CN105893329B - A kind of tide gauge consistent correction method based on moon yardstick - Google Patents
A kind of tide gauge consistent correction method based on moon yardstick Download PDFInfo
- Publication number
- CN105893329B CN105893329B CN201610249068.2A CN201610249068A CN105893329B CN 105893329 B CN105893329 B CN 105893329B CN 201610249068 A CN201610249068 A CN 201610249068A CN 105893329 B CN105893329 B CN 105893329B
- Authority
- CN
- China
- Prior art keywords
- point
- tide gauge
- value
- tidal level
- moon
- 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
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/14—Fourier, Walsh or analogous domain transformations, e.g. Laplace, Hilbert, Karhunen-Loeve, transforms
- G06F17/148—Wavelet transforms
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/10—Interfaces, programming languages or software development kits, e.g. for simulating neural networks
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computing Systems (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Operations Research (AREA)
- Probability & Statistics with Applications (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Other Liquid Machine Or Engine Such As Wave Power Use (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention relates to a kind of tide gauge consistent correction method based on moon yardstick, establish the relation between actual measurement tidal level series and trend term, periodic term and random factor item, by the expectation for contrasting the mould of coefficient of wavelet decomposition under two wavelet scales of residual error and calculating that different scale small echo is examined, the position that abnormity point and catastrophe point occur is determined, the interference of exceptional value elimination random factor is replaced with average value.The interference for eliminating periodic term is handled by moving average.Dependency relation is set up to the tide gauge and field data after trend term, elimination periodic term, nearly 5 days tide gauge training BP neural network is chosen, the network calculations obtained with training obtain correction value.The problem of solving the tide gauge consistent correction of moon yardstick, eliminates the interference of random factor during amendment, calculates revised tide gauge by the BP neural network model of training, improves the reliability and precision of amendment.
Description
Technical field
Field is revised the present invention relates to hydrological data, is repaiied more particularly, to a kind of tide gauge uniformity based on moon yardstick
Correction method.
Background technology
When carrying out related hydraulic engineering construction planning, generally we think the data such as water level, tidal level, rainfall institute
The environment of generation is constant, so as to regard the hydrographic features sequence that we are studied as simple stochastic variable, it influences
The uniformity of the factor causes data system also to have relevant coherency accordingly, is easy to our research.
For the tide gauge of each sea outfall door, with coastal area expanding economy, mankind's activity has a strong impact on
The land surface condition of tidal level formation, especially for sanidal improvement project etc., above-mentioned influence is changed in tide gauge
Correlation.Carrying out the hydraulic condition of river simulation in tidal level influence area or river course calculation and the tidal level to non-avaible mouthful door
When carrying out prediction research, accurate tide gauge is particularly important.
Current research carries out general amendment mainly on year yardstick to tide gauge；During amendment not
Mutation value or exceptional value to series are handled；Amendment series is fitted by least square method can not be anti-well
Mirror the relation between tide gauge variation tendency and measured value.
In a word, tide gauge tool is of great significance, either the River Construction of tidal reach, or city is anti-
Flood, a set of tide level data after amendment can help us more accurate forecast to go out the water level being likely to occur in river course,
Accomplish to prevent trouble before it happens.The method being modified at present to tidal level series is seldom, also fairly simple, often both for year tidal level
Data is come what is be modified, and this method applicability preferably, obtains revised moon tide gauge, can carry out regimen with upper river
Condition carries out more preferable combination research.
The content of the invention
The present invention devises a kind of tide gauge consistent correction method based on moon yardstick, and its technical problem solved is
Carry out general amendment, mutation value or exception during amendment not to series to tide gauge on year yardstick at present
Value is handled, and amendment series, which is fitted, by least square method can not reflect tide gauge variation tendency well
Relation between measured value, and ultimately result in both relation distortions.
In order to solve above-mentioned technical problem, present invention employs following scheme：
A kind of tide gauge consistent correction method based on moon yardstick, comprises the following steps：
Step 1, the moon tide gauge of a certain website is sequentially arranged into tidal level series；The tide gauge is real
Moon sight highest water level data or actual measurement moon lowest water level data；
Step 2, the catastrophe point and abnormity point eliminated in tidal level series：
Step 3, the periodic term factor eliminated in tidal level series：Step 3.1, the tidal level series to elimination random factor are entered
Row wavelet analysis, obtains wavelet variance diagram, the serial main energy cycle shown in analysis wavelet variogram, so that it is determined that tidal level is serial
Primary period n, n is natural number；Step 3.2, the moving average that proceeded by from n/2 serial to tidal level are handled, and are slided every time flat
N are all referred to, the result W (t) after the periodic term that is eliminated；If n/2 is non-integer, it is [n]+1 to take n/2, that is, takes n integer
It is n/2 that part, which Jia 1,；
Step 4, nearly N tide gauge may be considered what is produced under tale quale, and N is >=5 natural number, is calculated
The moon tide gauge of nearly 1 year is averaged A, and average value A represents the trend term under tale quale, is calculated and is surveyed under tale quale
With average value A difference, but the data of nearest 1 year are not involved in calculating, only calculate N-1 actual measurement item and average value A it
Difference；
Step 5, the calculating chart to nearly N under tale quale carry out collect statistics, but do not include providing the moon of nearest 1 year
Material, statistical items include the tidal level value X (t) of each time point elimination abnormity point and catastrophe point under tale quale, eliminated after periodic term
As a result W (t), actual measurement item and average value A difference α (t)；Wherein,
α (t)=A-X (t)；
Step 6, selection BP neural network ask for correction value, choose nearly N moon data, do not include the data of nearest 1 year；
(N-1) * 12 samples are as training set altogether, and input item is defeated to eliminate the difference of the result after periodic term and actual measurement item and average value A
Go out item to eliminate the tidal level value of abnormity point and catastrophe point, and the neutral net after must training；
Step 7, the correction value at the neural computing each time point obtained by step 6 training, input item are all to eliminate
The difference α of result and actual measurement item and average value A after phaset, output is the correction value in each year；
Step 8, the tide gauge of nearest 1 year use measured value for correction value, the amendment before historical summary starting n/2
Value is replaced with n/2 correction values, and n is natural number, if n/2 is non-integer, it is [n]+1 to take n/2, that is, takes n integer part
It is n/2 plus 1.
Further, the tide gauge actual measurement sequence in step 1 is regarded as to the combination of periodic term and random factor, formula
For：
Z (t)=f [A (t), P (t), R (t)],
Wherein Z (t) represent actual measurement tidal level sequence, what A (t), P (t), R (t) were represented respectively be trend term, periodic term and with
The machine factor, t represents the time.
Further, step 2 is made up of step by step following：
Step 2.1, the average value for seeking taken tidal level series；
Step 2.2, with wavelet analysis method, the regression criterion of tidal level series is carried out under two wavelet scales online
Wavelet decomposition；
Step 2.3, the mould for calculating coefficient of wavelet decomposition under two wavelet scales, calculating difference obtain the expectation of each single item
Et；
Step 2.4, detection abnormity point and catastrophe point；
Step 2.5, the time that abnormity point and catastrophe point occur chosen according to detection principle, replaced with average value exceptional value or
Mutation value.
Further, if not occurring modulus maximum point in step 2.2, and in step 2.3 without being mutated, illustrate at this both
It is not abnormity point nor catastrophe point；If occur modulus maximum point in step 2.2, and in step 2.3 without being mutated, explanation
It is abnormity point at this；If occurring modulus maximum point in step 2.2, and there is mutation in step 2.3, illustrate it is mutation at this
Point.
Further, to the value at abnormity point and catastrophe point with tidal level series average value replaces, by abnormity point and dash forward
The processing of height, eliminates interference of the random factor R (t) to tidal level series, is expressed as：
X (t)=g [A (t), P (t)]；What A (t), P (t), R (t) were represented respectively is trend term, periodic term and random factor
；X (t) represents the tidal level series after elimination random factor, and t represents the time.
A kind of application that tide gauge is reduced using above-mentioned modification method, it is characterised in that：Change is eliminated by amendment
Influence of the environment to data, reduces the uniformity of tide gauge, it is ensured that carry out hydrologic(al) frequency analysis with tide gauge
Reliability, reliable foundation is provided using the tide gauge of revised moon yardstick for urban flood defence.For example：Limit flood bank
Build height.
The tide gauge consistent correction method based on moon yardstick has the advantages that：
(1) the problem of the inventive method has been solved currently without being modified on moon yardstick to tide gauge, method
In the exceptional value of tide gauge is handled, solve influence of the exceptional value to correction result.
(2) the inventive method solves round-off error greatly, accuracy is not with bp neural network models to forecast correction value
High the problems such as.
(3) the initiative tide level data to moon yardstick of the present invention has carried out consistent correction, establishes actual measurement tidal level system
Relation between row and trend term, periodic term and random factor item, by contrasting residual error and the meter that different scale small echo is examined
The expectation of the mould of coefficient of wavelet decomposition under two wavelet scales is calculated, the position that abnormity point and catastrophe point occur is determined, with average
Value replaces the interference of exceptional value elimination random factor.The interference for eliminating periodic term is handled by moving average.To trend term,
Eliminate tide gauge and field data after periodic term and set up dependency relation, choose nearly 5 days tide gauge training BP nerves
Network, the network calculations obtained with training obtain correction value.The problem of solving the tide gauge consistent correction of moon yardstick,
The interference of random factor is eliminated during amendment, calculating revised tidal level by the BP neural network model of training provides
Material, improves the reliability and precision of amendment.
Embodiment
With reference to embodiment, the present invention will be further described：
Step 1, by a certain website actual measurement highest water level data or lowest water level data according to year order statistics arrange,
Arranged in year according to the order of the moon.
Step 2, the catastrophe point and abnormity point eliminated in tidal level series：
(1) average value of taken tidal level series is sought.
(2) wavelet analysis method is used, online wavelet is carried out to the regression criterion of tidal level series under two wavelet scales
Decompose, online wavelet recursive decomposition method can be chosen small echo is decomposed.
(3) mould of coefficient of wavelet decomposition under two wavelet scales is calculated, calculating difference obtains the expectation E of each single itemt。
(4) detection abnormity point and catastrophe point
1) do not occur modulus maximum point in (if 2), and in (3) without being mutated, illustrate at this neither abnormity point
It is not catastrophe point.
2) occur modulus maximum point in (if 2), and in (3) without being mutated, illustrate it is abnormity point at this.
3) occur modulus maximum point in (if 2), and there is mutation in (3), illustrate it is catastrophe point at this.
4) value at abnormity point and catastrophe point is replaced with the average value of tidal level series, by abnormity point and catastrophe point
Processing, can eliminate interference of the random factor R (t) to tidal level series, be expressed as：X (t)=g [A (t), P (t)].
Step 3, the periodic term factor eliminated in tidal level series：
3.1st, wavelet analysis is carried out to the tidal level series for eliminating random factor, obtains wavelet variance diagram, analysis wavelet side
The main energy cycle of series shown in difference figure, so that it is determined that the primary period n of tidal level series.
3.2nd, moving average processing is proceeded by from n/2 to tidal level series, each moving average is related to n, disappeared
Except the tidal level series W (t) of periodic term.
Step 4, the tide gauge of nearly 5 years may be considered what is produced under tale quale, calculate the moon tidal level money of nearly 1 year
Expect that the A that averages, A can represent the trend term under tale quale, calculate the difference of actual measurement item and average value A under tale quale：
α (t)=A-X (t).
Step 5, under tale quale nearly 5 years calculating charts carry out collect statistics (not including the moon data of nearest 1 year),
Statistical items include the tidal level value X (t) of each time point elimination abnormity point and catastrophe point under tale quale, eliminate the result after periodic term
W (t), actual measurement item and average value A difference α (t).
Step 6, selection BP neural network ask for correction value, choose the moon money of nearly 5 years (not including the data of nearest 1 year)
Material, totally 48 samples are as training set, and input item is the result W (t) and actual measurement item and average value A difference α after elimination periodic term
(t), output item is the tidal level value X (t) for eliminating abnormity point and catastrophe point.
Step 7, the BP neural network obtained by step 6 training calculate the correction value at each time point, and input item is elimination
Result W (t) and actual measurement item and average value A and actual measurement item and average value A difference α after periodic termt, output is the amendment in each year
Value.
Step 8, the data of nearest 1 year use measured value for correction value, and the correction value before historical summary starting year n/2 is equal
Replaced with n/2 correction values.
Exemplary description is carried out to the present invention above in conjunction with embodiment, it is clear that realization of the invention is not by above-mentioned side
The limitation of formula, as long as employing the various improvement of inventive concept and technical scheme of the present invention progress, or not improved sends out this
Bright design and technical scheme directly applies to other occasions, within the scope of the present invention.
Claims (5)
1. a kind of tide gauge consistent correction method based on moon yardstick, comprises the following steps：
Step 1, the moon tide gauge of a certain website is sequentially arranged into tidal level series；The tide gauge is the actual measurement moon
Highest water level data or actual measurement moon lowest water level data；
Step 2, the catastrophe point and abnormity point eliminated in tidal level series；
Step 3, the periodic term factor eliminated in tidal level series：Step 3.1, the tidal level series progress to eliminating random factor are small
Wave analysis, obtains wavelet variance diagram, the serial main energy cycle shown in analysis wavelet variogram, so that it is determined that the master of tidal level series
Cycle n, n is natural number；Step 3.2, the moving average that proceeded by from n/2 serial to tidal level are handled, and each moving average is related to
And n, the result W (t) after the periodic term that is eliminated；If n/2 is non-integer, it is n/2 to take n plus 1；
Step 4, nearly N tide gauge may be considered what is produced under tale quale, and N is >=5 natural number, calculates nearly one
Year moon tide gauge average A, average value A represents the trend term under tale quale, calculate under tale quale actual measurement item with
Average value A difference, but the data of nearest 1 year are not involved in calculating, only calculate N-1 actual measurement and average value A difference；
Step 5, the calculating chart to nearly N under tale quale carry out collect statistics, but do not include the moon data of nearest 1 year, unite
Counting item includes the tidal level value X (t) of each time point elimination abnormity point and catastrophe point under tale quale, eliminates the result W after periodic term
(t), actual measurement item and average value A difference α (t)；Wherein,
α (t)=A-X (t)；
Step 6, selection BP neural network ask for correction value, choose nearly N moon data, do not include the data of nearest 1 year；Altogether
(N-1) * 12 samples are as training set, and input item exports to eliminate the difference of the result after periodic term and actual measurement item and average value A
To eliminate the tidal level value of abnormity point and catastrophe point, and the neutral net after must training；
Step 7, the correction value at the neural computing each time point obtained by step 6 training, input item are elimination periodic term
Result afterwards and actual measurement item and average value A difference α (t), output is the correction value in each year；
Step 8, the tide gauge of nearest 1 year use measured value for correction value, and the correction value before historical summary starting n/2 is equal
Replaced with n/2 correction values, n is natural number, if n/2 is non-integer, it is n/2 to take n plus 1.
2. the tide gauge consistent correction method based on moon yardstick according to claim 1, it is characterised in that：By step 1
In tide gauge actual measurement sequence regard the combination of periodic term and random factor as, formula is：
Z (t)=f [A (t), P (t), R (t)],
Wherein Z (t) represent actual measurement tidal level sequence, what A (t), P (t), R (t) were represented respectively be trend term, periodic term and it is random because
Subitem, t represents the time.
3. the tide gauge consistent correction method based on moon yardstick according to claim 1, it is characterised in that：
Step 2 is made up of step by step following：
Step 2.1, the average value for seeking taken tidal level series；
Step 2.2, with wavelet analysis method, online wavelet is carried out to the regression criterion of tidal level series under two wavelet scales
Decompose；
Step 2.3, the mould for calculating coefficient of wavelet decomposition under two wavelet scales, calculating difference obtain the expectation E of each single itemt；
Step 2.4, detection abnormity point and catastrophe point；
Step 2.5, the time occurred according to detection principle selection abnormity point and catastrophe point, exceptional value or mutation are replaced with average value
Value.
4. the tide gauge consistent correction method based on moon yardstick according to claim 3, it is characterised in that：If step
Do not occur modulus maximum point in 2.2, and in step 2.3 without being mutated, illustrate at this neither abnormity point is nor mutation
Point；If occur modulus maximum point in step 2.2, and in step 2.3 without being mutated, illustrate it is abnormity point at this；If step
Occur modulus maximum point in 2.2, and have mutation in step 2.3, illustrate it is catastrophe point at this.
5. the tide gauge consistent correction method based on moon yardstick, its feature according to any one of claim 1-4
It is：Value at abnormity point and catastrophe point is replaced with the average value of tidal level series, by the processing to abnormity point and catastrophe point,
Interference of the random factor R (t) to tidal level series is eliminated, is expressed as：
X (t)=g [A (t), P (t)]；What A (t), P (t), R (t) were represented respectively is trend term, periodic term and random factor；X
(t) represent to eliminate the tidal level series after random factor, t represents the time.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610249068.2A CN105893329B (en) | 2016-04-20 | 2016-04-20 | A kind of tide gauge consistent correction method based on moon yardstick |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610249068.2A CN105893329B (en) | 2016-04-20 | 2016-04-20 | A kind of tide gauge consistent correction method based on moon yardstick |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105893329A CN105893329A (en) | 2016-08-24 |
| CN105893329B true CN105893329B (en) | 2017-08-25 |
Family
ID=56705249
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610249068.2A Active CN105893329B (en) | 2016-04-20 | 2016-04-20 | A kind of tide gauge consistent correction method based on moon yardstick |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105893329B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109190254B (en) * | 2018-09-05 | 2023-07-25 | 上海交通大学 | Tidal level change forecasting method based on real-time measurement data |
| CN113806959B (en) * | 2021-09-26 | 2024-03-15 | 上海市水利工程设计研究院有限公司 | Quantitative research method for estuary design high tide level in future situation |
| CN114693002B (en) * | 2022-05-23 | 2022-08-26 | 中国海洋大学 | Tide level prediction method, device, electronic equipment and computer storage medium |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102214262A (en) * | 2010-04-02 | 2011-10-12 | 上海海洋大学 | Tide predicting method |
-
2016
- 2016-04-20 CN CN201610249068.2A patent/CN105893329B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN105893329A (en) | 2016-08-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Xu et al. | Review on regional water resources assessment models under stationary and changing climate | |
| CN107885951B (en) | A kind of Time series hydrological forecasting method based on built-up pattern | |
| US10439594B2 (en) | Actually-measured marine environment data assimilation method based on sequence recursive filtering three-dimensional variation | |
| CN110689179A (en) | Water bloom prediction method based on space-time sequence mixed model | |
| CN103268525B (en) | A kind of Hydrological Time Series simulating and predicting method based on WD-RBF | |
| CN105893329B (en) | A kind of tide gauge consistent correction method based on moon yardstick | |
| CN101718775A (en) | Spatial variability layout plan generation method of heavy metal content in reclamation land soil | |
| CN113569438B (en) | Urban flood model construction method based on multisource rainfall fusion and real-time correction | |
| Lin et al. | Rainfall prediction using innovative grey model with the dynamic index | |
| CN114357716A (en) | Basin hydrological model parameter dynamic estimation method based on digital twinning technology | |
| CN114861840A (en) | Multi-source precipitation data fusion method | |
| Abdolmanafi et al. | Evaluation of global ensemble prediction models for forecasting medium to heavy precipitations | |
| CN117113854B (en) | Salt tide forecasting method based on ConvLSTM and three-dimensional numerical simulation | |
| Van Gent et al. | Neural network modelling of forces on vertical structures | |
| CN107905270A (en) | A kind of Deformation Prediction in Deep Foundation Pit method | |
| Pham et al. | Spatiotemporal variations of extreme sea levels around the South China Sea: assessing the influence of tropical cyclones, monsoons and major climate modes | |
| Rajabi-Kiasari et al. | Forecasting of absolute dynamic topography using deep learning algorithm with application to the Baltic Sea | |
| CN109828270B (en) | Method for representing ground settlement time sequence evolution | |
| CN106403908B (en) | Prediction of water depth method and system based on time series | |
| CN109101759A (en) | A kind of parameter identification method based on forward and reverse response phase method | |
| Warmink et al. | Uncertainty in water level predictions due to various calibrations | |
| CN113158556A (en) | Short-time high-precision forecasting method for regional water level | |
| CN113673145A (en) | Surface subsidence prediction method based on PCA-LSTM neural network | |
| Watson | How to improve estimates of real-time acceleration in the mean sea level signal | |
| Nawab et al. | Comparison of ANN global horizontal irradiation predictions with satellite global horizontal irradiation using statistical evaluation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |