CN115130540A - Bridge abnormal point identification method and device based on multi-granularity time window - Google Patents

Bridge abnormal point identification method and device based on multi-granularity time window Download PDF

Info

Publication number
CN115130540A
CN115130540A CN202210436017.6A CN202210436017A CN115130540A CN 115130540 A CN115130540 A CN 115130540A CN 202210436017 A CN202210436017 A CN 202210436017A CN 115130540 A CN115130540 A CN 115130540A
Authority
CN
China
Prior art keywords
sample
abnormal
linear
time window
bridge
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
Application number
CN202210436017.6A
Other languages
Chinese (zh)
Inventor
谢海龙
郑建兴
刘一鸣
巩跃龙
韩之江
郝晨先
罗鹏
李洁
陈伟
梁立江
郝仰玥
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Jincheng Expressway Branch Shanxi Transporation Holdings Group Co ltd
Shanxi Province Traffic Construction Project Quality Testing Center (co Ltd)
Shanxi University
Original Assignee
Jincheng Expressway Branch Shanxi Transporation Holdings Group Co ltd
Shanxi Province Traffic Construction Project Quality Testing Center (co Ltd)
Shanxi University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Jincheng Expressway Branch Shanxi Transporation Holdings Group Co ltd, Shanxi Province Traffic Construction Project Quality Testing Center (co Ltd), Shanxi University filed Critical Jincheng Expressway Branch Shanxi Transporation Holdings Group Co ltd
Priority to CN202210436017.6A priority Critical patent/CN115130540A/en
Publication of CN115130540A publication Critical patent/CN115130540A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • G06Q10/06395—Quality analysis or management
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/08—Construction

Landscapes

  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Engineering & Computer Science (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • General Business, Economics & Management (AREA)
  • Educational Administration (AREA)
  • Marketing (AREA)
  • Tourism & Hospitality (AREA)
  • Physics & Mathematics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • General Physics & Mathematics (AREA)
  • Development Economics (AREA)
  • Theoretical Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • General Health & Medical Sciences (AREA)
  • Game Theory and Decision Science (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)

Abstract

The invention discloses a bridge abnormal point identification method and device based on a multi-granularity time window, wherein the method comprises the following steps: detecting and acquiring a plurality of sample data of the bridge by using a real-time monitoring system; constructing sample initial characteristic vector representation of a sliding time window, and fusing the sample initial characteristic vector representation to construct sample initial characteristic matrix representation; fitting linear relation and nonlinear relation of sample neighborhood characteristics of a sliding time window; based on the linear relation and the nonlinear relation, fusing the abnormal probability values of the linear residual errors and the nonlinear residual error block characteristics identification samples, and calculating the abnormal probability values of the samples to mark abnormal samples; and identifying abnormal values in the plurality of sample data through the marked abnormal samples so as to identify the bridge abnormal points. According to the method, the abnormal probability of the sample is predicted from the potential neighborhood characteristics of the sample in the multi-granularity sliding time window, the identification of the abnormal data points is realized, and great support is provided in the aspect of abnormal diagnosis of the bridge structure health monitoring data.

Description

Bridge abnormal point identification method and device based on multi-granularity time window
Technical Field
The invention relates to the technical field of bridge structure health monitoring abnormity identification, in particular to a bridge abnormity point identification method and device based on a multi-granularity time window.
Background
In the bridge structure health monitoring process, data collected by sensors are often accompanied with omission, noise and abnormal data, and the identification of the abnormal data is important for the structural safety diagnosis and evaluation of the bridge. The data scale collected by the sensor is large, so that the difficulty of manually identifying abnormal data is increased, the data preprocessing is performed on the time sequence monitoring data automatically collected by the sensor, the identification of the abnormal data is developed, the identification and trend analysis of the structural damage of the bridge are important links of intelligent bridge health monitoring, the correlation between bridge response and load can be effectively analyzed, and the cost of manual inspection in the bridge structural health monitoring is reduced.
The information of the neighbor sample needs to be considered for the abnormal recognition of the sample points in the time sequence data, the neighborhood characteristics of the sample points overcome the problem of less information of the current sample to a certain extent, the initial characteristic information of the sample is favorably expanded, the method has important help for the abnormal recognition, and the accuracy of the abnormal recognition of the sample can be improved.
Disclosure of Invention
The present invention is directed to solving, at least to some extent, one of the technical problems in the related art.
Therefore, one objective of the present invention is to provide a bridge abnormal point identification method based on a multi-granularity time window, which can improve the accuracy of bridge structure health monitoring abnormal data point identification. According to the method, the abnormal probability of the sample is predicted from the potential neighborhood characteristics of the sample in the multi-granularity sliding time window, the identification of the abnormal data points is realized, and great support is provided in the aspect of abnormal diagnosis of the bridge structure health monitoring data.
The invention also aims to provide a bridge abnormal point identification device based on the multi-granularity time window.
In order to achieve the above object, the present invention provides a bridge abnormal point identification method based on a multi-granularity time window, including:
detecting and acquiring a plurality of sample data of the bridge by using a real-time monitoring system; wherein the plurality of sample data comprises bridge displacement and strain; constructing a sample initial feature vector representation of a sliding time window based on the plurality of sample data, fusing the sample initial feature vector representation to construct a sample initial feature matrix representation; fitting linear relation and nonlinear relation of sample neighborhood characteristics of the sliding time window based on the sample initial characteristic matrix representation; based on the linear relation and the nonlinear relation, fusing linear residual and nonlinear residual block characteristics to identify the abnormal probability value of the sample, and calculating the abnormal probability value of the sample to mark an abnormal sample; and identifying abnormal values in the plurality of sample data through the marked abnormal samples so as to identify bridge abnormal points.
According to the bridge abnormal point identification method based on the multi-granularity time window, disclosed by the embodiment of the invention, the abnormal probability of the sample is predicted from the potential neighborhood characteristics of the sample in the multi-granularity sliding time window, the identification of the abnormal data point is realized, and great support is provided in the aspect of bridge structure health monitoring data abnormal diagnosis.
In addition, the bridge abnormal point identification method based on the multi-granularity time window according to the above embodiment of the present invention may further have the following additional technical features:
further, in an embodiment of the present invention, the constructing a sample initial feature vector representation of a sliding time window based on the plurality of sample data includes: initialization characteristic x of preset sample x at time t t Taking the neighborhood initialization feature vector of the sample x as: x is the number of 0 ＝(x t-2 ,x t-1 ，x t ，x t+1 ，x t+2 )。
Further, in an embodiment of the present invention, the fusing the sample initial feature vector representation to construct a sample initial feature matrix representation includes: according to a sliding time window δ, the δ left-neighborhood initialization feature vector of the sample x is represented as: x is a radical of a fluorine atom - ＝(x t-2-δ ,x t-1-δ ，x t-δ ，x t+1-δ ，x t+2-δ ) δ right neighborhood initialization feature vector is expressed as: x is the number of + ＝(x t-2+δ ,x t-1+δ ，x t+δ ，x t+1+δ ，x t+2+δ ) (ii) a Fusing the delta left neighborhood initialization feature and the right neighborhood initialization feature of the sample x, and constructing an initial feature matrix of the sample x to be expressed as: x ═ x 0 ；x - ；x + ]。
Further, in one embodiment of the present invention, the fitting of the linear relationship and the nonlinear relationship of the sample neighborhood features of the sliding time window based on the sample initial feature matrix representation includesComprises the following steps: representing x based on the initial feature matrix of the sample x by a linear classifier
Figure BDA0003612866470000021
Fitting a linear relation between the sample x and the classification label y; and, by means of a non-linear classifier
Figure BDA0003612866470000022
Fitting a non-linear relationship of the sample x and the classification label y.
Further, in an embodiment of the present invention, the fusing linear residual and non-linear residual block features to identify abnormal probability values of the samples based on the linear relationship and the non-linear relationship includes: combining the linear residuals of the samples x
Figure BDA0003612866470000023
And non-linear residual error
Figure BDA0003612866470000024
Feature of calculating the abnormal probability value of the sample x
Figure BDA0003612866470000025
Based on the anomaly probability value
Figure BDA0003612866470000026
Defining an anomaly probability value for the sample
Figure BDA0003612866470000027
Has an objective function of
Figure BDA0003612866470000028
Optimizing a weight parameter w 1 ，w 2 And W.
Further, in an embodiment of the present invention, the calculating the abnormal probability value of the sample to label an abnormal sample includes: for the sample s, according to the optimized weight parameter w 1 ，w 2 And W, calculating the abnormal probability value of the sample s
Figure BDA0003612866470000031
Comparing the abnormal probability values of the samples s
Figure BDA0003612866470000032
If the number is more than or equal to 0.5, marking the normal labeled sample, and marking the classification label as 1; otherwise, marking the sample as abnormal and marking the classification label as 0.
In order to achieve the above object, another aspect of the present invention provides a bridge anomaly point identification device based on a multi-granularity time window, including:
the data acquisition module is used for detecting and acquiring a plurality of sample data of the bridge by using the real-time monitoring system; wherein the plurality of sample data comprises bridge displacement and strain; the characteristic representation module is used for constructing sample initial characteristic vector representation of a sliding time window based on the plurality of sample data, and fusing the sample initial characteristic vector representation to construct sample initial characteristic matrix representation; the relation fitting module is used for fitting the linear relation and the nonlinear relation of the sample neighborhood characteristics of the sliding time window based on the sample initial characteristic matrix representation; the probability calculation module is used for fusing linear residual errors and nonlinear residual error block characteristics to identify the abnormal probability value of the sample based on the linear relation and the nonlinear relation, and calculating the abnormal probability value of the sample to label an abnormal sample; and the anomaly identification module is used for identifying an abnormal value in the plurality of sample data through the marked abnormal sample so as to identify the bridge abnormal point.
The bridge abnormal point identification device based on the multi-granularity time window predicts the abnormal probability of the sample from the potential neighborhood characteristics of the sample in the multi-granularity sliding time window, realizes the identification of the abnormal data point, and provides great support in the aspect of bridge structure health monitoring data abnormal diagnosis.
Additional aspects and advantages of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.
Drawings
The foregoing and/or additional aspects and advantages of the present invention will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:
FIG. 1 is a flowchart of a bridge anomaly point identification method based on a multi-granularity time window according to an embodiment of the present invention;
FIG. 2 is a sample diagram of annotation data according to an embodiment of the present invention;
FIG. 3 is a sample diagram of predicted anomalies according to an embodiment of the present invention;
fig. 4 is a schematic structural diagram of a bridge anomaly point identification device in a multi-granularity time window according to an embodiment of the present invention.
Detailed Description
It should be noted that, in the present application, the embodiments and features of the embodiments may be combined with each other without conflict. The present invention will be described in detail below with reference to the embodiments with reference to the attached drawings.
In order to make the technical solutions of the present invention better understood, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the 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 invention.
Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of the present invention, "a plurality" means two or more unless specifically defined otherwise.
In the present invention, unless otherwise expressly stated or limited, "above" or "below" a first feature means that the first and second features are in direct contact, or that the first and second features are not in direct contact but are in contact with each other via another feature therebetween. Also, the first feature being "on," "above" and "over" the second feature includes the first feature being directly on and obliquely above the second feature, or merely indicating that the first feature is at a higher level than the second feature. A first feature being "under," "below," and "beneath" a second feature includes the first feature being directly above and obliquely above the second feature, or simply meaning that the first feature is at a lesser level than the second feature.
The following describes a bridge anomaly point identification method and device based on a multi-granularity time window according to an embodiment of the present invention with reference to the accompanying drawings.
Fig. 1 is a flowchart of a bridge anomaly point identification method based on a multi-granularity time window according to an embodiment of the present invention.
As shown in fig. 1, the method includes, but is not limited to, the following steps:
step S1, detecting and acquiring a plurality of sample data of the bridge by using a real-time monitoring system; wherein the plurality of sample data comprises bridge displacement and strain.
It can be understood that the invention can obtain the relevant data of bridge health monitoring by installing sensors and other devices on the bridge.
Specifically, in the process of acquiring actual data of the bridge, the road surface change conditions of the bridge, such as displacement, strain and the like, are mainly detected through a wireless displacement sensor, an acceleration sensor, a vibration sensor and other real-time monitoring systems; in addition, data such as pictures and videos can be monitored through a camera, collected through an internet of things gateway and transmitted to a server and a measurement and control terminal display platform through a base station and the like.
And step S2, constructing sample initial characteristic vector representation of the sliding time window based on a plurality of sample data, and fusing the sample initial characteristic vector representation to construct sample initial characteristic matrix representation.
As an example, let the initialization characteristic x of a sample x at time t t The neighborhood initialization feature vector of the sample x is denoted as x 0 ＝(x t-2 ,x t-1 ，x t ，x t+1 ，x t+2 ) To construct a sample initial feature vector representation of the sliding time window.
As an example, according to a sliding time window δ, the left neighborhood initialization feature vector of δ for a sample book x is denoted as x - ＝(x t-2-δ ,x t-1-δ ，x t-δ ，x t+1-δ ，x t+2-δ ) δ right neighborhood initialization feature vector is denoted as x + ＝(x t-2+δ ,x t-1+δ ，x t+δ ，x t+1+δ ，x t+2+δ ) To enhance the input characteristics of sample x.
And an initial characteristic matrix of the sample x is constructed to express x ═ x by fusing delta left neighborhood initialization characteristic and right neighborhood initialization characteristic of the sample x 0 ；x - ；x + ]A sample initial feature matrix representation is constructed with the sample initial feature vector representation fused to the multi-granularity sliding time window.
And step S3, fitting the linear relation and the nonlinear relation of the sample neighborhood characteristics of the sliding time window based on the sample initial characteristic matrix representation.
Specifically, x is represented based on an initial feature matrix of the sample x, and is classified by a linear classifier
Figure BDA0003612866470000051
And fitting the linear relation between the sample x and the classification label y so as to fit the linear relation of the sample neighborhood characteristics of the multi-granularity sliding time window based on the linear classifier.
Further, x is represented based on the initial feature matrix of the sample x, and is classified by a nonlinear classifier
Figure BDA0003612866470000052
Fitting the nonlinear relation between the sample x and the classification label y so as to fit the nonlinear relation of the sample neighborhood characteristics of the multi-granularity sliding time window based on the nonlinear classifier.
And step S4, fusing the abnormal probability values of the linear residual error and the nonlinear residual error block characteristic identification samples based on the linear relation and the nonlinear relation, and calculating the abnormal probability values of the samples to label the abnormal samples.
In particular, linear residuals of combined samples x
Figure BDA0003612866470000053
And non-linear residual error
Figure BDA0003612866470000054
Feature, calculating the probability of class anomaly of sample x
Figure BDA0003612866470000055
Further, the probability value of the sample classification abnormality
Figure BDA0003612866470000056
On the basis of (1), an objective function defining the anomaly probability of the sample is
Figure BDA0003612866470000057
Optimizing a weight parameter w 1 ，w 2 And W.
Further, for the new sample s, according to the learned weight parameter w 1 ，w 2 And W, calculating the classification abnormal probability of the new sample
Figure BDA0003612866470000058
Further, sample-based classification anomaly probability
Figure BDA0003612866470000059
If the sample is larger than or equal to 0.5, marking the sample as normal, and marking the classification label as 1, otherwise, marking the sample as abnormal as 0.
Further, as shown in fig. 2, part of the outliers are marked manually based on the sigma principle, the sample is marked as 1 normally, and the sample is marked as 0 abnormally. The annotation data sample is shown, for example, in fig. 2.
Further, through the identification of the abnormal point of the multi-granularity time window, the label of the abnormal point is predicted. 70000 samples were input, the samples are marked as abnormal, red is abnormal, and green is normal, as shown in fig. 3.
Therefore, abnormal values in the displacement monitoring data are identified through analyzing and processing the bridge health monitoring data, potential safety hazards of the bridge are found, and decision support is provided for comprehensive structural state evaluation, bridge damage identification and management maintenance of the bridge. The abnormal value mainly refers to abnormal data in data analysis.
According to the bridge abnormal point identification method based on the multi-granularity time window, disclosed by the embodiment of the invention, the abnormal probability of the sample is predicted from the potential neighborhood characteristics of the sample in the multi-granularity sliding time window, the identification of an abnormal data point is realized, and great support is provided in the aspect of bridge structure health monitoring data abnormal diagnosis.
In order to implement the foregoing embodiment, as shown in fig. 4, a bridge abnormal point identification apparatus 10 based on a multi-granularity time window is further provided in this embodiment, where the apparatus 10 includes: data acquisition module 100, feature representation module 200, relationship fitting module 300, probability calculation module 400, and anomaly identification module 500.
The data acquisition module 100 is used for detecting and acquiring a plurality of sample data of the bridge by using a real-time monitoring system; wherein the plurality of sample data comprises bridge displacement and strain;
the feature representation module 200 is configured to construct a sample initial feature vector representation of a sliding time window based on a plurality of sample data, and fuse the sample initial feature vector representation to construct a sample initial feature matrix representation;
a relation fitting module 300, configured to fit a linear relation and a nonlinear relation of a sample neighborhood characteristic of the sliding time window based on the sample initial characteristic matrix representation;
a probability calculation module 400, configured to fuse the linear residual and the nonlinear residual block feature identification sample abnormal probability values based on a linear relationship and a nonlinear relationship, and calculate an abnormal probability value of the sample to label the abnormal sample;
and the anomaly identification module 500 is used for identifying an anomaly value in the plurality of sample data through the labeled anomaly sample so as to identify the bridge anomaly point.
According to the bridge abnormal point identification device based on the multi-granularity time window, provided by the embodiment of the invention, the abnormal probability of the sample is predicted from the potential neighborhood characteristics of the sample in the multi-granularity sliding time window, the identification of the abnormal data point is realized, and great support is provided in the aspect of bridge structure health monitoring data abnormal diagnosis.
It should be noted that the foregoing explanation of the embodiment of the method for identifying a bridge anomaly point based on a multi-granularity time window is also applicable to the apparatus for identifying a bridge anomaly point based on a multi-granularity time window in this embodiment, and is not repeated here.
In the description of the specification, reference to the description of "one embodiment," "some embodiments," "an example," "a specific example," or "some examples" or the like means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention, and that variations, modifications, substitutions and alterations can be made in the above embodiments by those of ordinary skill in the art without departing from the principle and spirit of the present invention.

Claims (10)

1. A bridge abnormal point identification method based on a multi-granularity time window is characterized by comprising the following steps:
detecting and acquiring a plurality of sample data of the bridge by using a real-time monitoring system; wherein the plurality of sample data comprises bridge displacement and strain;
constructing a sample initial feature vector representation of a sliding time window based on the plurality of sample data, fusing the sample initial feature vector representation to construct a sample initial feature matrix representation;
fitting a linear relation and a nonlinear relation of sample neighborhood characteristics of the sliding time window based on the sample initial characteristic matrix representation;
based on the linear relation and the nonlinear relation, fusing linear residual and nonlinear residual block characteristics to identify the abnormal probability value of the sample, and calculating the abnormal probability value of the sample to mark an abnormal sample;
and identifying abnormal values in the plurality of sample data through the marked abnormal samples so as to identify bridge abnormal points.
2. The method of claim 1, wherein said constructing a sample initial feature vector representation of a sliding time window based on said plurality of sample data comprises:
initialization characteristic x of preset sample x at time t t Taking the neighborhood initialization feature vector of the sample x as: x is the number of 0 ＝(x t-2 ,x t-1 ，x t ，x t+1 ，x t+2 )。
3. The method of claim 2, wherein said fusing the sample initial eigenvector representations to construct a sample initial eigenvector matrix representation comprises:
according to a sliding time window δ, the δ left-neighborhood initialization feature vector of the sample x is represented as: x is the number of - ＝(x t-2-δ ,x t-1-δ ，x t-δ ，x t+1-δ ，x t+2-δ ) δ right neighborhood initialization feature vector is expressed as: x is the number of + ＝(x t-2+δ ,x t-1+δ ，x t+δ ，x t+1+δ ，x t+2+δ )；
Fusing the delta left neighborhood initialization feature and the right neighborhood initialization feature of the sample x, and constructing an initial feature matrix of the sample x to be expressed as: x ═ x 0 ；x - ；x + ]。
4. The method of claim 3, wherein fitting linear and non-linear relationships of sample neighborhood features of the sliding time window based on the sample initial feature matrix representation comprises:
representing x based on the initial feature matrix of the sample x by a linear classifier
Figure FDA0003612866460000011
Fitting a linear relation between the sample x and the classification label y; and the number of the first and second groups,
by means of non-linear classifiers
Figure FDA0003612866460000021
Fitting a non-linear relationship of the sample x and the classification label y.
5. The method of claim 4, wherein the fusing linear residual and non-linear residual block features to identify abnormal probability values for the samples based on the linear relationship and the non-linear relationship comprises:
combining the linear residuals of the samples x
Figure FDA0003612866460000022
And non-linear residual error
Figure FDA0003612866460000023
Feature of calculating the abnormal probability value of the sample x
Figure FDA0003612866460000024
Based on the abnormal probability value
Figure FDA0003612866460000025
Defining an anomaly probability value for the sample
Figure FDA0003612866460000026
Has an objective function of
Figure FDA0003612866460000027
Optimizing a weight parameter w 1 ，w 2 And W.
6. The method of claim 5, wherein the calculating the anomaly probability values for the samples to label anomalous samples comprises:
for the sample s, according to the optimized weight parameter w 1 ，w 2 And W, calculating the abnormal probability value of the sample s
Figure FDA0003612866460000028
Comparing the abnormal probability values of the samples s
Figure FDA0003612866460000029
If the number is more than or equal to 0.5, marking the normal labeled sample, and marking the classification label as 1; otherwise, marking the sample as abnormal and marking the classification label as 0.
7. A bridge outlier recognition device based on multi-granularity time window is characterized by comprising:
the data acquisition module is used for detecting and acquiring a plurality of sample data of the bridge by using the real-time monitoring system; wherein the plurality of sample data comprises bridge displacement and strain;
the characteristic representation module is used for constructing sample initial characteristic vector representation of a sliding time window based on the plurality of sample data, and fusing the sample initial characteristic vector representation to construct sample initial characteristic matrix representation;
the relation fitting module is used for fitting a linear relation and a nonlinear relation of the sample neighborhood characteristics of the sliding time window based on the sample initial characteristic matrix representation;
the probability calculation module is used for fusing linear residual errors and nonlinear residual error block characteristics to identify the abnormal probability value of the sample based on the linear relation and the nonlinear relation, and calculating the abnormal probability value of the sample to label an abnormal sample;
and the anomaly identification module is used for identifying an anomaly value in the plurality of sample data through the marked anomaly sample so as to identify a bridge anomaly point.
8. The apparatus of claim 7, wherein the feature representation module is further configured to:
initialization characteristic x of preset sample x at time t t And expressing the neighborhood initialization feature vector of the sample x as: x is the number of 0 ＝(x t-2 ,x t-1 ，x t ，x t+1 ，x t+2 )。
9. The apparatus of claim 8, wherein the feature representation module comprises:
a feature representation submodule, configured to represent, according to a sliding time window δ, a δ left-neighbor initialization feature vector of the sample x as: x is the number of - ＝(x t-2-δ ,x t-1-δ ，x t-δ ，x t+1-δ ，x t+2-δ ) δ right neighborhood initialization feature vector is expressed as: x is a radical of a fluorine atom + ＝(x t-2+δ ,x t-1+δ ，x t+δ ，x t+1+δ ，x t+2+δ )；
A fusion construction module, configured to fuse the δ left neighborhood initialization feature and the right neighborhood initialization feature of the sample x, and construct an initial feature matrix of the sample x to be represented as: x ═ x 0 ；x - ；x + ]。
10. The apparatus of claim 9, wherein the relationship fitting module comprises:
a linear fitting module for representing x by a linear classifier based on the initial feature matrix of the sample x
Figure FDA0003612866460000031
Fitting a linear relation between the sample x and the classification label y; and the number of the first and second groups,
a non-linear fitting module for passing through a non-linear classifier
Figure FDA0003612866460000032
Fitting a non-linear relationship of the sample x and the classification label y.
CN202210436017.6A 2022-04-24 2022-04-24 Bridge abnormal point identification method and device based on multi-granularity time window Pending CN115130540A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210436017.6A CN115130540A (en) 2022-04-24 2022-04-24 Bridge abnormal point identification method and device based on multi-granularity time window

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210436017.6A CN115130540A (en) 2022-04-24 2022-04-24 Bridge abnormal point identification method and device based on multi-granularity time window

Publications (1)

Publication Number Publication Date
CN115130540A true CN115130540A (en) 2022-09-30

Family

ID=83377031

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210436017.6A Pending CN115130540A (en) 2022-04-24 2022-04-24 Bridge abnormal point identification method and device based on multi-granularity time window

Country Status (1)

Country Link
CN (1) CN115130540A (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117216701A (en) * 2023-09-13 2023-12-12 广州桐富科技发展有限公司 Intelligent bridge monitoring and early warning method and system
CN117521084A (en) * 2023-12-13 2024-02-06 北京熵度科技有限责任公司 Active safety early warning method for complex system
CN117708643A (en) * 2023-11-07 2024-03-15 中交公路长大桥建设国家工程研究中心有限公司 Bridge monitoring abnormal data identification method and system based on fusion sequence characteristics
CN117708643B (en) * 2023-11-07 2024-07-09 中交公路长大桥建设国家工程研究中心有限公司 Bridge monitoring abnormal data identification method and system based on fusion sequence characteristics

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117216701A (en) * 2023-09-13 2023-12-12 广州桐富科技发展有限公司 Intelligent bridge monitoring and early warning method and system
CN117216701B (en) * 2023-09-13 2024-03-15 华夏安信物联网技术有限公司 Intelligent bridge monitoring and early warning method and system
CN117708643A (en) * 2023-11-07 2024-03-15 中交公路长大桥建设国家工程研究中心有限公司 Bridge monitoring abnormal data identification method and system based on fusion sequence characteristics
CN117708643B (en) * 2023-11-07 2024-07-09 中交公路长大桥建设国家工程研究中心有限公司 Bridge monitoring abnormal data identification method and system based on fusion sequence characteristics
CN117521084A (en) * 2023-12-13 2024-02-06 北京熵度科技有限责任公司 Active safety early warning method for complex system
CN117521084B (en) * 2023-12-13 2024-06-11 北京熵度科技有限责任公司 Active safety early warning method for complex system

Similar Documents

Publication Publication Date Title
CN115130540A (en) Bridge abnormal point identification method and device based on multi-granularity time window
JP7017861B2 (en) Anomaly detection system and anomaly detection method
EP2905665B1 (en) Information processing apparatus, diagnosis method, and program
TWI782231B (en) Deep auto-encoder for equipment health monitoring and fault detection in semiconductor and display process equipment tools
JP2020098646A (en) Device, method, and system for converting cross domain time-series data
CN105579922A (en) Information processing device and analysis method
WO2018104985A1 (en) Abnormality analysis method, program, and system
KR20170122043A (en) Real-time indoor air quality outlier smoothing method and apparatus
KR20130100266A (en) Method for inspecting the quality of a solder joint
EP2963552B1 (en) System analysis device and system analysis method
JP7308717B2 (en) Platform door diagnostic device, platform door system, and platform door diagnostic program
JP2000259223A (en) Plant monitoring device
KR102618023B1 (en) Failure prediction diagnosis system and method through pattern analysis according to failure type
CN108921305B (en) Component life period monitoring method
CN109313442B (en) Automated visual and acoustic analysis for event detection
JP6823025B2 (en) Inspection equipment and machine learning method
JP4284322B2 (en) A method for rating and temporal stabilization of classification results
JP6587950B2 (en) Program, apparatus, and method capable of detecting time series change point by scalar feature
CN116907772A (en) Self-diagnosis and fault source identification method and system of bridge structure monitoring sensor
JP6579163B2 (en) Process condition diagnosis method and condition diagnosis apparatus
CN114970712A (en) Time sequence data abnormal point monitoring method based on neighborhood sample feature contrast learning
US20220405620A1 (en) Generation Device, Data Analysis System, Generation Method, and Generation Program
CN116649850B (en) Intelligent floor washing machine control system
CN115496931B (en) Industrial robot health monitoring method and system
CN117191126B (en) Container self-checking system, method, device and storage medium

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