CN106570259B - A kind of elimination of rough difference method of dam displacement data - Google Patents
A kind of elimination of rough difference method of dam displacement data Download PDFInfo
- Publication number
- CN106570259B CN106570259B CN201610957664.6A CN201610957664A CN106570259B CN 106570259 B CN106570259 B CN 106570259B CN 201610957664 A CN201610957664 A CN 201610957664A CN 106570259 B CN106570259 B CN 106570259B
- Authority
- CN
- China
- Prior art keywords
- measured value
- rough
- sequence
- criterion
- doubtful
- 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
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/13—Architectural design, e.g. computer-aided architectural design [CAAD] related to design of buildings, bridges, landscapes, production plants or roads
Abstract
The invention discloses a kind of elimination of rough difference methods of dam displacement data, set Dixon criterion, Grubbs criterion, Chauvenet criterion, quartile criterion these four outlier decision criterias weight coefficient, and set rough error decision threshold, then determined respectively using above four rules for each measured value, if the judgement result under certain criterion is doubtful rough error, the corresponding weight coefficient of the criterion that then adds up, finally obtain the measured value weight coefficient and, with rough error decision threshold compare, up to or over i.e. determine the measured value for rough error.A possibility that present invention has comprehensively considered a variety of outlier data identification methods, has weighed erroneous judgement and has failed to judge provides a kind of scientific and rational method to reject the rough error in dam displacement data.
Description
Technical field
The present invention relates to a kind of elimination of rough difference methods of dam displacement data, belong to dam safety monitoring technical field.
Background technique
Current China dam safety monitoring is totally in the special stage, and many old dams will reach design reference period, and
Large quantities of new dams just put into operation, therefore carry out prediction to dam safety monitoring index to provide safety monitoring decision assistant, just show
It obtains necessary.The displacement of dam is then one of the main indicator of dam safety monitoring, is occupied in the assessment of dam integrity state
Important meaning, abnormal value more particularly should be paid attention to and be analyzed.And drawn by factors such as measuring system and manual operations
The exceptional value risen, should be regarded as rough error and be rejected, to avoid subsequent analysis and evaluation operation is interfered.
So-called rough error refers to that in the time series of dam displacement amount, there are notable differences with the measured value of adjacent time
Catastrophe point, variation of the origin cause of formation generally with the true condition of dam and environment parameter is without direct correlation.It is universal thick in industry at present
Poor elimination method mainly includes method of expertise and statistical model method.Method of expertise relies primarily on hydraulic safety monitoring expert
Engineering experience determine rough error, more dependence Subjective;Statistical model method depends on a certain strategy determined, example
Such as according to extreme value, luffing, the size with the residual error of match value, to data using not comprehensive enough with deeply, there are biggish
The risk judged by accident and failed to judge.
Summary of the invention
The technical problem to be solved by the present invention is to overcome the deficiencies of existing technologies, a kind of the thick of dam displacement data is provided
Poor elimination method has comprehensively considered a variety of outlier data identification methods, rejects the rough error in dam displacement data.
In order to solve the above technical problems, the present invention provides a kind of elimination of rough difference method of dam displacement data, including following
Step:
1) Dixon criterion, Grubbs criterion, Chauvenet criterion, these four outliers of quartile criterion are set and determine standard
Weight coefficient then, and setting rough error decision threshold;
2) mean value and standard deviation in situ for moving measured value sequence is calculated, and residual sequence is calculated based on regression model;
3) doubtful rough error is determined according to Dixon criterion;
4) doubtful rough error is determined according to Grubbs criterion;
5) doubtful rough error is determined according to Chauvenet criterion;
6) doubtful rough error is determined according to quartile criterion;
7) for each measured value, 4 doubtful rough errors determine as a result, by the weight for the criterion that result is doubtful rough error
Coefficient is added, then compared with rough error decision threshold, up to or over determining the measured value for rough error.
In step 1) above-mentioned, the value range of the weight coefficient of every kind of outlier decision criteria is 1-100, four weights
The summation of coefficient is 100.
In step 2 above-mentioned, residual sequence is calculated based on regression model and is referred to, according to the dam horizontal displacement basis origin cause of formation
Theory determines the factor and displacement data for influencing dam displacement, then carries out regression fit using least square method, obtains
Regression model and corresponding fitting value sequence, and move measured value sequence work difference with original position and move the residual of measured value sequence to calculate original position
Difference sequence.
In step 3) above-mentioned, determines that doubtful rough error refers to according to Dixon criterion and move measured value sequence ascending order row in situ
Sequence obtains new sequence X, remembers D=(Xn-Xn-2)/ (Xn-X3), for k-th of measured value in sequence, if meeting D(k,α)> D, i.e.,
Determine that the measured value is doubtful rough error, wherein n is the total length in situ for moving measured value sequence, XnFor n-th of numerical value of sequence X, α is
Significance.
Level of significance α value above-mentioned is 0.01.
In step 4) above-mentioned, determine that doubtful rough error refers to residual error V corresponding for each measured value according to Grubbs criterion,
Meet when with the standard deviation S in situ for moving measured value sequence: | V | >=G(n,a)* when S, that is, determine that the measured value is doubtful rough error, wherein n
For the total length in situ for moving measured value sequence, a is significance.
Significance a value above-mentioned is 0.05.
In step 5) above-mentioned, it is corresponding for each measured value residual to determine that doubtful rough error refers to according to Chauvenet criterion
Poor V meets when with the standard deviation S in situ for moving measured value sequence: | V | >=Z(n)* when S, that is, determine that the measured value is doubtful rough error,
In, n is the total length in situ for moving measured value sequence.
In step 6) above-mentioned, determines that doubtful rough error refers to according to quartile criterion and residual sequence is carried out according to ascending order
Q is remembered in sequence1For new sequence intermediate value, Q2For new sequence head value to Q1Intermediate value, Q3For Q1To the intermediate value of new sequence end value, then each
The corresponding residual error V of measured value meets: V >=4* Q3 - Q1Or V≤4* Q1 - 3* Q3When, that is, determine that the measured value is doubtful
Rough error.
Advantageous effects of the invention:
A possibility that this law has comprehensively considered a variety of outlier data identification methods, has weighed erroneous judgement and has failed to judge, to reject
Rough error in dam displacement data provides a kind of scientific and rational method, so that monitoring materials can more accurately react
The actual work condition of dam provides more reliable decision assistant for the operational management of dam safety.
Detailed description of the invention
Fig. 1 is the method for the present invention flow chart.
Specific embodiment
The invention will be further described below in conjunction with the accompanying drawings.Following embodiment is only used for clearly illustrating the present invention
Technical solution, and not intended to limit the protection scope of the present invention.
As shown in Figure 1, the elimination of rough difference method of dam displacement data of the invention, comprising the following steps:
1, the weight coefficient and rough error decision threshold of four criterion are set
Set Dixon criterion, Grubbs criterion, Chauvenet criterion, quartile criterion these four outlier decision criterias
Weight coefficient, the value range of the weight coefficient of each criterion is 1-100, and the summation of four weight coefficients is 100.
Set judge a measured value as doubtful big error judgment threshold value, threshold range 1-100, in this way, for each survey
Value, if the judgement result under certain criterion is doubtful rough error, the corresponding weight coefficient of the criterion that adds up finally obtains this
The weight coefficient of measured value and, with rough error decision threshold compare, up to or over as rough error.
The weight coefficient of decision criteria and the setting mechanism of rough error threshold value, can be flexibly according to the data characteristic of actual sample
The weight coefficient of each rule is adjusted with quality, and can preferably combine expertise and historical experience, and it is right to realize the present invention
In different engineerings in ground ability.
2, the mean value of Displacement Sequence, standard deviation, and the residual sequence based on regression model are calculated
Firstly, moving measured value sequence according in situ, its mean value and standard deviation are calculated.Secondly, determining that the sample of regression model is empty
Between, according to dam horizontal displacement basis formation theory, the factor for influencing dam displacement includes reservoir level H, temperature T, timeliness t etc.,
The factor that the present invention chooses is to be averaged the same day to yesterday, be averaged within first 2 to 5 days, first 6 to 15 days average H, H2、H3、H4, T, temperature
Harmonic factor sin (2 π t/365), cos (2 π t/365), sin (4 π t/365), cos (4 π t/365) and t, ㏑ t amount to 21
A factor, effect quantity are displacement data, carry out regression fit using least square method, obtain regression model and corresponding quasi-
Value sequence is closed, and moves measured value sequence with original position and makees difference to calculate the residual sequence that measured value sequence is moved in original position.Finally, relative to
The difference of normal value, rough error and its match value is more significant, therefore can determine whether measured value is rough error according to residual error.
3, doubtful rough error is determined according to Dixon criterion
Measured value sequence ascending sort is moved in situ, new sequence X is obtained, remembers D=(Xn-Xn-2)/ (Xn-X3), for sequence
K-th of measured value in column, if meeting D(k,α)> D determines that the measured value is doubtful rough error.Wherein, n is shifting measured value sequence in situ
Total length, XnFor n-th of numerical value of sequence X, α is significance, and the present invention uses 0.01, D(k,α)It is examined by looking into Dixon
Tables of critical values can obtain.
4, doubtful rough error is determined according to Grubbs criterion
Residual error V corresponding for each measured value meets when with the standard deviation S in situ for moving measured value sequence: | V | >=G(n,a)*S
When, that is, determine that the measured value is doubtful rough error.Wherein, n is the total length in situ for moving measured value sequence, and a is significance, the present invention
Using 0.05, G(n,a)Value is then obtained by looking into Grubbs tables of critical values.
5, doubtful rough error is determined according to Chauvenet criterion
Residual error V corresponding for each measured value meets when with the standard deviation S in situ for moving measured value sequence: | V | >=Z(n)*S
When, that is, determine that the measured value is doubtful rough error.Wherein, n is the total length in situ for moving measured value sequence, Z(n)Pass through Cha Xiaoweile coefficient
Biao Ke get.
6, doubtful rough error is determined according to quartile rule
Residual sequence is ranked up according to ascending order, remembers Q1For new sequence intermediate value, Q2For new sequence head value to Q1In
Value, Q3For Q1To the intermediate value of new sequence end value, then the corresponding residual error V of each measured value, meets: V >=4* Q3 - Q1Or V≤
4* Q1 - 3* Q3When, that is, determine that the measured value is doubtful rough error.
7, the judgement result aggregative weighted of four kinds of criterion
Finally, for each measured value, there are 4 doubtful rough errors to determine as a result, by criterion that result is doubtful rough error
Weight coefficient is added, then compared with rough error decision threshold, up to or over determining the measured value for rough error.
The above is only a preferred embodiment of the present invention, it is noted that for the ordinary skill people of the art
For member, without departing from the technical principles of the invention, several improvement and deformations can also be made, these improvement and deformations
Also it should be regarded as protection scope of the present invention.
Claims (9)
1. a kind of elimination of rough difference method of dam displacement data, which comprises the following steps:
1) Dixon criterion, Grubbs criterion, Chauvenet criterion, quartile criterion these four outlier decision criterias are set
Weight coefficient, and setting rough error decision threshold;
2) mean value and standard deviation in situ for moving measured value sequence is calculated, and residual sequence is calculated based on regression model;
3) doubtful rough error is determined according to Dixon criterion;
4) doubtful rough error is determined according to Grubbs criterion;
5) doubtful rough error is determined according to Chauvenet criterion;
6) doubtful rough error is determined according to quartile criterion;
7) for each measured value, 4 doubtful rough errors determine as a result, by the weight coefficient for the criterion that result is doubtful rough error
It is added, then compared with rough error decision threshold, up to or over determining the measured value for rough error.
2. a kind of elimination of rough difference method of dam displacement data according to claim 1, which is characterized in that the step 1)
In, the value range of the weight coefficient of every kind of outlier decision criteria is 1-100, and the summation of four weight coefficients is 100.
3. a kind of elimination of rough difference method of dam displacement data according to claim 1, which is characterized in that the step 2
In, residual sequence is calculated based on regression model and is referred to, according to dam horizontal displacement basis formation theory, determining influences dam displacement
The factor and displacement data, then using least square method carry out regression fit, obtain regression model and corresponding fitting
Value sequence, and and the poor residual sequence to calculate shifting measured value sequence in situ of shifting measured value sequence work in situ.
4. a kind of elimination of rough difference method of dam displacement data according to claim 1, which is characterized in that the step
3) in, determine that doubtful rough error refers to according to Dixon criterion and move measured value sequence ascending sort in situ, obtain new sequence X, remember
D=(Xn-Xn-2)/ (Xn-X3), for k-th of measured value in sequence, if meeting D(k,α)> D determines that the measured value is doubtful thick
Difference, wherein n is the total length in situ for moving measured value sequence, XnFor n-th of numerical value of sequence X, α is significance.
5. a kind of elimination of rough difference method of dam displacement data according to claim 4, which is characterized in that the conspicuousness
Horizontal α value is 0.01.
6. a kind of elimination of rough difference method of dam displacement data according to claim 3, which is characterized in that the step
4) in, determine that doubtful rough error refers to residual error V corresponding for each measured value according to Grubbs criterion, when moving measured value sequence in situ
Standard deviation S meet: | V | >=G(n,a)* when S, that is, determine that the measured value is doubtful rough error, wherein n is shifting measured value sequence in situ
Total length, a are significance.
7. a kind of elimination of rough difference method of dam displacement data according to claim 6, which is characterized in that the conspicuousness
Horizontal a value is 0.05.
8. a kind of elimination of rough difference method of dam displacement data according to claim 3, which is characterized in that the step
5) in, determine that doubtful rough error refers to residual error V corresponding for each measured value according to Chauvenet criterion, when moving measured value in situ
The standard deviation S of sequence meets: | V | >=Z(n)* when S, that is, determine that the measured value is doubtful rough error, wherein n is shifting measured value sequence in situ
The total length of column.
9. a kind of elimination of rough difference method of dam displacement data according to claim 3, which is characterized in that the step
6) in, determine that doubtful rough error refers to according to quartile criterion and residual sequence is ranked up according to ascending order, remember Q1For new sequence
Intermediate value, Q2For new sequence head value to Q1Intermediate value, Q3For Q1To the intermediate value of new sequence end value, then the corresponding residual error V of each measured value,
Meet: V >=4* Q3 - Q1Or V≤4* Q1 - 3* Q3When, that is, determine that the measured value is doubtful rough error.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610957664.6A CN106570259B (en) | 2016-11-03 | 2016-11-03 | A kind of elimination of rough difference method of dam displacement data |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610957664.6A CN106570259B (en) | 2016-11-03 | 2016-11-03 | A kind of elimination of rough difference method of dam displacement data |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106570259A CN106570259A (en) | 2017-04-19 |
CN106570259B true CN106570259B (en) | 2019-08-23 |
Family
ID=58535768
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610957664.6A Active CN106570259B (en) | 2016-11-03 | 2016-11-03 | A kind of elimination of rough difference method of dam displacement data |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106570259B (en) |
Families Citing this family (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107229594B (en) * | 2017-05-05 | 2020-11-17 | 深圳市建设综合勘察设计院有限公司 | Multi-beam sounding data trend surface filtering method and system |
CN107862338B (en) * | 2017-11-11 | 2021-07-02 | 四创科技有限公司 | Marine environment monitoring data quality management method and system based on double inspection method |
CN108319664A (en) * | 2018-01-17 | 2018-07-24 | 国电南瑞科技股份有限公司 | A kind of dam and the recognition methods of project security monitoring data error and system |
CN108245154B (en) * | 2018-01-24 | 2020-10-09 | 福州大学 | Method for accurately determining blink interval in electroencephalogram or electrooculogram by using abnormal value detection |
CN109783939A (en) * | 2019-01-16 | 2019-05-21 | 武汉楚云端信息科技有限责任公司 | A kind of data processing method of combination Grubbs method and 3 σ methods |
CN110320434B (en) * | 2019-07-03 | 2020-09-25 | 山东大学 | High-resistance fault identification method and system based on zero-sequence current waveform interval slope curve |
CN111177218B (en) * | 2019-12-25 | 2022-08-30 | 深圳市东深电子股份有限公司 | Dam safety analysis method based on big data analysis |
CN111508216B (en) * | 2020-04-28 | 2021-12-03 | 水利部交通运输部国家能源局南京水利科学研究院 | Intelligent early warning method for dam safety monitoring data |
CN111709465B (en) * | 2020-06-04 | 2022-12-20 | 中国电建集团华东勘测设计研究院有限公司 | Intelligent identification method for rough difference of dam safety monitoring data |
CN111896842A (en) * | 2020-07-27 | 2020-11-06 | 国网上海市电力公司 | Power distribution network arc high-resistance fault section positioning method based on interval slope |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2011209900A (en) * | 2010-03-29 | 2011-10-20 | Nippon Telegr & Teleph Corp <Ntt> | Data center network design method and program |
CN104331841A (en) * | 2014-10-29 | 2015-02-04 | 国家电网公司 | Orderly power utilization state monitoring method based on AHP (analytic hierarchy process) and Delphi method |
CN104462808A (en) * | 2014-12-04 | 2015-03-25 | 河海大学 | Method for fitting safe horizontal displacement and dynamic data of variable sliding window of water level |
CN105930870A (en) * | 2016-04-26 | 2016-09-07 | 中国电建集团昆明勘测设计研究院有限公司 | Engineering safety monitoring data outlier detection method |
-
2016
- 2016-11-03 CN CN201610957664.6A patent/CN106570259B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2011209900A (en) * | 2010-03-29 | 2011-10-20 | Nippon Telegr & Teleph Corp <Ntt> | Data center network design method and program |
CN104331841A (en) * | 2014-10-29 | 2015-02-04 | 国家电网公司 | Orderly power utilization state monitoring method based on AHP (analytic hierarchy process) and Delphi method |
CN104462808A (en) * | 2014-12-04 | 2015-03-25 | 河海大学 | Method for fitting safe horizontal displacement and dynamic data of variable sliding window of water level |
CN105930870A (en) * | 2016-04-26 | 2016-09-07 | 中国电建集团昆明勘测设计研究院有限公司 | Engineering safety monitoring data outlier detection method |
Non-Patent Citations (3)
Title |
---|
大坝安全监测数据粗差识别方法的比较与改进;李啸啸 等;《中国农村水利水电》;20110315;第102-105页 |
大坝自动化监测数据粗差处理方法研究;许贝贝 等;《测绘地理信息》;20150316;第59-61页 |
粗大误差四种判别准则的比较和应用;熊艳艳 等;《大学物理实验》;20100226;第66-68页 |
Also Published As
Publication number | Publication date |
---|---|
CN106570259A (en) | 2017-04-19 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106570259B (en) | A kind of elimination of rough difference method of dam displacement data | |
CN111508216B (en) | Intelligent early warning method for dam safety monitoring data | |
CN106447210B (en) | A kind of distribution net equipment health degree dynamic diagnosis method of meter and trust evaluation | |
CN104677997B (en) | A kind of transformer oil chromatographic on-line monitoring differentiation method for early warning | |
CN104462808B (en) | Level of security displacement and the slip variable window dynamic data approximating method of water level | |
CN105608842B (en) | A kind of damaged online monitoring alarm device of nuclear reactor fuel | |
CN108335021A (en) | A kind of method and maintenance decision optimization of wind energy conversion system state Reliability assessment | |
CN113762604B (en) | Industrial Internet big data service system | |
CN105005822A (en) | Optimal step length and dynamic model selection based ultrahigh arch dam response prediction method | |
CN104456092A (en) | Multidimensional assessment method of petroleum and natural gas pipeline warning priority | |
CN107657121B (en) | Aircraft structure performance prediction processing method and system based on corrosion level evaluation | |
CN113361958A (en) | Defect early warning method and system | |
CN107103425B (en) | Intelligent energy evaluation system for power generation equipment running state computer | |
Saputra et al. | Quality improvement of molding machine through statistical process control in plastic industry | |
CN108684051A (en) | A kind of wireless network performance optimization method, electronic equipment and storage medium based on cause and effect diagnosis | |
CN109559019B (en) | Power system dynamic security assessment method based on risk index | |
KR102041683B1 (en) | A method for defects | |
CN117212050A (en) | Wind turbine generator system headroom prediction control method | |
KR101945131B1 (en) | Method and Apparatus for Managing Very Small Fraction of Nonconforming under Non-Normal Process | |
CN111639813A (en) | Deep learning-based slag disposal site risk early warning method and system | |
CN106053984B (en) | The method for determining the distribution of antiskid brake control device high temperature failure | |
CN111177218B (en) | Dam safety analysis method based on big data analysis | |
CN110991487B (en) | Integrated coupling model generation method for multi-source monitoring detection data | |
Zhang et al. | Treatment of errors in dam safety monitoring data | |
CN106059065A (en) | Safety and stability adaptive emergency control decision method based on in-value control measures |
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 |