CN110287827B - Bridge strain data outlier identification method based on data correlation - Google Patents
Bridge strain data outlier identification method based on data correlation Download PDFInfo
- Publication number
- CN110287827B CN110287827B CN201910500098.XA CN201910500098A CN110287827B CN 110287827 B CN110287827 B CN 110287827B CN 201910500098 A CN201910500098 A CN 201910500098A CN 110287827 B CN110287827 B CN 110287827B
- Authority
- CN
- China
- Prior art keywords
- data
- strain
- value
- original
- acceleration
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Mathematical Physics (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Operations Research (AREA)
- Probability & Statistics with Applications (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Algebra (AREA)
- Evolutionary Biology (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Testing Or Calibration Of Command Recording Devices (AREA)
Abstract
The invention discloses a bridge strain data abnormal value identification method based on data relevance, which comprises the following steps: acquiring sensor original data installed at the same position; extracting long-period trend data from the original strain data through wavelet decomposition to obtain a difference value between the original strain data and the long-period trend data; obtaining the variation range of potential abnormal values in the original strain data according to the mean value and standard deviation of the strain difference data; extracting long-period trend data from the original acceleration data through wavelet decomposition to obtain a difference value between the original acceleration data and the long-period trend data; obtaining the variation range of the potential abnormal value in the original acceleration data according to the mean value and standard deviation of the acceleration difference data; comparing the appearance position of the potential abnormal value in the original strain data with the appearance position of the potential abnormal value in the original acceleration data, and judging and identifying the real abnormal value of the original strain data. The method can accurately and efficiently treat the problem of abnormal values of strain data in bridge health monitoring.
Description
Technical Field
The invention belongs to the field of bridge health monitoring data analysis and research, and particularly relates to a bridge strain data outlier identification method based on data correlation.
Background
The application of bridge health monitoring technology at home and abroad is mature, the functions of the large bridge structure health monitoring system are improved gradually, but due to the increasing of the component links and influence factors, the potential possibility of faults is increased gradually, the contradiction between the long-term performance of bridge health monitoring tasks and the limited service life of monitoring system equipment is more prominent, for example, the faults of individual parts often cause chain reactions, so that the whole system cannot run and even is paralyzed. Especially, if the distortion of the monitoring data caused by the fault of the monitoring system cannot be found in time, the accuracy of the safety evaluation result is affected, so that the health monitoring task of the bridge structure cannot be successfully completed. According to Houser and Aktan et al, more than 80% of false alarms in existing bridge health monitoring systems are due to data distortion [1] . Therefore, the monitoring data needs to be cleaned to realize the value of the sensor data and improve the usability and efficiency of the bridge health monitoring system.
The abnormal value of the strain data is mainly caused by power failure, instantaneous fluctuation of a meter or other signals, and pulse interference is formed, and the abnormal value is mainly caused by overlarge fluctuation amplitude of the signals. However, strain data itself fluctuates greatly, and when a vehicle passes, large signal fluctuation is caused, and such signal fluctuation and data outliers are difficult to distinguish from a single signal source. Therefore, in view of the above drawbacks of the abnormal data, there is an urgent need to develop a method for cleaning the abnormal data of the bridge monitoring dynamic strain in real time.
Reference is made to:
[1] he Dongdong, zhang Li, zhang Huili, li Xianke application of the "3σ criterion" to singular value inspection of bridge health monitoring data [ J ]. Extra highway, 2013,33 (6): 107-110.
Disclosure of Invention
The invention aims to solve the technical problem of providing a bridge strain data abnormal value identification method based on data correlation aiming at the defects of the prior art, and the bridge strain data abnormal value identification method based on the data correlation can accurately and efficiently treat the problem of the abnormal value of the strain data in bridge health monitoring.
In order to achieve the technical purpose, the invention adopts the following technical scheme:
a bridge strain data abnormal value identification method based on data correlation comprises the following steps:
step 1, acquiring original data of sensors installed at the same position;
step 2, extracting long-period trend data from the original strain data through wavelet decomposition to obtain a difference value between the original strain data and the long-period trend data;
step 3, obtaining the variation range of potential abnormal values in the original strain data according to the mean value and standard deviation of the strain difference data obtained in the step 2;
step 4, extracting long-period trend data from the original acceleration data through wavelet decomposition to obtain a difference value between the original acceleration data and the long-period trend data;
step 5, obtaining the variation range of the potential abnormal value in the original acceleration data according to the mean value and standard deviation of the acceleration difference value data obtained in the step 4;
and 6, comparing the appearance position of the potential abnormal value in the original strain data with the appearance position of the potential abnormal value in the original acceleration data, and judging and identifying the real abnormal value of the original strain data.
As a further improved technical scheme of the invention, the specific steps of the step 1 are as follows:
acquiring original strain data { x } acquired by strain sensors at the same location of a health monitoring system i I=1, 2, …, n and raw acceleration data { y ] acquired by the acceleration sensor i I=1, 2, …, n, where n represents the number of data.
As a further improved technical scheme of the invention, the specific steps of the step 2 are as follows:
step 2.1, extracting long period trend data { s } of the original strain data by wavelet decomposition i },i＝1,2,…,n；
Step 2.2, calculating x i And s i Is analyzed and denoted d i ＝x i -s i 。
As a further improved technical scheme of the invention, the specific steps of the step 3 are as follows:
step 3.1, calculating Strain Difference data { d } i I=1, 2, …, mean μ and standard deviation σ of n;
wherein:
step 3.2, calculating the fluctuation range (mu-asigma, mu+asigma) of the normal value of the strain difference value according to the interval estimation method, wherein a is a set threshold range coefficient;
step 3.3, searching strain difference data { d } according to the confidence interval of the normal value of the strain difference i Data location information with i=1, 2, …, n outside the confidence interval (μ -aσ, μ+aσ), note: { d i },i＝1,2,…,n<The sequence of data positions of mu-aσ isj＝1,2,…,m 1 Wherein m is 1 The number of the data with the strain difference value smaller than the lower split value; { d i },i＝1,2,…,n>The data position sequence of μ+aσ is +.>j＝1,2,…,m 2 Wherein m is 2 The number of the data with the strain difference value larger than the upper dividing value; the set of potential outlier positions in the original strain data sequence +.>j＝1,2,…,/>j＝1,2,…,m 2 。
As a further improved technical scheme of the invention, the specific steps of the step 4 are as follows:
step 4.1, extracting long period trend data { s } of the original acceleration data by wavelet decomposition i '},i＝1,2,…,n；
Step 4.2, calculating y i And s' i Is analyzed and is marked as d' i ＝y i -s’ i 。
As a further improved technical scheme of the invention, the specific steps of the step 5 are as follows:
step 5.1, calculating the acceleration difference data { d' i I=1, 2, …, mean μ 'and standard deviation σ' of n;
wherein:
step 5.2, calculating the fluctuation range (mu '-bsigma', mu '+bsigma') of the normal value of the acceleration difference value according to a section estimation method, wherein b is a set threshold range coefficient;
step 5.3 according to accelerationConfidence interval of normal value of the difference of degree, find the data { d' i Data position information for i=1, 2, …, n outside the confidence interval (μ '-bσ', μ '+bσ'), note: { d' i },i＝1,2,…,n<The sequence of data positions of mu '-bSigma' isj＝1,2,…,n 1 Wherein n is 1 The data number is the data number of which the acceleration difference value is smaller than the lower dividing value; { d' i },i＝1,2,…,n>The data position sequence of μ '+bσ' is +.>j＝1,2,…,n 2 Wherein n is 2 The data number is the data number of which the acceleration difference value is larger than the upper dividing value; the set of potential outlier positions in the original acceleration data sequence +.>j＝1,2,…,/>j＝1,2,…,n 2 。
As a further improved technical scheme of the invention, the specific steps of the step 6 are as follows:
traversing an element in a potential abnormal point position set in the original strain data sequence, if the element is also in the potential abnormal point position set in the original acceleration data sequence, the original strain data corresponding to the element is a normal value, otherwise, the original strain data corresponding to the element is an abnormal value.
The beneficial effects of the invention are as follows: aiming at the problem that abnormal values of dynamic strain data are difficult to separate from signal fluctuation caused by vehicle load in a single data source, the invention synthesizes multiple data sources to perform information fusion, and provides a method for identifying abnormal values of strain data based on data relevance, which has the following beneficial effects:
(1) The invention can accurately identify the abnormal value in the dynamic strain data, and has less misjudgment;
(2) The invention can carry out real-time and automatic cleaning on dynamic strain data, and has wide engineering application value.
Drawings
Fig. 1 is a one-dimensional time chart of data of a second channel of a strain sensor of the day 4, month 9 of the bridge 2017 of the virulent river.
Fig. 2 is a one-dimensional time chart of data of a second channel of the acceleration sensor of the bridge 2017, 4 months and 9 days of the virulent river.
Fig. 3 is a schematic diagram of strain differential.
Fig. 4 is a graph of normal data and potential outliers in strain data.
Fig. 5 is a schematic diagram of acceleration differences.
Fig. 6 is a graph of normal data and potential outliers in acceleration data.
Fig. 7 is an outlier identification chart in strain data.
Detailed Description
The following further describes embodiments of the present invention with reference to fig. 1 to 7:
a bridge strain data abnormal value identification method based on data correlation comprises the following steps:
1) And acquiring the sensor original data installed at the same position.
The method comprises the following specific steps:
acquiring original strain data { x } acquired by strain sensors at the same location of a health monitoring system i I=1, 2, …, n and raw acceleration data { y ] acquired by the acceleration sensor i I=1, 2, …, n, where n represents the number of data.
2) And extracting long-period trend data from the original strain data through wavelet decomposition to obtain a difference value between the original strain data and the long-period trend data.
The method comprises the following specific steps:
a) Extracting long period trend data { s } of original strain data by wavelet decomposition i },i＝1,2,…,n；
b) Calculating x i And s i Is analyzed and denoted d i ＝x i -s i 。
3) And obtaining the variation range of the potential abnormal value in the original strain data according to the mean value and the standard deviation of the strain difference data.
The method comprises the following specific steps:
a) Calculating strain difference data { d } i I=1, 2, …, mean μ and standard deviation σ of n; wherein:
b) Calculating a fluctuation range (mu-aσ, mu+aσ) of a normal value of the strain difference value according to the interval estimation method, wherein a is a set threshold range coefficient; the range of fluctuation of the normal value of the strain difference may also be referred to as the normal range of fluctuation of the strain difference;
c) Searching strain difference data { d }, according to the confidence interval of the normal value of the strain difference i Data location information with i=1, 2, …, n outside the confidence interval (μ -aσ, μ+aσ), note: { d i },i＝1,2,…,n<The sequence of data positions of mu-aσ isj＝1,2,…,m 1 Wherein m is 1 The number of the data with the strain difference value smaller than the lower split value; { d i },i＝1,2,…,n>The data position sequence of μ+aσ is +.>j＝1,2,…,m 2 Wherein m is 2 The number of the data with the strain difference value larger than the upper dividing value; the set of potential outlier positions in the original strain data sequence +.>j＝1,2,…,/>j＝1,2,…,m 2 . In this embodiment, the data collected by the sensor are all recorded according to time sequence, and there are formatted data positions, whichThe position in the embodiment refers to the data position when the abnormal point appears, that is, the time when the abnormal point appears is recorded correspondingly.
4) And extracting long-period trend data from the original acceleration data through wavelet decomposition to obtain a difference value between the original acceleration data and the long-period trend data.
The method comprises the following specific steps:
a) Extracting long period trend data { s 'of original acceleration data by wavelet decomposition' i },i＝1,2,…,n；
b) Calculating y i And s' i Is analyzed and is marked as d' i ＝y i -s’ i 。
5) And obtaining the variation range of the potential abnormal value in the original acceleration data according to the mean value and the standard deviation of the acceleration difference data.
The method comprises the following specific steps:
a) Calculating acceleration difference data { d } i ' i=1, 2, …, mean μ ' and standard deviation σ ' of n; wherein:
b) Calculating a fluctuation range (mu '-bσ', mu '+bσ') of a normal value of the acceleration difference value according to the interval estimation method, wherein b is a set threshold range coefficient; the range of fluctuation of the normal value of the acceleration difference value may also be referred to as the normal range of fluctuation of the acceleration difference value;
c) Searching acceleration difference data { d 'according to the confidence interval of the normal value of the acceleration difference' i Data position information for i=1, 2, …, n outside the confidence interval (μ '-bσ', μ '+bσ'), note: { d' i },i＝1,2,…,n<The sequence of data positions of mu '-bSigma' isj＝1,2,…,n 1 Wherein n is 1 The data number is the data number of which the acceleration difference value is smaller than the lower dividing value; { d' i },i＝1,2,…,n>The data position sequence of μ '+bσ' is +.>j＝1,2,…,n 2 Wherein n is 2 The data number is the data number of which the acceleration difference value is larger than the upper dividing value; the set of potential outlier positions in the original acceleration data sequence +.>j＝1,2,…,j＝1,2,…,n 2 。
And 6, comparing the appearance position of the potential abnormal value in the original strain data with the appearance position of the potential abnormal value in the original acceleration data, and judging and identifying the real abnormal value of the original strain data.
The method comprises the following specific steps: judging whether the potential abnormal value position in the original strain data also has the same abnormal point position in the original acceleration data, namely:
traversing elements in the set of potential outlier positions P in the original strain data sequence, noting: if P epsilon P, the original strain data corresponding to the element is normal, otherwise, the original strain data corresponding to the element is abnormal.
The following describes the implementation process of the invention by taking a large bridge of a virulent river as an example:
the original strain data in this embodiment is data of a second channel of a strain sensor of 2017, 4, 9 days of the bridge of the Sharp river, the original acceleration data is data of a second channel of the acceleration sensor at the same position, sampling frequency is 200Hz, and the total of 17280000 data, fig. 1 shows a one-dimensional time course curve of the original strain data, and fig. 2 shows a one-dimensional time course curve of the original acceleration data. Next, the original strain data is divided into a long period signal and a high frequency signal by wavelet decomposition, wherein the high frequency signal is a strain difference, and a schematic diagram thereof is shown in fig. 3. The long period signal is mainly caused by temperature effect and has positive correlation with temperature; the high-frequency signal is mainly generated by the load of the vehicle, but also contains data outliers caused by certain problems of the sensor or external interference.
The average value of the strain difference values is calculated to be 0.0021, the standard deviation is 1.4049, the threshold range parameter is set to be 8, the normal fluctuation interval of the strain difference values is (-11.2372,11.2415), and then 60 potential abnormal values of the original strain data are obtained. The distribution of the potential outliers of the raw strain data is shown in fig. 4.
Then, wavelet decomposition is performed on the original acceleration data, and the long period trend is extracted, so that the difference value between the original acceleration data and the long period trend data is obtained, as shown in fig. 5. The average value of the acceleration difference values is 0, the standard deviation is 6.8925, the threshold range parameter is set to be 8, the normal fluctuation interval (-55.1397,55.1397) of the acceleration difference values is obtained, and then 115 potential abnormal values of the original acceleration data are obtained. The distribution of the potential outliers of the raw acceleration data is shown in fig. 6.
Comparing the appearance position of the potential abnormal value of the strain with the appearance position of the potential abnormal value of the acceleration, and judging and identifying the real abnormal value of the original strain data. When searching the same position of the potential jump point value of the original strain data and the original acceleration data, taking the potential jump point position of the strain data as p into consideration that a certain response error exists in the bridge structure and the influence of a sensor exists i This section considers potential jump points of acceleration data [ p ] i -2,p i +2]The same position is the position. If the potential jump points of the two points are the same, the strain data corresponding to the points are normal and are not processed; otherwise, the strain data corresponding to the point is an outlier. After the comparative identification, the final original strain data outlier identification is shown in fig. 7, coexisting at 4 identical locations, i.e., 4 normal data.
The above calculation example shows that the method provided by the invention can effectively identify the abnormal value in the strain data, has less misjudgment on the abnormal value, can clean the abnormal value in the dynamic strain data in real time and automatically, and can be applied to the preprocessing of the actual engineering health monitoring data.
The scope of the present invention includes, but is not limited to, the above embodiments, and any alterations, modifications, and improvements made by those skilled in the art are intended to fall within the scope of the invention.
Claims (5)
1. The bridge strain data abnormal value identification method based on the data relevance is characterized by comprising the following steps of:
step 1, acquiring original data of sensors installed at the same position;
step 2, extracting long-period trend data from the original strain data through wavelet decomposition to obtain a difference value between the original strain data and the long-period trend data;
step 3, obtaining the variation range of potential abnormal values in the original strain data according to the mean value and standard deviation of the strain difference data obtained in the step 2;
step 4, extracting long-period trend data from the original acceleration data through wavelet decomposition to obtain a difference value between the original acceleration data and the long-period trend data;
step 5, obtaining the variation range of the potential abnormal value in the original acceleration data according to the mean value and standard deviation of the acceleration difference value data obtained in the step 4;
step 6, comparing the appearance position of the potential abnormal value in the original strain data with the appearance position of the potential abnormal value in the original acceleration data, and judging and identifying the real abnormal value of the original strain data;
the specific steps of the step 3 are as follows:
step 3.1, calculating Strain Difference data { d } i I=1, 2, …, mean μ and standard deviation σ of n;
wherein:
step 3.2, calculating the fluctuation range (mu-asigma, mu+asigma) of the normal value of the strain difference value according to the interval estimation method, wherein a is a set threshold range coefficient;
step 3.3,Searching strain difference data { d }, according to the confidence interval of the normal value of the strain difference i Data location information with i=1, 2, …, n outside the confidence interval (μ -aσ, μ+aσ), note: { d i },i＝1,2,…,n<The sequence of data positions of mu-aσ isWherein m is 1 The number of the data with the strain difference value smaller than the lower split value; { d i },i＝1,2,…,n>The data position sequence of μ+aσ is +.>Wherein m is 2 The number of the data with the strain difference value larger than the upper dividing value; the set of potential outlier positions in the original strain data sequence +.>
The specific steps of the step 5 are as follows:
step 5.1, calculating the acceleration difference data { d' i I=1, 2, …, mean μ 'and standard deviation σ' of n;
wherein:
step 5.2, calculating the fluctuation range (mu '-bsigma', mu '+bsigma') of the normal value of the acceleration difference value according to a section estimation method, wherein b is a set threshold range coefficient;
step 5.3, searching the acceleration difference data { d 'according to the confidence interval of the normal value of the acceleration difference' i Data position information for i=1, 2, …, n outside the confidence interval (μ '-bσ', μ '+bσ'), note: { d' i },i＝1,2,…,n<The sequence of data positions of mu '-bSigma' isWherein n is 1 Is the difference of accelerationThe number of data smaller than the lower split value; { d' i },i＝1,2,…,n>The data position sequence of μ '+bσ' is +.>Wherein n is 2 The data number is the data number of which the acceleration difference value is larger than the upper dividing value; then a set of potential outlier positions in the raw acceleration data sequence
2. The method for identifying abnormal values of bridge strain data based on data correlation as set forth in claim 1, wherein the specific steps of step 1 are as follows:
acquiring original strain data { x } acquired by strain sensors at the same location of a health monitoring system i I=1, 2, …, n and raw acceleration data { y ] acquired by the acceleration sensor i I=1, 2, …, n, where n represents the number of data.
3. The method for identifying abnormal values of bridge strain data based on data correlation as set forth in claim 1, wherein the specific steps of step 2 are as follows:
step 2.1, extracting long period trend data { s } of the original strain data by wavelet decomposition i },i＝1,2,…,n；
Step 2.2, calculating x i And s i Is analyzed and denoted d i ＝x i -s i 。
4. The method for identifying abnormal values of bridge strain data based on data correlation as set forth in claim 1, wherein the specific steps of step 4 are as follows:
step 4.1, extracting long period trend data { s 'of the original acceleration data by wavelet decomposition' i },i＝1,2,…,n；
Step 4.2, calculating y i And s' i Is analyzed for differences in (a)Is marked as d' i ＝y i -s′ i 。
5. The method for identifying abnormal values of bridge strain data based on data correlation as claimed in claim 1, wherein the specific steps of the step 6 are as follows:
traversing an element in a potential abnormal point position set in the original strain data sequence, if the element is also in the potential abnormal point position set in the original acceleration data sequence, the original strain data corresponding to the element is a normal value, otherwise, the original strain data corresponding to the element is an abnormal value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910500098.XA CN110287827B (en) | 2019-06-11 | 2019-06-11 | Bridge strain data outlier identification method based on data correlation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910500098.XA CN110287827B (en) | 2019-06-11 | 2019-06-11 | Bridge strain data outlier identification method based on data correlation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110287827A CN110287827A (en) | 2019-09-27 |
| CN110287827B true CN110287827B (en) | 2023-09-26 |
Family
ID=68003602
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910500098.XA Active CN110287827B (en) | 2019-06-11 | 2019-06-11 | Bridge strain data outlier identification method based on data correlation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110287827B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111189488B (en) * | 2019-12-13 | 2020-12-04 | 精英数智科技股份有限公司 | Sensor value abnormity identification method, device, equipment and storage medium |
| CN111707782A (en) * | 2020-03-30 | 2020-09-25 | 江苏方天电力技术有限公司 | Thermal power generating unit carbon dioxide emission concentration abnormity detection method based on oxygen amount |
| CN113866455A (en) * | 2021-09-30 | 2021-12-31 | 中铁桥隧技术有限公司 | Bridge acceleration monitoring data anomaly detection method, system and device based on deep learning |
| CN116308305B (en) * | 2023-05-25 | 2023-08-08 | 安徽省云鹏工程项目管理有限公司 | Bridge health monitoring data management system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102661716A (en) * | 2012-04-20 | 2012-09-12 | 武汉理工大学 | Method and system for detecting linearity and rigidity of bridges and tunnels on basis of fiber-optic gyroscope technology |
| CN105760934A (en) * | 2016-03-02 | 2016-07-13 | 浙江工业大学 | Bridge abnormity monitoring restoration method based on wavelet and BP neural network |
| CN107230150A (en) * | 2017-07-19 | 2017-10-03 | 西南大学 | A kind of method for detecting abnormality in stock market based on Non-negative Matrix Factorization |
| CN108170925A (en) * | 2017-12-21 | 2018-06-15 | 中国铁道科学研究院铁道建筑研究所 | A kind of bridge damnification method for early warning based on arma modeling |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9069608B2 (en) * | 2013-03-06 | 2015-06-30 | Vmware, Inc. | Method and system for providing a roaming remote desktop |
-
2019
- 2019-06-11 CN CN201910500098.XA patent/CN110287827B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102661716A (en) * | 2012-04-20 | 2012-09-12 | 武汉理工大学 | Method and system for detecting linearity and rigidity of bridges and tunnels on basis of fiber-optic gyroscope technology |
| CN105760934A (en) * | 2016-03-02 | 2016-07-13 | 浙江工业大学 | Bridge abnormity monitoring restoration method based on wavelet and BP neural network |
| CN107230150A (en) * | 2017-07-19 | 2017-10-03 | 西南大学 | A kind of method for detecting abnormality in stock market based on Non-negative Matrix Factorization |
| CN108170925A (en) * | 2017-12-21 | 2018-06-15 | 中国铁道科学研究院铁道建筑研究所 | A kind of bridge damnification method for early warning based on arma modeling |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110287827A (en) | 2019-09-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110287827B (en) | Bridge strain data outlier identification method based on data correlation | |
| EP3588222B1 (en) | Time-series data processing device, time-series data processing system, and time-series data processing method | |
| CN106888205B (en) | Non-invasive PLC anomaly detection method based on power consumption analysis | |
| US9146800B2 (en) | Method for detecting anomalies in a time series data with trajectory and stochastic components | |
| US10852357B2 (en) | System and method for UPS battery monitoring and data analysis | |
| US20110213577A1 (en) | Method and apparatus for analyzing waveform signals of a power system | |
| CN116992393B (en) | Safety production monitoring method based on industrial Internet of things | |
| CN116308305B (en) | Bridge health monitoring data management system | |
| CN115358721B (en) | Engineering supervision information supervision system and method based on big data | |
| CN105205113A (en) | System and method for excavating abnormal change process of time series data | |
| CN106247173A (en) | The method and device of pipeline leakage testing | |
| US20220179393A1 (en) | Machine tool evaluation method, machine tool evaluation system and medium | |
| US20210116905A1 (en) | Cause determination of anomalous events | |
| CN110598265A (en) | Bridge health data abnormity correction method and system based on wavelet analysis | |
| CN112861350A (en) | Temperature overheating defect early warning method for stator winding of water-cooled steam turbine generator | |
| CN115165079A (en) | Equipment alarm method based on multiple sound characteristic analysis | |
| CN114997313A (en) | Anomaly detection method for ocean online monitoring data | |
| CN117314020B (en) | Wetland carbon sink data monitoring system of plankton | |
| CN108508860B (en) | Process industrial production system data monitoring method based on coupling relation | |
| CN108333443B (en) | Method for alarming intermittent defects of power equipment | |
| CN112882898A (en) | Anomaly detection method, system, device and medium based on big data log analysis | |
| CN112528227A (en) | Sensor abnormal data identification method based on mathematical statistics | |
| CN115310499B (en) | Industrial equipment fault diagnosis system and method based on data fusion | |
| CN112486096A (en) | Machine tool operation state monitoring method | |
| CN114429148A (en) | Power equipment state detection method based on multi-source data fusion |
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 | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20210423 Address after: 200120 building C, No. 888, Huanhu West 2nd Road, Lingang New Area, China (Shanghai) pilot Free Trade Zone, Pudong New Area, Shanghai Applicant after: Shanghai Shenwu Intelligent Technology Co.,Ltd. Address before: 211200 Jiangsu, Nanjing, Lishui District, Yong Yang Street, Qinhuai Road, 288, Xing Zhuang science and Technology Industrial Park Applicant before: Nanjing Ruiyong Zhi Operations and Maintenance Engineering Technology Co.,Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |