CN109886461A - A kind of Runoff Forecast method and device - Google Patents
A kind of Runoff Forecast method and device Download PDFInfo
- Publication number
- CN109886461A CN109886461A CN201910047588.9A CN201910047588A CN109886461A CN 109886461 A CN109886461 A CN 109886461A CN 201910047588 A CN201910047588 A CN 201910047588A CN 109886461 A CN109886461 A CN 109886461A
- Authority
- CN
- China
- Prior art keywords
- period
- forecasting
- historical
- feature vector
- accumulative rainfall
- 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
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention provides a kind of Runoff Forecast method and device, by obtaining feature vector to forecasting period, this feature vector include to forecasting period accumulative rainfall amount, to forecasting period a upper period accumulative rainfall amount, to forecasting period a upper period average run-off, to forecasting period a upper period accumulative rainfall variable quantity and the average streamflow change amount of the upper period to forecasting period；Similarity calculation will be carried out to the feature vector of forecasting period feature vector corresponding with historical sample each historical period of concentration；Based on similarity calculation as a result, filtering out the historical period of the condition of satisfaction according to preset rules, each historical period for meeting condition run-off that is averaged is weighted and averaged processing, obtains the average run-off to forecasting period；It avoids through error present in the main stream approach such as transforming relationship between building impact factor and runoff, improves the precision and accuracy of Runoff Forecast.
Description
Technical field
The present invention relates to hydrologic forecast field more particularly to a kind of Runoff Forecast method and devices.
Background technique
The Medium-and Long-Term Runoff Forecasting of month scale is the important foundation and foundation for working out water resource long-term dispatch scheme, for water
Resource Management is significant.The variation of runoff is influenced by a large amount of factors, including precipitation, atmospheric circulation, sea temperature distribution, day
Literary activity, land surface state etc., current main-stream research is mostly around all kinds of impact factors, it is intended to by mathematical method find each influence because
Son and the correlation of runoff process, and high correlation factor and runoff are established directly or indirectly using method linearly or nonlinearly
Transforming relationship formula.
Current research focuses mostly in terms of the foundation of selection and statistical law to Key Influential Factors, and essence is desirable to
Reflect the Quantitative yield relationship of each impact factor and runoff by mathematical method.However, this transforming relationship is by a large amount of real
Measured data calibration obtains, the practical basin extremely high for a dimension, the field data calibration of limited length, limited quality
Obtained transforming relationship certainly exists larger defect in representativeness.And hydrology phenomenon is extremely complex, exists apparent non-linear
Feature is difficult to accurately be described using existing mathematical measure, limited so as to cause model generalization ability, forecast precision and
Reliability is poor.
Summary of the invention
A kind of Runoff Forecast method and device provided by the invention, mainly solving the technical problems that: how to improve runoff
The precision and reliability of forecast.
In order to solve the above technical problems, the present invention provides a kind of Runoff Forecast method, comprising:
The feature vector to forecasting period is obtained, described eigenvector includes described to forecasting period accumulative rainfall amount, institute
Stated the accumulative rainfall amount of the upper period to forecasting period, the average run-off of the upper period to forecasting period, described
The average diameter rheology of accumulative rainfall variable quantity and the upper period to forecasting period
Change amount；
The feature vector to forecasting period feature vector corresponding with historical sample each historical period of concentration is carried out
Similarity calculation；The corresponding feature vector of the historical period included current historical period accumulative rainfall amount, the accumulation of a upper period
Precipitation, upper period average diameter flow, upper period accumulative rainfall variable quantity and upper period average diameter stream variable quantity；
Based on similarity calculation as a result, filtering out the historical period of the condition of satisfaction according to preset rules, by each history
The period run-off that is averaged is weighted and averaged processing, obtains the average run-off to forecasting period.
Optionally, the period is as unit of month.
Optionally, it is described based on similarity calculation as a result, filtering out the historical period packet of the condition of satisfaction according to preset rules
It includes: the similarity calculation result is ranked up according to numerical values recited, the biggish K historical period of numerical value is selected, as institute
State the historical period of the condition of satisfaction；The K is more than or equal to 1.
Optionally, the K=7.
The present invention also provides a kind of Runoff Forecast devices, comprising:
Module is obtained, for obtaining the feature vector to forecasting period, described eigenvector includes described to forecasting period
Accumulative rainfall amount, the accumulative rainfall amount of the upper period to forecasting period, the upper period to forecasting period it is flat
Equal run-off, the accumulative rainfall variable quantity of the upper period to forecasting period and the upper period to forecasting period
Average streamflow change amount；It is also used to obtain historical sample collection, it includes the corresponding spy of each historical period that the historical sample, which is concentrated,
Vector is levied, the corresponding feature vector of the historical period includes current historical period accumulative rainfall amount, upper period accumulative rainfall
Amount, upper period average diameter flow, upper period accumulative rainfall variable quantity and upper period average diameter stream variable quantity；
Computing module concentrates each historical period corresponding for calculating the feature vector to forecasting period with historical sample
Feature vector between similarity；
Processing module, for based on similarity calculation as a result, filter out the historical period of the condition of satisfaction according to preset rules,
Each historical period run-off that is averaged is weighted and averaged processing, obtains the average run-off to forecasting period.
Optionally, the period is as unit of month.
Optionally, the processing module is selected for the similarity calculation result to be ranked up according to numerical values recited
The biggish K historical period of numerical value, as the historical period for meeting condition；The K is more than or equal to 1.
Optionally, the K=7.
The beneficial effects of the present invention are:
A kind of Runoff Forecast method and device provided according to the present invention, by obtaining the feature vector to forecasting period,
This feature vector include to forecasting period accumulative rainfall amount, to forecasting period a upper period accumulative rainfall amount, to give the correct time in advance
Section a upper period average run-off, to forecasting period a upper period accumulative rainfall variable quantity and to forecasting period
The average streamflow change amount of a upper period；Each historical period will be concentrated corresponding with historical sample to the feature vector of forecasting period
Feature vector carries out similarity calculation；The corresponding feature vector of historical period includes current historical period accumulative rainfall amount, upper one
Period accumulative rainfall amount, upper period average diameter flow, upper period accumulative rainfall variable quantity and upper period average diameter stream
Variable quantity；Based on similarity calculation as a result, filtering out the historical period of the condition of satisfaction according to preset rules, by each condition that meets
The historical period run-off that is averaged is weighted and averaged processing, obtains the average run-off to forecasting period；It is directly based upon wait forecast
The similitude of information and historical sample, the average run-off for treating forecasting period predicted, avoid by building influence because
Error present in the main stream approach such as transforming relationship, improves the precision and accuracy of Runoff Forecast between son and runoff.
Detailed description of the invention
Fig. 1 is the Runoff Forecast method flow schematic diagram of the embodiment of the present invention one；
Fig. 2 is that the attribute variable of the embodiment of the present invention one collects selection schematic diagram；
Fig. 3 is the Runoff Forecast apparatus structure schematic diagram of the embodiment of the present invention two.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, below by specific embodiment knot
Closing attached drawing, invention is further described in detail.It should be appreciated that specific embodiment described herein is only used to explain this
Invention, is not intended to limit the present invention.
Embodiment one:
The main stream approach of current Medium-long Term Prediction is the mathematical model by building reflection meteorological model rule or establishes more
Factorial regression prognostic equation, but since runoff mechanism of production is extremely complex, there is stronger chaotic property and nonlinear characteristic, therefore
In fact it is difficult accurately to retouch the Runoff Evolution process of the following long period with a mathematical model or one group of regression equation
It states and simulates.Closest (K-Nearest Neighbors, the abbreviation KNN) algorithm of K- is by Hart and Cover in last century 60
Age end is put forward for the first time, and is a kind of mode identification method of classics, its core concept is: objective world is with regularity and again
Existing property, often generates similar result under similar conditions.KNN algorithm is directly based upon historical sample and obtains to forecast information
Feature makes it avoid the process for establishing fixed prediction model or regression equation, so as to avoid model structure error and ginseng
Number error, making it under certain conditions has the preferable value of forecasting.
Referring to Figure 1, Fig. 1 is Runoff Forecast method flow schematic diagram provided in this embodiment, Runoff Forecast method master
Include:
S101, obtain feature vector to forecasting period, wherein feature vector include to forecasting period accumulative rainfall amount, to
The accumulative rainfall amount of a upper period for forecasting period, to forecasting period a upper period average run-off, to forecasting period
The average streamflow change amount of the accumulative rainfall variable quantity of a upper period and the upper period to forecasting period.
In the present embodiment, the period can be as unit of month.
Assuming that current time is the year two thousand twenty January 31, it is the year two thousand twenty 2 months to forecasting period, namely to 2 months 2020
Average run-off forecast, then 2 months accumulative rainfall amounts of available the year two thousand twenty first, specifically can be from Chinese national weather
Office obtains the year two thousand twenty 2 months Precipitation forecast amounts (namely the year two thousand twenty 2 months accumulative rainfall amount), it is of course possible to from other countries' meteorology
Office, such as the U.S., Japan, European weather bureau obtain the accumulative rainfall amount to forecasting period in region to be measured.Then it obtains again
The accumulative rainfall amount in the year two thousand twenty 2 months upper periods namely the year two thousand twenty January, average run-off, accumulative rainfall variable quantity and flat
Equal streamflow change amount.Wherein, the accumulative rainfall amount in the year two thousand twenty January, average run-off, accumulative rainfall variable quantity and average runoff
Variable quantity can be obtained by measured data.The accumulative rainfall variable quantity in the year two thousand twenty January that is to say the tired of the year two thousand twenty January
The difference of product precipitation and the accumulative rainfall amount in December, 2019；The average streamflow change amount in the year two thousand twenty January that is to say the year two thousand twenty 1
The average run-off of the moon and the difference of the average run-off in December, 2019.
Get the year two thousand twenty 2 months accumulative rainfall amounts, the accumulative rainfall amount in the year two thousand twenty January, the year two thousand twenty January are averaged
After run-off, the accumulative rainfall variable quantity in the year two thousand twenty January, the average streamflow change amount in the year two thousand twenty January, the year two thousand twenty can be obtained
2 months feature vectors.
In other embodiments, it can also obtain simultaneously from multiple and different data sources to forecasting period accumulative rainfall amount,
Such as obtaining certain waters from weather bureaus such as China, the U.S., Europe, Japan waits for forecasting period accumulative rainfall amount simultaneously, will also obtain
To four accumulative rainfall amount data (usually may be to have differences), it is based on this four accumulative rainfall amount data, it is assumed that respectively
It is Pi1、Pi2、Pi3、Pi4, four feature vectors for waiting for forecasting period, respectively (P can be formedi1,Pi-1,Qi-1,ΔPi-1,Δ
Qi-1)、(Pi2,Pi-1,Qi-1,ΔPi-1,ΔQi-1)、(Pi3, Pi-1,Qi-1,ΔPi-1,ΔQi-1)、(Pi4,Pi-1,Qi-1,ΔPi-1,Δ
Qi-1), each feature vector of forecasting period is waited for for this, can be executed following steps S102-S103 respectively, be obtained each spy
The Runoff Forecast of vector is levied as a result, obtaining finally, be weighted and averaged processing for the Runoff Forecast result of each feature vector
The final forecast result of forecasting period is waited for this.
Wherein in features described above vector, Pi-1For the accumulative rainfall amount of the upper period to forecasting period, Qi-1For wait forecast
The average run-off of the upper period of period, Δ Pi-1For the accumulative rainfall variable quantity of the upper period to forecasting period, Δ Qi-1
For the average streamflow change amount of the upper period to forecasting period.Pi-1、Qi-1、ΔPi-1、ΔQi-1Measured data is all based on to obtain
It arrives, therefore the P that four attribute variables concentratei-1、Qi-1、ΔPi-1、ΔQi-1It is identical.
S102, it will be carried out to the feature vector of forecasting period feature vector corresponding with historical sample each historical period of concentration
Similarity calculation；Wherein the corresponding feature vector of historical period included current historical period accumulative rainfall amount, the accumulation of a upper period
Precipitation, upper period average diameter flow, upper period accumulative rainfall variable quantity and upper period average diameter stream variable quantity.
It is given the correct time in advance in the average run-off to the year two thousand twenty 2 months, it is also necessary to obtain the feature vector (structure in history each month
At historical sample collection) namely the year two thousand twenty 2 months before each month feature vector.History each month, corresponding feature vector included
Current history month accumulative rainfall amount, accumulative rainfall amount of upper January, upper January are averaged run-off, accumulative rainfall of upper January
Variable quantity and upper January are averaged streamflow change amount.By taking the year two thousand twenty January as an example, corresponding feature vector includes: the year two thousand twenty
January accumulative rainfall amount, in December, 2019 accumulative rainfall amount, 12 monthly average run-offs in 2019, accumulative rainfall in December, 2019 become
The change amount difference of accumulative rainfall amount and in November, 2019 accumulative rainfall amount (in December, 2019), 12 monthly average diameter rheologies in 2019
Change amount (differences of 12 monthly average run-offs and 11 monthly average run-offs in 2019 in 2019).It should be appreciated that each history month is also
It is corresponding with the average run-off of of that month actual measurement, the of that month of each history month is surveyed average run-off as aim parameter by the present embodiment；
And incite somebody to action.The current history month accumulative rainfall amount of each month feature vector of history, accumulative rainfall amount of upper January, upper January are flat
Equal run-off, accumulative rainfall variable quantity of upper January and upper January are averaged streamflow change amount as attribute variable.
It should be appreciated that historical sample concentrates the numerical value of day part feature vector that can obtain based on measured data.
In the way of similarity calculation, by the feature vector of forecasting period respectively with each historical period character pair vector
Similarity calculation is carried out, the similarity to forecasting period and each historical period is obtained.Wherein, calculating formula of similarity is, for example, Europe
Formula distance, apart from it is smaller show it is more similar.It is of course possible to which other calculating formula of similarity, the present embodiment are without limitation.
S103, based on similarity calculation as a result, the historical period of the condition of satisfaction is filtered out according to preset rules, by each history
The period run-off that is averaged is weighted and averaged processing, obtains the average run-off to forecasting period.
Optionally, based on similarity calculation as a result, according to the historical period that preset rules filter out the condition of satisfaction include: by
Similarity calculation result is ranked up according to numerical values recited, the biggish K historical period of numerical value is selected, as going through for the condition that meets
The history period；Wherein K is more than or equal to 1.It should be understood that K should be less than the number of historical period.Optionally, K=7.Namely choosing
Select with historical period of 7 historical periods most like to the feature vector of forecasting period as the condition that meets, this 7 are gone through
History period corresponding aim parameter Ti (namely surveying average run-off) is weighted and averaged processing and obtains final predicted value.
Assuming that there is one group of historical sample collection to be defined as H.Wherein, H is made of n sample (i.e. historical period), and each sample
This is by m attribute variable XijAnd 1 aim parameter Ti is constituted.Its mathematic(al) representation such as formula (1):
Attribute variable X in the present embodimentijRespectively represent forecasting period accumulative rainfall amount Pi, upper period accumulative rainfall amount
Pi-1, upper period average diameter flow Qi-1, upper period accumulative rainfall amount variable quantity Pi-1_ Increment, a upper period are average
Changes in runoff amount Qi-1Five broadwise amount X of this 5 elements of _ Increment compositioni={ Pi, Pi-1, Qi-1, Pi-1_
Increment, Qi-1_Increment}。
The present embodiment is that case makees embodiment explanation with the storage forecast of Route Danjiangkou Reservoir moon scale,
Danjiangkou Reservoir is Upper Reaches of Hanjiang River outlet.Danjiangkou Reservoir is put in storage data from Water Year Book, in January, 1970 to 2014
The average diameter flow valuve of each moon in December in year.
Areal rainfall data: the rainfall station data of Upper Reaches of Hanjiang River derives from National Meteorological Bureau's shared data, is 1970 1
The accumulative rainfall amount of the moon in December, 2014 each moon.Upper Reaches of Hanjiang River precipitation station density is larger, and spatial distribution is more uniform, therefore face
The method that the calculating of rainfall uses arithmetic mean, as shown in formula (2).
Wherein: PjFor the accumulation areal rainfall of Upper Reaches of Hanjiang River j period, unit mm, pijIt is tired for i-th precipitation station jth period
Product areal rainfall, unit mm, n are the quantity of precipitation station.
It can be described as following mathematical model using the long-period runoff prediction that KNN algorithm carries out moon scale: assuming that wait forecast
It is forecast sample that the five broadwise amounts of period, which are Xi={ Pi, Pi-1, Qi-1, Pi-1_Increment, Qi-1_Increment },
When carrying out moon scale Runoff Forecast, found and pre- test sample first in historical sample collection H using euclidean distance method (formula (4))
This XiK most like neighbour；Then this K target duration set T={ t is found out1, t2..., tk}；It is finally right according to formula (3)
This K aim parameter t1, t2..., tkIt is weighted and averaged processing, obtains the average run-off Q to forecasting periodi。
Wherein, wkFor the weight of each historical sample target valueOrdinary circumstance takes: wk=1/K.
The present embodiment is using Euclidean distance (Euclid Distance) as judge forecast sample and different historical samples
Similitude or close degree.Assuming that two point A={ a in n-dimensional space1,a2,…,an, B={ b1,b2,…,bn, then point
Euclidean distance between A and point B is calculated by formula (4).It should be understood that in other embodiments of the invention, it can also
To calculate the similitude of forecast sample and historical sample using existing any way.
The selection of attribute variable's collection, refers to Fig. 2, " the producing Confluence Model " of certain determination is certainly existed in real world
(itself is a extremely complex model systems for nature), only its mechanism is not yet grasped by the mankind completely at present, or
The mankind are difficult to effectively portray this mechanism in quite a long time, but this " model " is objective reality, such
Under one model system, certainly exists mode input, model structure and model and export three parts.Different from conventional method, originally
Patent does not put forth effort to portray or intermediate description produces confluence mechanism, and thinks in real world this model of objective reality, and with
The numerical model of mankind's building is similar, and identical original state and drive condition will lead to identical output.Therefore, the present embodiment
The physical significance of the complexity and variable that are obtained according to data, the primary condition of mode input includes underlay noodles early period
Part, run-off early period, previous rainfall amount etc., drive condition mainly include precipitation etc..Primary condition and drive condition are simplified.
Within a certain period of time, influence of the land surface condition to stream is produced, is mainly shown as how many pairs of runoff coefficients of soil moisture content
It influences, if soil moisture content is more, then runoff coefficient is then higher in the period, otherwise less.Therefore, the drop of a upper period was chosen
The run-off of water and a upper period represents land surface condition early period.It is well known that precipitation and runoff are continuous process, tool
There is certain duration.Therefore, upper period changes and precipitation and the two attribute variables of upper period streamflow change is selected to participate in
Model calculates.As it can be seen that the underlying attribute variable to reflect primary condition chosen includes upper period accumulative rainfall amount Pi-1、
Upper period average diameter flow Qi-1, upper period accumulative rainfall amount variable quantity Pi-1_ Increment, upper period average diameter stream
Measure variable quantity Qi-1_Increment.To reflect that the attribute variable of boundary condition is mainly the accumulative rainfall amount P of forecasting periodi,
As shown in figure 2 above.
Wherein: Pi-1_ Increment=Pi-1-Pi-2；
Qi-1_ Increment=Qi-1-Qi-2；
The selection of this attribute variable's collection has certain physical significance:
Pi-1With Qi-1: upper period accumulative rainfall and runoff can be directly as the primary condition of model, in addition, their phase
To how much capable of reflecting the primary condition or basin water storage situation of underlying surface indirectly.
Pi-1_ Increment and Qi-1_ Increment: upper period accumulative rainfall variable quantity and upper period average diameter stream
Variable quantity represents the continuity information of precipitation runoff sequence.
Pi: the accumulative rainfall amount of prediction period represents the boundary information of model, cooperates Pi-1With Qi-1Enable model to
The prediction of a continuous sequence is completed in preceding calculating.Current main-stream forecasting model is overcome to generally require to model for each month
The drawbacks of, while expression of forecast result early period in later period forecast conclusion is also achieved, meet the successional feature of runoff.
Runoff Forecast method provided in this embodiment, by obtaining the feature vector to forecasting period, this feature vector packet
Containing to forecasting period accumulative rainfall amount, to forecasting period a upper period accumulative rainfall amount, to a upper period for forecasting period
Average run-off, to forecasting period a upper period accumulative rainfall variable quantity and upper period to forecasting period it is flat
Equal streamflow change amount；It will be carried out to the feature vector of forecasting period feature vector corresponding with historical sample each historical period of concentration
Similarity calculation；The corresponding feature vector of historical period includes current historical period accumulative rainfall amount, upper period accumulative rainfall
Amount, upper period average diameter flow, upper period accumulative rainfall variable quantity and upper period average diameter stream variable quantity；Based on phase
Like degree calculated result, the historical period of the condition of satisfaction is filtered out according to preset rules, each historical period for meeting condition is averaged
Run-off is weighted and averaged processing, obtains the average run-off to forecasting period；It is directly based upon to forecast information and history sample
This similitude, the average run-off for treating forecasting period are predicted, are avoided by between building impact factor and runoff
Error present in the main stream approach such as transforming relationship improves the precision and accuracy of Runoff Forecast.
Embodiment two:
The present embodiment on the basis of example 1, provides a kind of Runoff Forecast device, for realizing above-described embodiment one
The step of described Runoff Forecast method, Fig. 3 is referred to, which includes obtaining module 31, computing module 32
And processing module 33, in which:
It obtains module 31 and is used to obtain the feature vector to forecasting period, this feature vector includes to accumulate to drop to forecasting period
Water, to forecasting period a upper period accumulative rainfall amount, to forecasting period a upper period average run-off, wait forecast
The average streamflow change amount of the accumulative rainfall variable quantity of the upper period of period and the upper period to forecasting period；Obtain mould
Block 31 is also used to obtain historical sample collection, and it includes the corresponding feature vector of each historical period, historical period pair that historical sample, which is concentrated,
The feature vector answered include current historical period accumulative rainfall amount, upper period accumulative rainfall amount, upper period average diameter flow,
Upper period accumulative rainfall variable quantity and upper period average diameter stream variable quantity.
Computing module 32 is used to calculate concentrates each historical period corresponding to the feature vector of forecasting period with historical sample
Similarity between feature vector.
Processing module 33 is used for based on similarity calculation as a result, when filtering out the history of the condition of satisfaction according to preset rules
Each historical period for the condition that the meets run-off that is averaged is weighted and averaged processing, obtains the average runoff to forecasting period by section
Amount.
In the present embodiment, the period can be as unit of month.
Processing module 33 is based on similarity calculation as a result, filtering out the historical period packet of the condition of satisfaction according to preset rules
It includes: similarity calculation result is ranked up according to numerical values recited, the biggish K historical period of numerical value is selected, as meeting item
The historical period of part；Wherein K is more than or equal to 1.It should be understood that K should be less than the number of historical period.Optionally, K=7.
Namely select with historical period of 7 historical periods most like to the feature vector of forecasting period as the condition that meets, to this 7
A historical period is corresponding to survey average run-off and is weighted and averaged processing and obtain final predicted value.
Obviously, those skilled in the art should be understood that each module of aforementioned present invention or each step can be with general
Computing device realizes that they can be concentrated on a single computing device, or be distributed in constituted by multiple computing devices
On network, optionally, they can be realized with the program code that computing device can perform, it is thus possible to be stored in
It is performed by computing device in computer storage medium (ROM/RAM, magnetic disk, CD), and in some cases, it can be with not
The sequence being same as herein executes shown or described step, or they are fabricated to each integrated circuit modules, or
Person makes multiple modules or steps in them to single integrated circuit module to realize.So the present invention is not limited to appoint
What specific hardware and software combines.
The above content is specific embodiment is combined, further detailed description of the invention, and it cannot be said that this hair
Bright specific implementation is only limited to these instructions.For those of ordinary skill in the art to which the present invention belongs, it is not taking off
Under the premise of from present inventive concept, a number of simple deductions or replacements can also be made, all shall be regarded as belonging to protection of the invention
Range.
Claims (8)
1. a kind of Runoff Forecast method, which is characterized in that the Runoff Forecast method includes:
Obtain feature vector to forecasting period, described eigenvector include it is described to forecasting period accumulative rainfall amount, it is described to
It is the accumulative rainfall amount of a upper period for forecasting period, the average run-off of the upper period to forecasting period, described to pre-
The average streamflow change amount of the accumulative rainfall variable quantity of a upper period for section of giving the correct time and the upper period to forecasting period；
The feature vector to forecasting period feature vector corresponding with historical sample each historical period of concentration is carried out similar
Degree calculates；The corresponding feature vector of the historical period includes current historical period accumulative rainfall amount, upper period accumulative rainfall
Amount, upper period average diameter flow, upper period accumulative rainfall variable quantity and upper period average diameter stream variable quantity；
Based on similarity calculation as a result, filtering out the historical period of the condition of satisfaction according to preset rules, by each historical period
Average run-off is weighted and averaged processing, obtains the average run-off to forecasting period.
2. Runoff Forecast method as described in claim 1, which is characterized in that the period is as unit of month.
3. Runoff Forecast method as claimed in claim 1 or 2, which is characterized in that it is described be based on similarity calculation as a result, according to
The historical period that preset rules filter out the condition of satisfaction includes: to arrange the similarity calculation result according to numerical values recited
Sequence selects the biggish K historical period of numerical value, as the historical period for meeting condition；The K is more than or equal to 1.
4. Runoff Forecast method as claimed in claim 3, which is characterized in that the K=7.
5. a kind of Runoff Forecast device, which is characterized in that the Runoff Forecast device includes:
Module is obtained, for obtaining the feature vector to forecasting period, described eigenvector includes described to forecasting period accumulation
Precipitation, the accumulative rainfall amount of the upper period to forecasting period, the average diameter of the upper period to forecasting period
Flow, the accumulative rainfall variable quantity of the upper period to forecasting period and the upper period to forecasting period it is flat
Equal streamflow change amount；Be also used to obtain historical sample collection, the historical sample concentrate comprising the corresponding feature of each historical period to
Amount, the corresponding feature vector of the historical period include current historical period accumulative rainfall amount, upper period accumulative rainfall amount, on
One period average diameter flow, upper period accumulative rainfall variable quantity and upper period average diameter stream variable quantity；
Computing module, for calculating the feature vector to forecasting period spy corresponding with historical sample each historical period of concentration
Levy the similarity between vector；
Processing module, for based on similarity calculation as a result, filter out the historical period of the condition of satisfaction according to preset rules, will be each
The historical period run-off that is averaged is weighted and averaged processing, obtains the average run-off to forecasting period.
6. Runoff Forecast device as claimed in claim 5, which is characterized in that the period is as unit of month.
7. such as Runoff Forecast device described in claim 5 or 6, which is characterized in that the processing module is used for will be described similar
Degree calculated result is ranked up according to numerical values recited, is selected the biggish K historical period of numerical value, is met going through for condition as described
The history period；The K is more than or equal to 1.
8. Runoff Forecast device as claimed in claim 7, which is characterized in that the K=7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910047588.9A CN109886461A (en) | 2019-01-18 | 2019-01-18 | A kind of Runoff Forecast method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910047588.9A CN109886461A (en) | 2019-01-18 | 2019-01-18 | A kind of Runoff Forecast method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109886461A true CN109886461A (en) | 2019-06-14 |
Family
ID=66926219
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910047588.9A Pending CN109886461A (en) | 2019-01-18 | 2019-01-18 | A kind of Runoff Forecast method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109886461A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111401666A (en) * | 2020-04-28 | 2020-07-10 | 中国水利水电科学研究院 | Method for forecasting influence of upstream reservoir group on runoff by utilizing forecasting errors |
| CN113589404A (en) * | 2021-07-30 | 2021-11-02 | 郑州大学 | Method for predicting runoff of storm of field |
| CN113705931A (en) * | 2021-09-17 | 2021-11-26 | 中国长江电力股份有限公司 | Method for predicting runoff elements by using K nearest neighbor method |
| CN113762645A (en) * | 2021-10-11 | 2021-12-07 | 昆仑(重庆)河湖生态研究院(有限合伙) | Natural disaster forecasting method and device |
| CN116226687A (en) * | 2023-05-10 | 2023-06-06 | 长江三峡集团实业发展(北京)有限公司 | Reservoir daily initial water level estimation method and device based on data mining technology |
| CN116911467A (en) * | 2023-09-12 | 2023-10-20 | 浙江华云电力工程设计咨询有限公司 | Renewable energy output prediction method, device and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106650767A (en) * | 2016-09-20 | 2017-05-10 | 河海大学 | Flood forecasting method based on cluster analysis and real time correction |
| CN108876021A (en) * | 2018-05-31 | 2018-11-23 | 华中科技大学 | A kind of Medium-and Long-Term Runoff Forecasting method and system |
| CN109059875A (en) * | 2018-06-28 | 2018-12-21 | 中国水利水电科学研究院 | A method of drive perfect model to carry out moon scale Runoff Forecast |
-
2019
- 2019-01-18 CN CN201910047588.9A patent/CN109886461A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106650767A (en) * | 2016-09-20 | 2017-05-10 | 河海大学 | Flood forecasting method based on cluster analysis and real time correction |
| CN108876021A (en) * | 2018-05-31 | 2018-11-23 | 华中科技大学 | A kind of Medium-and Long-Term Runoff Forecasting method and system |
| CN109059875A (en) * | 2018-06-28 | 2018-12-21 | 中国水利水电科学研究院 | A method of drive perfect model to carry out moon scale Runoff Forecast |
Non-Patent Citations (2)
| Title |
|---|
| 谭乔凤等: "ANN、ANFIS和AR模型在日径流时间序列预测中的应用比较", 《南水北调与水利科技》 * |
| 陈云华等: "《多市场下流域水电定价理论与优化运营》", 31 January 2010, 中国电力出版社 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111401666A (en) * | 2020-04-28 | 2020-07-10 | 中国水利水电科学研究院 | Method for forecasting influence of upstream reservoir group on runoff by utilizing forecasting errors |
| CN111401666B (en) * | 2020-04-28 | 2021-07-27 | 中国水利水电科学研究院 | Method for forecasting influence of upstream reservoir group on runoff by utilizing forecasting errors |
| CN113589404A (en) * | 2021-07-30 | 2021-11-02 | 郑州大学 | Method for predicting runoff of storm of field |
| CN113589404B (en) * | 2021-07-30 | 2023-02-03 | 郑州大学 | Method for predicting runoff volume of storm at scene |
| CN113705931A (en) * | 2021-09-17 | 2021-11-26 | 中国长江电力股份有限公司 | Method for predicting runoff elements by using K nearest neighbor method |
| CN113762645A (en) * | 2021-10-11 | 2021-12-07 | 昆仑(重庆)河湖生态研究院(有限合伙) | Natural disaster forecasting method and device |
| CN116226687A (en) * | 2023-05-10 | 2023-06-06 | 长江三峡集团实业发展(北京)有限公司 | Reservoir daily initial water level estimation method and device based on data mining technology |
| CN116911467A (en) * | 2023-09-12 | 2023-10-20 | 浙江华云电力工程设计咨询有限公司 | Renewable energy output prediction method, device and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109886461A (en) | A kind of Runoff Forecast method and device | |
| Li et al. | Prediction for tourism flow based on LSTM neural network | |
| CN110084367B (en) | Soil moisture content prediction method based on LSTM deep learning model | |
| CN110555561B (en) | Medium-and-long-term runoff ensemble forecasting method | |
| CN106598917B (en) | A kind of upper ocean heat structure prediction technique based on deepness belief network | |
| CN108021773B (en) | DSS database-based distributed hydrological model multi-field secondary flood parameter calibration method | |
| CN108830430A (en) | Convolutional neural networks multiple spot regressive prediction model for traffic flow forecasting | |
| CN101480143B (en) | Method for predicating single yield of crops in irrigated area | |
| CN111767517B (en) | BiGRU multi-step prediction method, system and storage medium applied to flood prediction | |
| CN107145965B (en) | River flood prediction method based on similarity matching and extreme learning machine | |
| CN109711617A (en) | A kind of medium-term and long-term Runoff Forecast method based on BLSTM deep learning | |
| CN109816167A (en) | Runoff Forecast method and Runoff Forecast device | |
| CN109993372A (en) | One kind being based on the probabilistic flood probability forecasting procedure of multi-source | |
| Lafdani et al. | Research article daily rainfall-runoff prediction and simulation using ANN, ANFIS and conceptual hydrological MIKE11/NAM models | |
| EP4226473A1 (en) | Estimating energy consumption for a building using dilated convolutional neural networks | |
| CN109190810B (en) | TDNN-based prediction method for NDVI (normalized difference vegetation index) of northern grassland area of China | |
| CN109272144B (en) | BPNN-based prediction method for NDVI (normalized difference of variance) in northern grassland area of China | |
| CN114692981A (en) | Medium-and-long-term runoff forecasting method and system based on Seq2Seq model | |
| CN106295877B (en) | Method for predicting electric energy consumption of smart power grid | |
| Souhe et al. | Forecasting of electrical energy consumption of households in a smart grid | |
| CN112819244B (en) | Meteorological factor-based RF-HW water quality index hybrid prediction method | |
| CN110442836A (en) | Short-term wind speed forecasting method of wind farm and system | |
| CN117114190B (en) | River runoff prediction method and device based on mixed deep learning | |
| Wang et al. | Prediction of water quality in South to North Water Transfer Project of China based on GA-optimized general regression neural network | |
| CN112488820A (en) | Model training method and default prediction method based on noctilucent remote sensing data |
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 |
Application publication date: 20190614 |
|
| RJ01 | Rejection of invention patent application after publication |