order polynomial to obtain a denoised monitoring data sample, wherein the coefficient of the n-order polynomial is determined by a least square method criterion under the condition of the minimum fitting error, and n is an integer greater than or equal to 1.
In a third aspect, the present application provides an early warning model building apparatus, including: the device comprises a first acquisition module, a second acquisition module and a third acquisition module, wherein the first acquisition module is used for acquiring monitoring data samples of at least two linkage parts in the device; the calculation module is used for calculating correlation change parameters among monitoring data samples corresponding to different linkage parts, and the correlation change parameters are used for representing time sequence-based fault change trends; and the model construction module is used for constructing an early warning model according to the monitoring data samples and the related change parameters.
In a fourth aspect, the present application provides an equipment fault early warning device, including: the second acquisition module is used for acquiring at least two monitoring data corresponding to at least two linkage parts in the equipment; the detection module is used for inputting monitoring data into the fault early warning model to obtain an abnormal monitoring result, the abnormal monitoring model is obtained by constructing related change parameters between monitoring data samples corresponding to different linkage parts, the related change parameters are used for representing a time sequence-based fault change trend, and the monitoring data samples are obtained by monitoring at least two linkage parts of the equipment.
In a fifth aspect, the present application provides a computer-readable storage medium, which stores computer instructions for causing a computer to execute the method for constructing an early warning model according to any one of the above first aspects and/or execute the method for early warning equipment failure as described in the second aspect.
In a sixth aspect, the present application provides an electronic device, comprising: at least one processor; and a memory communicatively coupled to the at least one processor; the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor, so that the at least one processor executes the warning model construction method of any one of the above first aspects and/or executes the equipment fault warning method described in the above second aspect.
Compared with the prior art, the technical scheme provided by the embodiment of the application has the following advantages:
the early warning model construction method provided by the embodiment of the application calculates the correlation of the monitoring data samples corresponding to at least two linkage parts and the correlation change parameters thereof, because the correlation of the monitoring data samples is related to the linkage state of the linkage parts, because the correlation of the samples is the relative change quantity between the monitoring data and is unrelated to the monitoring data of a single device, the worse the linkage state is, the lower the correlation is, the linkage state is poor or even invalid, the working condition instruction of the device cannot be effectively executed, the device has faults, and the correlation change parameters can represent the equipment fault change trend based on time sequence, therefore, the correlation change parameters can be used as the parameters for judging the equipment abnormity and abnormity degree, and the fault early warning model is constructed based on the correlation change parameters of the time sequence between the monitoring data samples corresponding to different linkage parts, the precision of early warning the fault can be improved, and the influence of insufficient knowledge of a single signal and external interference on fault early warning is reduced.
The equipment fault early warning method provided by the embodiment of the application obtains early warning results by acquiring monitoring data corresponding to at least two linkage parts in equipment and inputting the monitoring data into a fault early warning model, wherein the early warning model is constructed by using correlation change parameters between monitoring data samples corresponding to different mutual linkage parts, because the correlation of the monitoring data is related to the linkage state of the linkage parts, the correlation of the samples is the relative change quantity between the monitoring data and is unrelated to the monitoring data of single equipment, the worse the linkage state is, the lower the correlation is, the linkage state is worse or even invalid, the working condition instruction of the equipment cannot be effectively executed, the equipment is in fault, and the correlation change parameters can represent the equipment fault change trend based on time sequence, so the correlation change parameters can be used as parameters for judging the abnormal degree and abnormal degree of the equipment, therefore, the model is used for carrying out fault early warning on newly acquired monitoring data, the precision of early warning on faults can be improved, and the influence of insufficient knowledge of a single signal and external interference on the fault early warning is reduced.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and together with the description, serve to explain the principles of the invention.
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without inventive exercise.
Fig. 1 is a schematic flow chart of a method for constructing an early warning model according to an embodiment of the present disclosure;
fig. 2 is a schematic flow chart of an equipment fault early warning method provided in an embodiment of the present application;
fig. 3 is a schematic flowchart of calculating a correlation variation parameter according to an embodiment of the present application;
4a-4c are graphs of the correlation change trend between the monitoring data provided by the embodiment of the present application;
FIG. 5 is a graph of a trend of a correlation coefficient provided in an embodiment of the present application;
fig. 6 is a schematic flowchart of a correlation coefficient envelope spectrum calculation according to an embodiment of the present application;
FIG. 7 is a diagram illustrating an envelope spectrum of a correlation coefficient provided in an embodiment of the present application;
FIG. 8 is a schematic diagram of an envelope spectrum after parameter adjustment according to an embodiment of the present application;
fig. 9 is a schematic diagram of a device for constructing an early warning model according to an embodiment of the present disclosure;
fig. 10 is a schematic diagram of equipment fault warning provided in an embodiment of the present application;
fig. 11 is a schematic view of an electronic device according to an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
As described in the background art, currently, a single data analysis method is usually adopted for the early warning of equipment faults, for example, a large amount of monitoring data of a single component, such as current data or rotation speed of a single flywheel, or other monitoring data capable of representing performance/faults of the flywheel, is collected, and an artificial intelligence model is trained by using the monitoring data, so as to predict the faults. The research of the invention finds that there are many reasons for the failure early warning effect which is difficult to achieve by adopting a single data analysis method, and the reasons are explained by taking a spacecraft as an example as follows:
1. the spacecraft equipment has insufficient fault data, generally 99% of monitoring data belong to normal data, and the fault prediction effect is poor under the condition of lacking fault data samples;
2. the working condition switching is complex, so that signals (current, vibration, temperature and the like) acquired by the spacecraft have corresponding changes along with the working condition, the changes can not be distinguished from abnormal signals, and the effect of a fault prediction algorithm is interfered.
3. The interference of outer space signals is large, and a plurality of noise signals and real abnormity are difficult to distinguish, so that effective fault characteristics cannot be extracted.
4. The air-to-ground bandwidth resources are limited, and sufficient monitoring points cannot be received, so that the data dimensionality is insufficient, and the fine fault positioning cannot be realized.
The inventor researches and discovers that when equipment fails, for example, component failure or assembly failure occurs, the phenomenon which can be shown by the equipment is not necessarily the fault phenomenon of a single current component or assembly, deviation or linkage failure can occur in linkage effect of other components which are linked with the equipment in the linkage process, so that the equipment can show the fault. When the linkage has deviation or linkage failure, the working condition instruction cannot be effectively executed, and the spacecraft can be indicated to have obvious faults. Based on the method and the device, the relevance of different parts in the linkage process of the linkage parts of the equipment is used, the change of the relevance of the corresponding collected features in the whole process is analyzed, and the abnormal degree of the relevance change is used as a standard for measuring whether the equipment is abnormal in operation or not.
Referring to fig. 1, the present application provides an equipment fault early warning method, which specifically includes the following steps:
s11, acquiring monitoring data of at least two linkage parts in the equipment. In the present embodiment, the linkage components may be functionally cooperating components, linkage components, functionally related components, and the like, the component referred to in this embodiment is a component on the equipment for implementing a certain function or functions, and specifically, the equipment failure may be divided into a transient failure and a non-transient failure, for example, a failure of an electronic component belongs to the transient failure, and faults of some mechanical parts, electromechanical systems and the like can be non-transient faults, and because transient fault failure has few symptoms, prediction is difficult, prediction of faults in this embodiment focuses on prediction of non-transient faults, and of course, those skilled in the art will appreciate that, prediction of transient faults for components with symptoms, such as reduced performance before the transient fault occurs, is also within the scope of the present application. Taking a spacecraft as an example, the collection of monitoring data for linkage components may be data that produces non-transitory faulty components.
For example, for a spacecraft control system, when an attitude angle of a spacecraft needs to be converted, different working condition codes can be reached, flywheels of a spacecraft X, Y, Z shaft can adjust the position of the spacecraft to a position meeting requirements through cooperation, and linkage of the flywheels of X, Y, Z shafts is required to realize the function, so that data of three flywheels of X, Y, Z shafts, such as attitude angle, angular rate, temperature, rotating speed, current and the like, can be acquired. Therefore, in the embodiment, the monitoring data of at least two linkage parts can be collected to represent whether the current equipment is abnormal or not and the abnormal degree based on the correlation of the monitoring data.
And S12, inputting the monitoring data into a fault early warning model to obtain an abnormal monitoring result. As an exemplary embodiment, the input of the fault warning model is monitoring data of the linkage part, and the output is the fault degree of the equipment. For example, current monitoring data may be input for a period of time, and the fault early warning model may calculate a time-series-based fault variation trend of the equipment based on a change in correlation of the input current data, and finally output the fault degree of the equipment, for example, may output a score representing the fault degree, a prompt message, or a degradation trend curve or a fault degree variation trend curve of the equipment as an output result.
As an exemplary embodiment, the fault early warning model is constructed based on monitoring of relevant change parameters between monitoring data samples corresponding to different linkage components, the relevant change parameters are used for representing a time sequence-based fault change trend, and the monitoring data samples are obtained based on monitoring of at least two linkage components in the equipment. Illustratively, the fault early warning model performs secondary conversion on the characteristics of the monitoring data to obtain the correlation of the monitoring data, and obtains a correlation change parameter capable of representing whether the current equipment is abnormal or not and the abnormal degree based on the correlation, wherein the correlation change parameter may be a correlation change curve, or may be data or a change result capable of representing the correlation change of the monitoring data corresponding to different components, such as a plurality of groups of values of the correlation change. The related change parameters are related to the fault change trend of the equipment, namely the worse the linkage state of the linkage part, the higher the fault degree of the equipment, the lower the correlation of the corresponding monitoring data, the worse or even invalid the linkage state, the failure condition instruction of the equipment can not be effectively executed, and the equipment has a fault, at the moment, the correlation of the monitoring data is in a descending trend along with the trend, and the trend is represented by the related change parameters, so that the fault degree of the equipment can be quantitatively expressed. Therefore, the model is used for carrying out fault early warning on newly acquired monitoring data, the monitoring data change caused by the change of the single data acquisition amount, the air-ground bandwidth resources and the working condition is irrelevant, the equipment fault can be predicted only by the relative change amount of a small amount of data (the trend of the overall change of the data correlation), the fault early warning precision can be improved, and the influences of insufficient knowledge of a single signal, the self working condition of the equipment and external interference on the fault early warning are reduced.
As an exemplary embodiment, since there is signal interference outside, for example, there is more signal interference outside space, and the interference is not fixed, it is necessary to perform denoising on the monitoring data to obtain a more accurate early warning result, and before inputting the monitoring data into the fault early warning model, it is necessary to perform outlier elimination on the on-orbit operation monitoring data, remove a current value that is obviously noise, and improve data quality. For example, the data may be filtered by using an s-g filtering algorithm, specifically, the data in each data neighborhood in the monitored data is fitted by using an nth-order polynomial, so as to obtain a denoised monitored data sample, and coefficients of the nth-order polynomial are determined by a least square method criterion under a condition that a fitting error is minimum, where n is an integer greater than or equal to 1. As will be described in detail below with respect to the filtering algorithm, a set of data x (i) (i ═ m.,. m) is set up, and now n-th order polynomial fit data is constructed:
and filtering a total of 2n +1 observed values before and after the current moment, and fitting the observed values by using a k-1 order polynomial. For the observed value at the current time, the following formula can be used for fitting:
xt＝a0+a1·t+a2·t2+...+ak-1·tk-1
similarly, the previous and subsequent time (e.g., t-1, t +1, t-2, t +3, etc.) can be calculated by the above formula, so that a total of 2n +1 formulas can be obtained to form a matrix:
to make the entire matrix have a solution, 2n +1 must be satisfied>k, parameter a can be determined by the least square method according to the above formula0,a1,a2,...,ak-1. The above matrix is simplified to the following formula:
X(2n+1)×1＝T(2n+1)×k·Ak×1+E(2n+1)×1
wherein T represents [ T-n, T + n [ ]]Coordinates of observations at time of day, subscripts of individual parameters indicating their respective dimensions, e.g. Ak×1Parameters with k rows and 1 columns are shown. By the least square method, A can be obtainedk×1The solution of (a) is:
A＝(Ttrans·T)-1·Ttrans·X
wherein, superscript trans represents transposition, and the filtering value of the model is:
P＝T·A＝T·(Ttrans·T)-1·Ttrans·X＝B·X
finally, a relation matrix between the filtering value and the observed value can be obtained:
B＝T·(Ttrans·T)-1·Ttrans，
and then can be fast according to the observed value and obtain the filtering value. Wherein a is to-be-evaluated and is expressed as a coefficient of each order in a fitting formula; k represents the order of the fitting formula; t represents the time point identification of the data.
And the method of denoising and smoothing filtering is used for eliminating monitoring data which are obviously noise, improving the data quality and increasing the accuracy of the output result of the model.
The invention provides a construction method of an early warning model, which comprises the following steps as shown in figure 2:
and S21, acquiring monitoring data samples of at least two linkage parts in the equipment. As an exemplary embodiment, for the collection of the monitoring data samples, reference may be made to the description of step S11 in the above embodiment, in this embodiment, data of the samples of the monitoring data may be collected at a preset sampling frequency within a period of time, for example, data of the monitoring data within a period of one year, two years, or longer or shorter may be collected at a sampling frequency of one day, two days, or longer or shorter, for example, data of the monitoring data within a period of one year, two years, or longer or shorter may be collected, for example, a flywheel failure prediction of a spacecraft control system from a year to a +1 year is taken as an example, sampling is performed at a certain sampling frequency, for example, in a unit of one day, and data of the monitoring signals of the flywheel operating in orbit from a year to a +1 year of a. Wherein, table 1 gives exemplary monitoring data:
in this embodiment, a current may be taken as an example for illustration, and of course, it should be understood by those skilled in the art that other data may be modeled as monitoring data.
And S22, calculating correlation change parameters among monitoring data samples corresponding to different linkage parts, wherein the correlation change parameters are used for representing time sequence-based fault change trends. In this embodiment, the correlation change parameter may describe a trend of correlation change between the monitored data samples corresponding to different linkage components, for example, the correlation change parameter may be a curve describing the trend of correlation change of the monitored data samples, or may also be a parameter describing data or a result of change that can represent the correlation change of the monitored data samples corresponding to different components, such as a plurality of sets of numerical values of correlation change. The correlation change parameters are related to the fault change trend of the equipment, namely the worse the linkage state of the linkage component, the higher the fault degree of the equipment, the lower the correlation of the corresponding monitoring data, the worse or even invalid the linkage state, the working condition instruction of the equipment can not be effectively executed, and the equipment is in fault. For the correlation between the monitoring samples, a covariance matrix, a Pearson correlation coefficient, a Spearman correlation coefficient, etc. can be used for calculation, wherein the correlation coefficient value for the correlation is between [ -1,1], 0 represents complete irrelevance, 1 represents complete positive correlation, and-1 represents complete negative correlation. After obtaining the correlation, the correlations of all sampled monitoring data samples in the preset time period may be listed, for example, the correlation coefficients between the monitoring data samples of different linkage components each day may be calculated by taking the day as a unit, and a correlation coefficient matrix is constructed for listing, or the correlation coefficients may be listed in a distribution diagram manner, so as to obtain a correlation change trend, and the correlation change trend is described by a correlation variation parameter, for example, a correlation change function may be calculated by the correlation coefficient matrix, or a correlation change curve may be drawn according to the correlation coefficients each day in the graph, so as to obtain the correlation change parameter.
And S23, constructing an early warning model according to the monitoring data samples and the related change parameters. As an exemplary embodiment, a fault evaluation threshold value is constructed based on the relevant change parameters and the known equipment fault change trend, an abnormal degree judgment standard is formed, a calculation mapping between the monitoring data sample and the relevant change parameters is established, the monitoring data sample is used as input, and the abnormal degree grading result is used as output to construct a fault early warning model.
By calculating the correlation and the correlation change parameters of the monitoring data samples corresponding to at least two linkage parts, because the correlation of the monitoring data samples is related to the linkage state of the linkage parts, because the correlation of the samples is the relative variation between the monitoring data and is unrelated to the monitoring data of a single device, the linkage state is worse, the correlation is lower, the linkage state is poor or even fails, the working condition instruction of the device cannot be effectively executed, the device has faults, and the correlation change parameters can represent the fault change trend of the device based on time sequence, therefore, the correlation change parameters can be used as the parameters for judging the abnormity and the abnormity degree of the device, therefore, a fault early warning model is constructed based on the correlation between the monitoring data samples corresponding to different linkage parts and based on the correlation change parameters of the time sequence, and the early warning precision of the faults can be improved, the influence of insufficient knowledge of a single signal and external interference on fault early warning is reduced.
As an exemplary embodiment, in calculating the correlation variation parameter, the correlation between the monitoring data samples may be calculated, and the correlation variation parameter may be obtained by using a time-series-based variation trend of the correlations of the plurality of samples, and for example, as shown in fig. 3, calculating the correlation variation parameter may include the following steps:
and S221, calculating a plurality of correlation coefficients among the monitoring data samples of different parts corresponding to a plurality of unit times. For example, pearson correlation coefficients between daily monitoring data samples may be calculated in units of data acquisition cycles, such as days. Specifically, the calculation principle of the correlation coefficient may be:
let (X, Y) be two monitoring data samples, if covariance cov (X, Y) of X and Y exists, and variances DX and DY satisfy DX > 0 and DY > 0, then
I.e. the correlation coefficient of X and Y. Namely:
wherein R isXYFor the correlation coefficient of the monitored data sample X and the monitored data sample Y, DX、DYIs the variance, EX、EYIs a mathematical expectation.
As an exemplary embodiment, taking three flywheels of the spacecraft control system X, Y, Z as an example, a correlation coefficient between the X-axis flywheel and the Y-axis flywheel monitoring data sample, a correlation coefficient between the X-axis flywheel and the Z-axis flywheel monitoring data sample, and a correlation coefficient between the Y-axis flywheel and the Z-axis flywheel monitoring data sample are calculated respectively.
S222, constructing a correlation coefficient matrix in a preset time period based on the plurality of correlation coefficients. And counting the daily correlation coefficient of the whole period of the whole preset time period, establishing a correlation coefficient matrix by using a plurality of correlation coefficients, and respectively constructing a correlation coefficient matrix between X-axis flywheel monitoring data samples and Y-axis flywheel monitoring data samples, a correlation coefficient matrix between X-axis flywheel monitoring data samples and Z-axis flywheel monitoring data samples and a correlation coefficient matrix between Y-axis flywheel monitoring data samples and Z-axis flywheel monitoring data samples by taking three flywheels of a spacecraft control system X, Y, Z as an example.
And S223, obtaining a correlation change parameter by using the correlation coefficient matrix. The correlation change trend graphs shown in fig. 4a-4c may be obtained based on the calculated correlation coefficient matrix, wherein the abscissa is the sampling period and the ordinate is the correlation coefficient. The area with the changed correlation can be obtained through the correlation change matrix and the correlation change trend graph. As shown in fig. 5, the correlation of the area a is not significantly changed, and may be defined as a normal area, the overall trend of the correlation of the area B is decreased, and may be defined as an abnormal area, and the overall trend of the correlation of the area C is decreased severely, and may be defined as a fault area.
As an exemplary embodiment, in the process of acquiring the monitoring data sample, the monitoring data sample may be interfered by the working condition of the equipment itself or the outside, so that the sampled data at one time or several times fluctuates greatly, and the finally obtained change trend changes greatly in the local time, for example, the change trends of the correlation coefficient change graphs shown in fig. 4a to 4C and 5 are in a peak and valley staggered distribution, and it can be determined from the overall trend of the whole period that the overall trend is in a falling region in the region B and in a severe falling trend in the region C, in this embodiment, the overall change trend of the whole period of the correlation coefficient can be described by using an envelope spectrum, the overall distribution condition of the correlation coefficient in the whole period can be described, the upper and lower threshold change features of the feature change can be extracted from the global perspective, and the feature change of the local instability can be removed, the implementation of a relatively long-scale fault prediction, specifically, as shown in fig. 6, may include the following steps:
s2231, constructing a time-sequence-based correlation coefficient change curve of the correlation coefficient based on the correlation coefficient matrix. 4a-4c, wherein fig. 4a is a correlation coefficient variation trend between current data samples of an X-axis flywheel and a Y-axis flywheel of a spacecraft in one year; FIG. 4b is a correlation coefficient variation trend between current data samples of an X-axis flywheel and a Z-axis flywheel of the spacecraft in one year; FIG. 4c is a correlation coefficient variation trend between current data samples of a Y-axis flywheel and a Z-axis flywheel of the spacecraft in one year.
And S2232, calculating an envelope spectrum of the correlation coefficient change curve to obtain a correlation change parameter. For example, the envelope spectrum may be constructed by referring to each of the correlation coefficient trend curves of fig. 4a to 4c, and the upper envelope T1 and the lower envelope T2 may be constructed by referring to the envelope spectrum shown in fig. 7. And adjusting the hyper-parameters of the envelope spectrum to obtain a curve form capable of meeting the degradation trend of the equipment. The envelope spectrogram after the hyper-parameter adjustment can be referred to as the envelope spectrogram after the parameter adjustment shown in fig. 8. Specifically, a hilbert envelope spectrum is constructed. The principle of the hilbert envelope spectrum is as follows:
in the implementation of the signal envelope, the Hilbert transform cannot be used directly, and is a continuous-time signal, given a continuous-time signal x (t), the Hilbert transform of which is given
Can be defined as
In the formula
Can be viewed as x (t) output through a filter with a unit impulse response h (t) of 1/t. The Hilbert transform of the signal is the convolution of the original signal with 1/t in the time domain. The result of the hilbert transform is to provide the original real signal with a signal that is constant in amplitude and frequency, but shifted in phase by 90 °. The Hilbert transform is the basis for generating the analytic signal, the original signal x (t) and its transform
Respectively form an analytic signal x+(t) real and imaginary components, as follows:
the amplitude A (t) is the envelope of the signal x (t), and the envelope signal is subjected to Fourier transform to obtain an envelope spectrum.
The hyper-parameters of the envelope spectrum are adjusted to form envelope curves as suitable as possible, including an upper envelope and a lower envelope. As an exemplary embodiment, the hyper-parameter of the envelope spectrum may be adjusted according to a preset target, and specifically, the length of the fir (finite Impulse response) filter may be adjusted according to a preset form of envelope to make the envelope tend to be smooth. The envelope after referencing can be seen in fig. 8.
As an exemplary embodiment, the lower threshold of the correlation coefficient change curve is similar in performance under normal and abnormal conditions, and may cause interference to the entire model, and for this reason, the upper envelope curve of the entire correlation coefficient curve may be extracted by combining with the hilbert envelope spectrum model, and the lower envelope spectrum is discarded, as shown in fig. 8, which can ensure that fewer interference factors for model early warning are caused, so as to improve the accuracy of model early warning.
Referring to fig. 8, the x-axis coordinate represents days, the y-axis coordinate represents correlation coefficients, 0-350 represents data of a year, the correlation shows a relatively stable trend, and the envelope curve on the envelope spectrum is approximately around 1. When the data of 5-6 months in A +1 year are subjected to correlation calculation, the trend that the correlation is reduced begins to appear in 5 months, and early warning can be performed. The 6 month correlation has dropped below 0.3 and the flywheel has failed severely. Compared with the prior knowledge (the flywheel fails in 5-6 months in the year A + 1) and the actual model curve, the model can perform abnormity early warning by utilizing the correlation of the flywheel current, and has a good effect.
The embodiment of the invention provides a device for constructing an early warning model, and as shown in fig. 9, the device may include: the first acquisition module 10 is used for acquiring monitoring data samples of at least two linkage parts in the equipment; the calculation module 20 is used for calculating correlation change parameters among monitoring data samples corresponding to different linkage parts, and the correlation change parameters are used for representing time sequence-based fault change trends; and the model construction module 30 is used for constructing an early warning model according to the monitoring data samples and the related change parameters.
An embodiment of the present invention provides an equipment fault early warning device, as shown in fig. 10, the device may include: the second acquisition module 40 is used for acquiring at least two monitoring data corresponding to at least two linkage parts in the equipment; the detection module 50 is configured to input the monitoring data into the fault early warning model to obtain an abnormal monitoring result, the abnormal monitoring model is constructed based on monitoring relevant change parameters between monitoring data samples corresponding to different linkage components, the relevant change parameters are used for representing a time sequence-based fault change trend, and the monitoring data samples are obtained based on monitoring at least two linkage components of the equipment.
An embodiment of the present invention provides an electronic device, as shown in fig. 11, the electronic device includes one or more processors 101 and a memory 102, where one processor 101 is taken as an example in fig. 11.
The controller may further include: an input device 103 and an output device 104.
The processor 101, the memory 102, the input device 103, and the output device 104 may be connected by a bus or other means, and fig. 11 illustrates an example of connection by a bus.
The processor 101 may be a Central Processing Unit (CPU). The processor 101 may also be other general purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field Programmable Gate Arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or combinations thereof. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
The memory 102, which is a non-transitory computer readable storage medium, may be used to store non-transitory software programs, non-transitory computer executable programs, and modules, such as program instructions/modules corresponding to the control methods in the embodiments of the present application. The processor 101 executes various functional applications and data processing of the server by running the non-transitory software programs, instructions and modules stored in the memory 32, that is, implementing the construction method of the warning model and/or the equipment fault warning method of the above method embodiments.
The memory 102 may include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function; the storage data area may store data created according to use of a processing device operated by the server, and the like. Further, the memory 102 may include high speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid state storage device. In some embodiments, memory 102 may optionally include memory located remotely from processor 101, which may be connected to a network connection device via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The input device 103 may receive input numeric or character information and generate key signal inputs related to user settings and function control of the processing device of the server. The output device 104 may include a display device such as a display screen.
One or more modules are stored in the memory 102 and, when executed by the one or more processors 101, perform a method as shown in any of fig. 1, 2, 4, 6.
It is noted that, in this document, relational terms such as "first" and "second," and the like, may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
The foregoing are merely exemplary embodiments of the present invention, which enable those skilled in the art to understand or practice the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (11)
1. A construction method of an early warning model is characterized by comprising the following steps:
acquiring monitoring data samples of at least two linkage parts in equipment;
calculating correlation change parameters among monitoring data samples corresponding to different linkage parts, wherein the correlation change parameters are used for representing time sequence-based fault change trends;
and constructing an early warning model according to the monitoring data samples and the related change parameters.
2. The method for constructing an early warning model according to claim 1, wherein the calculating of correlation variation parameters between monitoring data samples of different monitored components comprises:
calculating a plurality of correlation coefficients between the monitoring data samples of the different components corresponding to a plurality of unit times;
constructing a correlation coefficient matrix in the preset time period based on the plurality of correlation coefficients;
and obtaining the correlation change parameter by using the correlation coefficient matrix.
3. The method for constructing an early warning model according to claim 2, wherein the obtaining the relevant change parameter by using the correlation coefficient matrix comprises:
constructing a time-sequence-based correlation coefficient change curve of the correlation coefficient based on the correlation coefficient matrix;
and calculating the envelope spectrum of the correlation coefficient change curve to obtain the correlation change parameter.
4. The method for constructing an early warning model according to claim 3, wherein the calculating the envelope spectrum of the correlation coefficient change curve to obtain the correlation change parameter comprises:
constructing a Hilbert envelope spectrum of the correlation coefficient change curve;
and screening the Hilbert envelope spectrum based on a preset fault change trend to obtain an envelope line meeting the preset fault change trend.
5. The method for constructing an early warning model according to claim 4, wherein before or after the hilbert envelope spectrum is screened based on a preset fault variation trend, the method further comprises:
and adjusting the hyper-parameters of the envelope spectrum according to a preset target.
6. The method for constructing an early warning model according to claim 4, wherein the step of constructing the calculation mapping from the monitoring data samples to the correlation parameters to obtain the early warning model comprises the steps of:
extracting the envelope characteristics;
and determining the evaluation standard of the fault early warning model according to the envelope characteristic.
7. The method for constructing an early warning model according to claim 1, wherein after calculating correlation variation parameters between monitoring data samples corresponding to different monitored components, the method further comprises:
fitting the data in each data neighborhood in the monitoring data sample by utilizing an n-order polynomial to obtain a denoised monitoring data sample, wherein the coefficient of the n-order polynomial is determined by a least square method criterion under the condition of the minimum fitting error, and n is an integer greater than or equal to 1.
8. An equipment fault early warning method is characterized by comprising the following steps:
acquiring monitoring data of at least two linkage parts in equipment;
and inputting the monitoring data into a fault early warning model to obtain an abnormal monitoring result, wherein the fault early warning model is constructed on the basis of monitoring correlation change parameters between monitoring data samples corresponding to different linkage parts, the correlation change parameters are used for representing a time sequence-based fault change trend, and the monitoring data samples are obtained on the basis of monitoring at least two linkage parts of the equipment.
9. The equipment fault early warning method of claim 8, wherein prior to inputting the monitoring data into a fault early warning model, the method further comprises:
fitting the data in each data neighborhood in the monitoring data by utilizing an n-order polynomial to obtain a denoised monitoring data sample, wherein the coefficient of the n-order polynomial is determined by a least square method criterion under the condition of the minimum fitting error, and n is an integer greater than or equal to 1.
10. An early warning model construction device is characterized by comprising:
the device comprises a first acquisition module, a second acquisition module and a third acquisition module, wherein the first acquisition module is used for acquiring monitoring data samples of at least two linkage parts in the device;
the calculation module is used for calculating correlation change parameters among monitoring data samples corresponding to different linkage parts, and the correlation change parameters are used for representing time sequence-based fault change trends;
and the model construction module is used for constructing an early warning model according to the monitoring data samples and the related change parameters.
11. An equipment fault early warning device, comprising:
the second acquisition module is used for acquiring at least two monitoring data corresponding to at least two linkage parts in the equipment;
the detection module is used for inputting the monitoring data into a fault early warning model to obtain an abnormal monitoring result, the abnormal monitoring model is obtained by constructing correlation change parameters on the basis of monitoring the correlation change parameters between monitoring data samples corresponding to different linkage parts, the correlation change parameters are used for representing the fault change trend based on time sequence, and the monitoring data samples are obtained by monitoring at least two linkage parts of the equipment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011182775.7A CN112307619B (en) | 2020-10-29 | 2020-10-29 | Construction method of early warning model, equipment fault early warning method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011182775.7A CN112307619B (en) | 2020-10-29 | 2020-10-29 | Construction method of early warning model, equipment fault early warning method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112307619A true CN112307619A (en) | 2021-02-02 |
| CN112307619B CN112307619B (en) | 2024-06-11 |
Family
ID=74331703
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011182775.7A Active CN112307619B (en) | 2020-10-29 | 2020-10-29 | Construction method of early warning model, equipment fault early warning method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112307619B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112836990A (en) * | 2021-03-02 | 2021-05-25 | 浙江数智交院科技股份有限公司 | Tunnel monitoring equipment fault judgment method and device and electronic equipment |
| CN113027746A (en) * | 2021-03-12 | 2021-06-25 | 苏州德姆斯信息技术有限公司 | Fault monitoring method for reciprocating equipment |
| CN113345128A (en) * | 2021-05-31 | 2021-09-03 | 中铁工程装备集团有限公司 | Tunnel boring machine key component abnormity alarm method and device |
| CN115492692A (en) * | 2022-09-21 | 2022-12-20 | 厦门大学 | Method, device and equipment for detecting faults of fuel metering valve |
| CN116305531A (en) * | 2023-01-13 | 2023-06-23 | 北京控制工程研究所 | Spacecraft health evolution model modeling method, device, equipment and medium |
| CN116832385A (en) * | 2023-08-04 | 2023-10-03 | 贝滨（广东）科技有限公司 | Fire hydrant water monitoring system based on Internet of things |
| CN117874713A (en) * | 2024-03-12 | 2024-04-12 | 山东沪金精工科技股份有限公司 | Mechanical equipment fault early warning system and using method thereof |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH09205429A (en) * | 1996-01-29 | 1997-08-05 | Toshiba Corp | Network fault diagnostic device, fault prediction device, and its diagnostic and prediction method |
| CN103699118A (en) * | 2013-12-18 | 2014-04-02 | 北京航天测控技术有限公司 | Method and device for analyzing abnormal state of spacecraft in operating process |
| JP2016018526A (en) * | 2014-07-11 | 2016-02-01 | 株式会社荏原製作所 | Failure diagnosis system and failure diagnosis method |
| US20160274551A1 (en) * | 2015-03-18 | 2016-09-22 | Accenture Global Services Limited | Method and system for predicting equipment failure |
| CN107291991A (en) * | 2017-05-25 | 2017-10-24 | 华侨大学 | A kind of Wind turbines early defect method for early warning based on dynamic network mark |
| CN107642347A (en) * | 2017-09-11 | 2018-01-30 | 中国石油大学(北京) | Accident forecast method for early warning and device under shale gas fractured well |
| CN108036941A (en) * | 2017-12-26 | 2018-05-15 | 浙江大学 | A kind of steam turbine bearing abnormal vibration analysis method based on correlation visual analysis |
| CN108983749A (en) * | 2018-07-10 | 2018-12-11 | 福州大学 | Diagnosing failure of photovoltaic array method based on K-SVD training sparse dictionary |
| US20190265303A1 (en) * | 2018-02-28 | 2019-08-29 | Denso Corporation | Battery monitoring system |
| CN110414741A (en) * | 2019-08-01 | 2019-11-05 | 优必爱信息技术（北京）有限公司 | A kind of polynary prediction technique of automobile parts demand, system and storage medium |
| CN111579731A (en) * | 2020-06-28 | 2020-08-25 | 华侨大学 | Early transformer defect early warning method based on combined model |
-
2020
- 2020-10-29 CN CN202011182775.7A patent/CN112307619B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH09205429A (en) * | 1996-01-29 | 1997-08-05 | Toshiba Corp | Network fault diagnostic device, fault prediction device, and its diagnostic and prediction method |
| CN103699118A (en) * | 2013-12-18 | 2014-04-02 | 北京航天测控技术有限公司 | Method and device for analyzing abnormal state of spacecraft in operating process |
| JP2016018526A (en) * | 2014-07-11 | 2016-02-01 | 株式会社荏原製作所 | Failure diagnosis system and failure diagnosis method |
| US20160274551A1 (en) * | 2015-03-18 | 2016-09-22 | Accenture Global Services Limited | Method and system for predicting equipment failure |
| CN107291991A (en) * | 2017-05-25 | 2017-10-24 | 华侨大学 | A kind of Wind turbines early defect method for early warning based on dynamic network mark |
| CN107642347A (en) * | 2017-09-11 | 2018-01-30 | 中国石油大学(北京) | Accident forecast method for early warning and device under shale gas fractured well |
| CN108036941A (en) * | 2017-12-26 | 2018-05-15 | 浙江大学 | A kind of steam turbine bearing abnormal vibration analysis method based on correlation visual analysis |
| US20190265303A1 (en) * | 2018-02-28 | 2019-08-29 | Denso Corporation | Battery monitoring system |
| CN108983749A (en) * | 2018-07-10 | 2018-12-11 | 福州大学 | Diagnosing failure of photovoltaic array method based on K-SVD training sparse dictionary |
| CN110414741A (en) * | 2019-08-01 | 2019-11-05 | 优必爱信息技术（北京）有限公司 | A kind of polynary prediction technique of automobile parts demand, system and storage medium |
| CN111579731A (en) * | 2020-06-28 | 2020-08-25 | 华侨大学 | Early transformer defect early warning method based on combined model |
Non-Patent Citations (2)
| Title |
|---|
| 房红征，邹克旭，伊大伟: "一种与状态和事件相关的卫星实时故障预警方法", 《电子测量与仪器学报》, pages 49 - 55 * |
| 王再英;白华宁;: "基于相关系数的过程系统故障检测与诊断方法", 化工学报, no. 12, pages 352 - 358 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112836990A (en) * | 2021-03-02 | 2021-05-25 | 浙江数智交院科技股份有限公司 | Tunnel monitoring equipment fault judgment method and device and electronic equipment |
| CN112836990B (en) * | 2021-03-02 | 2023-08-01 | 浙江数智交院科技股份有限公司 | Tunnel monitoring equipment fault judging method and device and electronic equipment |
| CN113027746A (en) * | 2021-03-12 | 2021-06-25 | 苏州德姆斯信息技术有限公司 | Fault monitoring method for reciprocating equipment |
| CN113027746B (en) * | 2021-03-12 | 2022-06-14 | 苏州德姆斯信息技术有限公司 | Fault monitoring method for reciprocating equipment |
| CN113345128A (en) * | 2021-05-31 | 2021-09-03 | 中铁工程装备集团有限公司 | Tunnel boring machine key component abnormity alarm method and device |
| CN115492692A (en) * | 2022-09-21 | 2022-12-20 | 厦门大学 | Method, device and equipment for detecting faults of fuel metering valve |
| CN115492692B (en) * | 2022-09-21 | 2024-06-04 | 厦门大学 | Fuel metering valve fault detection method, device and equipment |
| CN116305531A (en) * | 2023-01-13 | 2023-06-23 | 北京控制工程研究所 | Spacecraft health evolution model modeling method, device, equipment and medium |
| CN116305531B (en) * | 2023-01-13 | 2023-09-15 | 北京控制工程研究所 | Spacecraft health evolution model modeling method, device, equipment and medium |
| CN116832385A (en) * | 2023-08-04 | 2023-10-03 | 贝滨（广东）科技有限公司 | Fire hydrant water monitoring system based on Internet of things |
| CN117874713A (en) * | 2024-03-12 | 2024-04-12 | 山东沪金精工科技股份有限公司 | Mechanical equipment fault early warning system and using method thereof |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112307619B (en) | 2024-06-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112307619A (en) | Construction method of early warning model, and equipment fault early warning method and device | |
| Xu | Separable multi-innovation Newton iterative modeling algorithm for multi-frequency signals based on the sliding measurement window | |
| US11146445B2 (en) | Time series decomposition | |
| Olhede et al. | A generalized demodulation approach to time-frequency projections for multicomponent signals | |
| Long et al. | Diagnostics of filtered analog circuits with tolerance based on LS-SVM using frequency features | |
| Boateng et al. | Improving the effectiveness of the median filter | |
| CN114285728B (en) | Predictive model training method, traffic prediction device and storage medium | |
| EP3493109A1 (en) | System and methods for faster processor comparisons of visual graph features | |
| CN112380774B (en) | Dynamic modeling method and system based on residual echo state network | |
| Rahim | Applications of multitaper spectral analysis to nonstationary data | |
| CN116257742B (en) | Landslide displacement dynamic prediction method and device, electronic equipment and storage medium | |
| Elorrieta et al. | Discrete-time autoregressive model for unequally spaced time-series observations | |
| Omelchuk et al. | A closed-form ARMA-based ML-estimator of a single-tone frequency | |
| Rudner et al. | Inter-domain deep Gaussian processes | |
| CN108228959A (en) | Using the method for Random censorship estimating system virtual condition and using its wave filter | |
| Zhang et al. | A new regularized adaptive windowed Lomb periodogram for time–frequency analysis of nonstationary signals with impulsive components | |
| Liu et al. | Bearing failure diagnosis at time-varying speed based on adaptive clustered fractional Gabor transform | |
| Chan et al. | A new transform-domain regularized recursive least M-estimate algorithm for a robust linear estimation | |
| CN113127803A (en) | Method and device for establishing service cluster capacity estimation model and electronic equipment | |
| CN116046968A (en) | Liquid chromatography workstation data processing method, system and storable medium | |
| Eckley et al. | A test for the absence of aliasing or local white noise in locally stationary wavelet time series | |
| CN112988536B (en) | Data anomaly detection method, device, equipment and storage medium | |
| CN112561121B (en) | Rework trend prediction method and system based on mobile phone signaling data | |
| Zhang et al. | Recursive parametric frequency/spectrum estimation for nonstationary signals with impulsive components using variable forgetting factor | |
| Hu et al. | Incipient fault diagnostics and remaining useful life prediction of analog filters |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |