CN108921345A - The river flood-peak stage Real-time Forecasting Method that precipitation forecast is coupled with support vector machines - Google Patents
The river flood-peak stage Real-time Forecasting Method that precipitation forecast is coupled with support vector machines Download PDFInfo
- Publication number
- CN108921345A CN108921345A CN201810686997.9A CN201810686997A CN108921345A CN 108921345 A CN108921345 A CN 108921345A CN 201810686997 A CN201810686997 A CN 201810686997A CN 108921345 A CN108921345 A CN 108921345A
- Authority
- CN
- China
- Prior art keywords
- precipitation
- flood
- rainfall
- river
- support vector
- 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
Links
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
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- 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
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A10/00—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE at coastal zones; at river basins
- Y02A10/40—Controlling or monitoring, e.g. of flood or hurricane; Forecasting, e.g. risk assessment or mapping
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Game Theory and Decision Science (AREA)
- Development Economics (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Educational Administration (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention discloses the river flood-peak stage Real-time Forecasting Method that precipitation forecast is coupled with support vector machines.The present invention forecasts river future flood-peak stage using Quantitative Precipitation Forecast data as impact factor, using support vector machines.This method forecasts flood peak using Quantitative Precipitation Forecast, can significantly shift to an earlier date leading time, predicts the following flood-peak stage in real time with current precipitation forecast information, can correct in time prediction result according to Changes in weather, the precision and reliability of flood peak forecast is continuously improved.The relationship that this method establishes rainfall based on support vector machines, acts the rise factors such as position and flood-peak stage compares flood routing for river channel method, and entire analytic process is quick and convenient, is easy to be grasped using by user.
Description
Technical field
The invention belongs to observation and control technology field, the river that specifically a kind of Quantitative Precipitation Forecast is coupled with support vector machines is big vast
Peak water level Real-time Forecasting Method.
Background technique
Forecast can be carried out to flood damage in advance, formulate various flood control measures in time, upstream carries out reasonable water
Library scheduling, dyke is suitably increased in downstream, and can carry out flood diversion in time in the case of necessary, possibly to subtract greatly to the greatest extent
Few flood damage bring loss, is of great significance.Flood decision needs to know flood-peak stage information, to be best carried out
Effective Flood Control Dispatch and the urgent flood control measure of schedule ahead reduce the loss of flood bring.Therefore, seek a kind of simple reality
It is very necessary with, precision is high, reliable, real time correction river flood-peak stage forecasting procedure.
Tradition establishes river according to river cross-section and roughness data based on the method for hydrological model (flow flood routing for river channel)
Road water level process hydrodynamics forecasting model simulates flood motion process with this and carries out flood forecasting.Such method is by river mould
Type precision and River course change can all have an immense impact on to forecast result, and forecasting process complicated difficult is to be normally applied personnel's palm
It holds, forecast cost is big, it is therefore foreseen that the phase is short.Flood forecasting is carried out based on machine learning method, forecasting model is established according to historical data,
It is simple and practical without excessively considering peb process physical significance.
To by the river of rebuilding and improving, the history hydrological data that can be used for referring to is less more in the near future.Supporting vector
Machine (Support Vector Machine-SVM) is that a kind of statistical learning suitable for small sample that developed recently gets up is managed
By.In recent years, numerical weather forecast technology is fast-developing, Quantitative Precipitation Forecast is introduced in flood forecasting can extend forecast
Valid time.The present invention is using Quantitative Precipitation Forecast data as impact factor, using support vector machines to the following flood in river
Peak water level is forecast.Simultaneously as precipitation forecast information itself has uncertainty, using real-time correction mode to flood peak water
It is forecast position.
Summary of the invention
In view of the deficiencies of the prior art, the present invention proposes the river floods that a kind of Quantitative Precipitation Forecast is coupled with support vector machines
Peak water level Real-time Forecasting Method, it is therefore intended that accurate real-time prediction is carried out to river flood-peak stage.
This method comprises the concrete steps that：
Step 1：Classified according to precipitation event to historical data using pedigree clustering algorithm.
1. reading the relevant information of N Precipitation Process in historical data, including near zone each rainfall measurement station in river is every
Hour precipitation, upstream and downstream river water level amount per hour.
2. calculating the unit time precipitation such as formula (1) of each field Precipitation Process, upstream and downstream website according to reading historical data
With the average water potential difference such as formula (2) of prediction website (Gong Chenqiao).
Wherein i indicates i-th rainfall in history rainfall；K indicates the duration (hour) of i-th rainfall；Rr,i(k)
For i-th rainfall kth hourly rainfall depth survey station r precipitation；For i-th rainfall rainfall measurement station r unit time precipitation；
dZs,iIt (k) is the water-head of i-th rainfall kth hour upstream and downstream river water level measuring point s and prediction website；It is dropped for i-th
The average water potential difference of rain upstream and downstream river water level measuring point s and prediction website.
3. defining different field Precipitation Process i, j using area unit time precipitation, water levels of upstream and downstream difference as clustering factor
The distance between dijIt is as follows：
Wherein L indicates that rainfall measurement station sum relevant to website to be measured, S indicate upstream and downstream relevant to website to be measured river
Road water level measuring point sum.
After distance matrix foundation finishes, the Precipitation Process of all plays can be clustered.Set distance threshold
Value completes cluster, wherein using the mean value of each all objects of cluster as cluster centre；Each cluster has similar rainfall
Journey during subsequent analysis, can select corresponding cluster to train prediction model to be predicted according to precipitation character.
Step 2：By the prior art obtain above-mentioned all rainfall measurement stations the following Quantitative Precipitation Forecast and it is current when
Engrave the water level of each water level measuring point of downstream river course；Each cluster centre obtained with step 1 calculates distance, and wherein distance is minimum for selection
Cluster as training sample.
Step 3：The moment river n acts the Z that rises in training sample obtained in read step 2i(n), the corresponding moment is related
4 hours precipitation R before regionr,i(n-1)、Rr,i(n-2)、Rr,i(n-3)、Rr,i(n-4), the following precipitation at corresponding moment
Rr,i,f(n) and the duration T of corresponding future at moment precipitationr,i,f(n), calculating flood peak and the water level at the moment rise poor (following
Abbreviation flood peak rises difference) dZf(n)；
dZf(n)=Zf-Z(n) (4)
Step 4：Using principal component analytical method to precipitation in training sample, act rise position, rainy persistent time data
Dimensionality reduction.
1. establishing eigenmatrix X
Wherein m is characteristic, xlIndicate the characteristic parameter vector of first of historical data sample, xnowExpression will currently be predicted
The characteristic parameter vector (precipitation, act rise position, rainy persistent time can be obtained by the prior art) at moment.
2. carrying out average value processing by column to matrix X, eigenvalue matrix X is obtained0；
3. seeking its covariance matrix：M=X0·X'0, wherein X'0For X0Transposed matrix.
4. seeking the eigenvalue λ of covariance matrix Mi(i=1,2 ..., m), λ1> λ2> ... > λmWith corresponding feature to
Measure p1, p2..., pm。
5. according to preceding v principal component amount p1, p2..., pvAccumulation contribution rate, obtain v.
95% or more θ value, that is, taking v pivot includes the information of 95% or more initial data, original with v pivot characterization
Information.
Step 5：Construct training sample.
Using the input matrix P of the preceding N row building support vector machines of the feature vector after step 4 dimensionality reduction, risen using water level
The objective matrix Y of difference building support vector machines.
Wherein EN=[IN 0N],
Therefore input matrix P v data of every behavior, objective matrix 1 data of every behavior.
Step 6：Utilize the training sample Training Support Vector Machines model constructed with step 5.
The svmtrain function in Matlab7.10.0 function library is called to be supported vector machine model training, major parameter
It is chosen including kernel function, the determination of loss function ε and penalty factor parameter C.Support vector machines often includes multinomial with kernel function
Kernel function, radial basis function (RBF) kernel function, Sigmoid kernel function usually select RBF function；Loss function ε is determined back
Error is returned it is expected, value size will will affect the supporting vector quantity and generalization ability of corresponding model, and ε value is bigger, phase
Model supports vector is answered to reduce, precision of prediction is lower, and vice versa, and ε is generally taken as (0.0001~0.01)；Penalty factor parameter
C is mainly used for balancing approximate error and model complexity, and the value of C is bigger, and corresponding model error of fitting is then smaller, to data
Fitting degree is higher, but model complexity is also bigger, and C is generally taken as (1~1000)；
Step 7：The supporting vector machine model obtained using step 6 combines input parameter to obtain prediction data.
Using the N+1 row of the feature vector after step 4 dimensionality reduction as input, call in Matlab7.10.0 function library
The current flood peak of svmpredict function prediction rises difference
Flood-peak stage Zf(j) it is
The beneficial effects of the invention are as follows：
1. this method forecasts flood peak using Quantitative Precipitation Forecast, leading time can be shifted to an earlier date, significantly with current precipitation
Forecast information predicts the following flood-peak stage in real time, and prediction result can be corrected in time according to Changes in weather, and flood peak forecast is continuously improved
Precision and reliability.
2. this method be based on pedigree clusters classify to historical data, according to weather report when meteorology and water regime choosing
Historical data is selected, the model for obtaining training is more accurate；Using Principal Component Analysis to influence factor dimensionality reduction, data are reduced
Redundancy improves efficiency of algorithm.
3. the relationship that this method establishes rainfall based on support vector machines, acts the rise factors such as position and flood-peak stage, is one
Kind is suitable for the strong forecasting procedure of small sample, generalization ability.Entire analytic process is quick and convenient, is easy to be grasped using by user.
Detailed description of the invention
Fig. 1 is the river flood-peak stage Real-time Forecasting Method process that a kind of Quantitative Precipitation Forecast is coupled with support vector machines
Figure.
Specific embodiment
Method of the invention is further described below in conjunction with attached drawing.
By taking great mansion bridge channel stage forecasting is encircleed in Hangzhou canal as an example.
As shown in Figure 1, the river flood-peak stage real-time prediction side coupled for a kind of Quantitative Precipitation Forecast with support vector machines
Method flow chart, specific implementation step are as follows：
Step 1：Classified according to precipitation event to historical data using pedigree clustering algorithm.
1. the history relevant information of the Hangzhou canal arch field great mansion bridge N (N=65) Precipitation Process between reading 1981~2016,
Including each rainfall measurement station of river near zone (Tang Xi, Gong Chenqiao, seven forts, sluice gate) hourly precipitation, upstream and downstream river is (high
Moral, pool northwest, Tongxiang) water level amount per hour.
2. calculating the unit time precipitation such as formula (1) of each field Precipitation Process, upstream and downstream website according to reading historical data
With the average water potential difference such as formula (2) of prediction website (Gong Chenqiao).
Wherein i indicates i-th rainfall in history rainfall；K indicates the duration (hour) of i-th rainfall；Rr,i(k)
For kth hourly rainfall depth survey station r precipitation；For rainfall measurement station r unit time precipitation；dZs,i(k) on kth hour
The water-head of downstream river course water level measuring point s and prediction website；It is averaged for upstream and downstream river water level measuring point s and prediction website
Water-head.
3. defining different field Precipitation Process i, j using area unit time precipitation, water levels of upstream and downstream difference as clustering factor
The distance between dijIt is as follows：
Wherein L indicates that rainfall measurement station sum relevant to website to be measured, S indicate upstream and downstream relevant to website to be measured river
Road water level measuring point sum.
After distance matrix foundation finishes, the Precipitation Process of all plays can be clustered.Set distance threshold
Value is 1, completes cluster, obtains 3 clusters, and using the mean value of each all objects of cluster as cluster centre, each cluster has
Similar rainfall during subsequent analysis, can select corresponding cluster to train prediction model to carry out according to precipitation character
Prediction.
Step 2：For encircleing great mansion bridge 4 days 17 July in 2017:On the 00-7 month 4 20:The flood-peak stage of rainfall during 00
It is forecast, reads on July 4 17:The water of each water level measuring point of the following Quantitative Precipitation Forecast and upstream and downstream river at 00 moment
Position；Each cluster centre obtained with step 1 calculates distance, selects wherein apart from the smallest cluster as training sample (cluster sample
This is 39).
Step 3：The moment river n acts the Z that rises in training sample obtained in read step 2i(n), the corresponding moment is related
4 hours precipitation R before regionr,i(n-1)、Rr,i(n-2)、Rr,i(n-3)、Rr,i(n-4), the following precipitation at corresponding moment
Rr,i,f(n) and the duration T of corresponding future at moment precipitationr,i,f(n), calculating flood peak and the water level at the moment rise poor (following
Abbreviation flood peak rises difference) dZf(n)；
dZf(n)=Zf-Z(n) (4)
Step 4：Using principal component analytical method to precipitation in training sample, act rise position, rainy persistent time data
Dimensionality reduction.
1. establishing eigenmatrix X
Wherein m is characteristic, xlIndicate the characteristic parameter vector of first of historical data sample, xnowExpression will currently be predicted
The characteristic parameter vector (precipitation, act rise position, rainy persistent time can be obtained by the prior art) at moment.
2. carrying out average value processing by column to matrix X, eigenvalue matrix X is obtained0；
3. seeking its covariance matrix：M=X0·X'0, wherein X'0For X0Transposed matrix.
4. seeking the eigenvalue λ of covariance matrix Mi(i=1,2 ..., m), λ1> λ2> ... > λmWith corresponding feature to
Measure p1, p2..., pm。
5. according to preceding v principal component amount p1, p2..., pvAccumulation contribution rate, obtain v.
95% or more θ value, that is, taking v pivot includes the information of 95% or more initial data, original with 5 pivot characterizations
Information.It is 5 dimensional feature information by original 16 dimensional feature information dimensionality reduction.
Step 5：Construct training sample.
Using the input matrix P of 39 rows building support vector machines before the feature vector after step 4 dimensionality reduction, risen using water level
The objective matrix Y of difference building support vector machines.
Wherein EN=[IN 0N],
Therefore input matrix P 5 data of every behavior, objective matrix 1 data of every behavior.
Step 6：Utilize the training sample Training Support Vector Machines model constructed with step 5.
The svmtrain function in Matlab7.10.0 function library is called to be supported vector machine model training, major parameter
It is chosen including kernel function, the determination of loss function ε and penalty factor parameter C.Support vector machines often includes multinomial with kernel function
Kernel function, radial basis function (RBF) kernel function, Sigmoid kernel function usually select RBF function；Loss function ε is determined back
Error is returned it is expected, value size will will affect the supporting vector quantity and generalization ability of corresponding model, and ε value is bigger, phase
Model supports vector is answered to reduce, precision of prediction is lower, and vice versa, and ε is generally taken as (0.01)；Penalty factor parameter C is mainly used
In balance approximate error and model complexity, the value of C is bigger, and corresponding model error of fitting is then smaller, to the fitting journey of data
Degree is higher, but model complexity is also bigger, and C is generally taken as (2.2)；
Step 7：The supporting vector machine model obtained using step 6 combines input parameter to obtain prediction data.
Using the 40th row of the feature vector after step 4 dimensionality reduction as input, call in Matlab7.10.0 function library
The current flood peak of svmpredict function prediction rises difference
Flood-peak stage Zf(j) it is
Based on 4 days 17 July in 2017:The arch great mansion bridge river future that the input data at 00 moment uses this patent model to obtain
Flood-peak stage predicted value is 1.93m, and practical 4 days 17 July in 2017:On the 00-7 month 4 20:Flood-peak stage during 00 is
2.43m, absolute error 0.08m, relative error 3.29%, and the prediction flood-peak stage obtained with empirical model is 0.22m,
Absolute error is 0.15m, relative error 6.17%.Therefore this model prediction result has higher precision.
Claims (2)
1. the river flood-peak stage Real-time Forecasting Method that precipitation forecast is coupled with support vector machines, it is characterised in that this method includes
Following steps：
Step 1：Classified according to precipitation event to historical data using pedigree clustering algorithm：
1. reading the relevant information of N Precipitation Process in historical data, including each rainfall measurement station of river near zone is per hour
Precipitation, upstream and downstream river water level amount per hour；
2. calculating the unit time precipitation such as formula (1) of each field Precipitation Process according to historical data is read, upstream and downstream website and pre-
The average water potential difference such as formula (2) of survey station point；
Wherein i indicates i-th rainfall in history rainfall；K indicates the duration (hour) of i-th rainfall；Rr,iIt (k) is i-th
Field rainfall kth hourly rainfall depth survey station r precipitation；For i-th rainfall rainfall measurement station r unit time precipitation；dZs,i
It (k) is the water-head of i-th rainfall kth hour upstream and downstream river water level measuring point s and prediction website；For in i-th rainfall
The average water potential difference of downstream river course water level measuring point s and prediction website；
3. being defined between different field Precipitation Process i, j using area unit time precipitation, water levels of upstream and downstream difference as clustering factor
Distance dijIt is as follows：
Wherein L indicates that rainfall measurement station sum relevant to website to be measured, S indicate upstream and downstream urban river water relevant to website to be measured
Position measuring point sum；
4. being clustered to the Precipitation Process of all plays；
Set distance threshold value completes cluster, wherein using the mean value of each all objects of cluster as cluster centre；Each cluster tool
There is similar rainfall, selects corresponding cluster to train prediction model to be predicted according to precipitation character；
Step 2：Obtain each water of the following Quantitative Precipitation Forecast and current time upstream and downstream river of above-mentioned all rainfall measurement stations
The water level of position measuring point；Each cluster centre obtained with step 1 calculates distance, selects wherein apart from the smallest cluster as training sample
This；
Step 3：The moment river n acts the Z that rises in training sample obtained in read step 2i(n), before corresponding moment relevant range
4 hours precipitation Rr,i(n-1)、Rr,i(n-2)、Rr,i(n-3)、Rr,i(n-4), the following precipitation R at corresponding momentr,i,f(n)
And the duration T of corresponding future at moment precipitationr,i,f(n), the water level for calculating flood peak and the moment rises poor dZf(n)；
dZf(n)=Zf-Z(n) (4)
Step 4：Using principal component analytical method to precipitation in training sample, act rise position, rainy persistent time Data Dimensionality Reduction；
Step 5：Construct training sample：
Using the input matrix P of the preceding N row building support vector machines of the feature vector after step 4 dimensionality reduction, risen poor structure using water level
Build the objective matrix Y of support vector machines；
Wherein EN=[IN 0N], yk=[dZf(k)]；
Step 6：Utilize the training sample Training Support Vector Machines model constructed with step 5；
Step 7：The supporting vector machine model obtained using step 6 combines input parameter to obtain prediction data；
Using the N+1 row of the feature vector after step 4 dimensionality reduction as input, call in Matlab7.10.0 function library
The current flood peak of svmpredict function prediction rises difference
Flood-peak stage Zf(j) it is
2. the river flood-peak stage Real-time Forecasting Method that precipitation forecast as described in claim 1 is coupled with support vector machines,
It is characterized in that step (4) are specifically：
1. establishing eigenmatrix X
Wherein m is characteristic, xlIndicate the characteristic parameter vector of first of historical data sample, xnowExpression currently wants prediction time
Characteristic parameter vector；
2. carrying out average value processing by column to matrix X, eigenvalue matrix X is obtained0；
3. seeking its covariance matrix：M=X0·X'0, wherein X'0For X0Transposed matrix；
4. seeking the eigenvalue λ of covariance matrix Mi(i=1,2 ..., m), λ1> λ2> ... > λmWith corresponding feature vector p1,
p2..., pm；
5. according to preceding v principal component amount p1, p2..., pvAccumulation contribution rate, obtain v；
95% or more θ value, that is, taking v pivot includes the information of 95% or more initial data, characterizes original letter with v pivot
Breath.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810686997.9A CN108921345A (en) | 2018-06-28 | 2018-06-28 | The river flood-peak stage Real-time Forecasting Method that precipitation forecast is coupled with support vector machines |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810686997.9A CN108921345A (en) | 2018-06-28 | 2018-06-28 | The river flood-peak stage Real-time Forecasting Method that precipitation forecast is coupled with support vector machines |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108921345A true CN108921345A (en) | 2018-11-30 |
Family
ID=64423207
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810686997.9A Pending CN108921345A (en) | 2018-06-28 | 2018-06-28 | The river flood-peak stage Real-time Forecasting Method that precipitation forecast is coupled with support vector machines |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108921345A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109978235A (en) * | 2019-03-04 | 2019-07-05 | 宁波市气象服务中心 | A kind of flooded water level prediction method of product based on sample learning |
| CN109978263A (en) * | 2019-03-27 | 2019-07-05 | 上海市园林设计研究总院有限公司 | A kind of parks' water system water level method for early warning |
| CN110009002A (en) * | 2019-03-13 | 2019-07-12 | 河海大学 | The quick early alarming and forecasting method of mountain torrents based on multidimensional characteristics of rainfall space optimizing decision |
| CN110059766A (en) * | 2019-04-26 | 2019-07-26 | 兰州大学 | A method of algorithm, which is closed on, with K carries out type of precipitation classification |
| CN110298498A (en) * | 2019-06-14 | 2019-10-01 | 水利部交通运输部国家能源局南京水利科学研究院 | A kind of Flood of small drainage area forecasting procedure for establishing data-driven with support vector machines |
| CN112529252A (en) * | 2020-11-18 | 2021-03-19 | 贵州电网有限责任公司 | Small hydropower station forebay water level prediction method and prediction system |
| CN112926786A (en) * | 2021-03-10 | 2021-06-08 | 太湖流域管理局水利发展研究中心 | Shallow lake target water level reverse prediction method and system based on association rule model and numerical simulation |
| CN113378484A (en) * | 2021-07-12 | 2021-09-10 | 中国水利水电科学研究院 | Plain river network area flood process forecasting method based on machine learning |
| CN113435631A (en) * | 2021-06-04 | 2021-09-24 | 南京河海南自水电自动化有限公司 | Flood forecasting method and system, readable storage medium and computing device |
-
2018
- 2018-06-28 CN CN201810686997.9A patent/CN108921345A/en active Pending
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109978235B (en) * | 2019-03-04 | 2022-12-27 | 宁波市气象服务中心 | Waterlogging water level prediction method based on sample learning |
| CN109978235A (en) * | 2019-03-04 | 2019-07-05 | 宁波市气象服务中心 | A kind of flooded water level prediction method of product based on sample learning |
| CN110009002A (en) * | 2019-03-13 | 2019-07-12 | 河海大学 | The quick early alarming and forecasting method of mountain torrents based on multidimensional characteristics of rainfall space optimizing decision |
| CN110009002B (en) * | 2019-03-13 | 2020-02-07 | 河海大学 | Mountain torrent rapid early warning and forecasting method based on multi-dimensional rainfall characteristic space optimal decision |
| CN109978263A (en) * | 2019-03-27 | 2019-07-05 | 上海市园林设计研究总院有限公司 | A kind of parks' water system water level method for early warning |
| CN109978263B (en) * | 2019-03-27 | 2023-06-09 | 上海市园林设计研究总院有限公司 | Garden water system water level early warning method |
| CN110059766A (en) * | 2019-04-26 | 2019-07-26 | 兰州大学 | A method of algorithm, which is closed on, with K carries out type of precipitation classification |
| CN110298498A (en) * | 2019-06-14 | 2019-10-01 | 水利部交通运输部国家能源局南京水利科学研究院 | A kind of Flood of small drainage area forecasting procedure for establishing data-driven with support vector machines |
| CN112529252A (en) * | 2020-11-18 | 2021-03-19 | 贵州电网有限责任公司 | Small hydropower station forebay water level prediction method and prediction system |
| CN112529252B (en) * | 2020-11-18 | 2022-05-03 | 贵州电网有限责任公司 | Small hydropower station forebay water level prediction method and prediction system |
| CN112926786B (en) * | 2021-03-10 | 2022-01-04 | 太湖流域管理局水利发展研究中心 | Shallow lake target water level reverse prediction method and system based on association rule model and numerical simulation |
| CN112926786A (en) * | 2021-03-10 | 2021-06-08 | 太湖流域管理局水利发展研究中心 | Shallow lake target water level reverse prediction method and system based on association rule model and numerical simulation |
| CN113435631A (en) * | 2021-06-04 | 2021-09-24 | 南京河海南自水电自动化有限公司 | Flood forecasting method and system, readable storage medium and computing device |
| CN113435631B (en) * | 2021-06-04 | 2024-02-02 | 南京河海南自水电自动化有限公司 | Flood forecasting method, flood forecasting system, readable storage medium and computing device |
| CN113378484A (en) * | 2021-07-12 | 2021-09-10 | 中国水利水电科学研究院 | Plain river network area flood process forecasting method based on machine learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108921345A (en) | The river flood-peak stage Real-time Forecasting Method that precipitation forecast is coupled with support vector machines | |
| CN111310968B (en) | LSTM neural network circulating hydrologic forecasting method based on mutual information | |
| CN107992961B (en) | Adaptive drainage basin medium and long term runoff forecasting model architecture method | |
| CN106168991B (en) | A kind of tidal river network tide prediction method based on hydrodynamic simulation | |
| CN103729550B (en) | Multiple-model integration Flood Forecasting Method based on propagation time cluster analysis | |
| CN105139093B (en) | Flood Forecasting Method based on Boosting algorithm and support vector machines | |
| CN108830423B (en) | Annual runoff prediction method based on forecast domain | |
| CN111767517B (en) | BiGRU multi-step prediction method, system and storage medium applied to flood prediction | |
| CN111665575B (en) | Medium-and-long-term rainfall grading coupling forecasting method and system based on statistical power | |
| CN109143408B (en) | Dynamic region combined short-time rainfall forecasting method based on MLP | |
| CN111553394B (en) | Reservoir water level prediction method based on cyclic neural network and attention mechanism | |
| CN110598352B (en) | Drainage basin water supply forecasting method | |
| CN108647425B (en) | K-means high flow or low flow time forecasting procedure based on particle group optimizing | |
| CN113705931B (en) | Method for predicting runoff elements by using K nearest neighbor method | |
| CN116362419A (en) | Urban flood control early warning system and method | |
| CN112084461A (en) | KNN real-time correction method based on historical flood learning | |
| CN109977569B (en) | Multi-factor fused MOS storm surge process disaster simulation method | |
| CN111815043B (en) | Flood flow prediction method and system based on storm characteristics | |
| CN109190797A (en) | A kind of large-scale road network state Forecasting Approach for Short-term based on improvement k arest neighbors | |
| CN109919362A (en) | A kind of Medium-and Long-Term Runoff Forecasting method for considering hydraulic engineering scheduling and influencing | |
| CN116167513A (en) | River basin flood response time calculation method based on univariate optimization DMCA model | |
| CN115730739A (en) | Urban waterlogging prediction method based on LSTM neural network | |
| CN115330036A (en) | GRU-Seq2 Seq-based multistep long flood forecasting method and device | |
| CN115127556A (en) | Ship position prediction method based on LSTM neural network | |
| CN114325879A (en) | Quantitative precipitation correction method based on classification probability |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181130 |