CN115905963A - Flood forecasting method and system based on support vector machine model - Google Patents
Flood forecasting method and system based on support vector machine model Download PDFInfo
- Publication number
- CN115905963A CN115905963A CN202211363780.7A CN202211363780A CN115905963A CN 115905963 A CN115905963 A CN 115905963A CN 202211363780 A CN202211363780 A CN 202211363780A CN 115905963 A CN115905963 A CN 115905963A
- Authority
- CN
- China
- Prior art keywords
- rainfall
- runoff
- model
- scs
- water
- 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
Images
Classifications
-
- 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
Abstract
The invention discloses a flood forecasting method and system based on a support vector machine model, belonging to the field of runoff forecasting; the method comprises the steps of S1, collecting hydrological meteorological data of a research area; s2, determining a current generation module parameter CN value of the SCS-CN model according to the collected hydrological meteorological data; s3, improving the flow generation module of the SCS-CN model by utilizing the rainfall characteristic factor and the flow generation module parameter CN value; s4, establishing an SCS-CN three-water-source confluence model by coupling the improved production flow module and the improved three-water-source confluence module of the S3; s5, constructing a dynamic convergence time parameter through a support vector machine model to improve an SCS-CN three-water-source convergence model; s6, forecasting a flood process line of the outlet section of the drainage basin by using the actually measured rainfall data; the influence of rainfall characteristic factors is considered, and the prediction capability of the model in a region with a relatively dry climate is improved; factors such as a spatial distribution condition of rainfall in a field and early rainfall are considered, so that the risk of prediction deviation of the hydrological model caused by uneven spatial distribution of rainfall is reduced.
Description
Technical Field
The invention belongs to the field of runoff forecasting, and particularly relates to a flood forecasting method and system based on a support vector machine model.
Background
In recent years, machine learning models are rapidly developed and widely applied to flood variable prediction and flood sensitivity evaluation. Since the construction of the machine learning model does not require a large amount of data, the method based on machine learning has certain advantages in predicting flood variables. In previous related studies, machine learning models have been successfully used to predict problems associated with flood variables.
The SCS-CN model is a conceptual model developed by the United states department of agriculture, and can predict the runoff quantity only by two parameters of curve number and initial loss rate. The method has the characteristics of less required input data, relatively simple model structure, higher prediction precision and the like, and is widely applied to estimating the surface runoff of the rainfall event in the small watershed. However, the model is a conceptual model developed based on runoff data of the U.S. basin, and when the model is applied to research areas under different climate types, factors such as local vegetation, terrain, soil and the like need to be comprehensively considered. Through the continuous improvement of domestic scholars, the SCS-CN model has a good effect in domestic watersheds, but the traditional SCS-CN model only has a flow generating module and does not comprise a flow converging module, so that the change trend of the outlet section flow of the watershed cannot be drawn after rainfall occurs.
Most conceptual hydrological models take a watershed as a whole and study the operation and change of a runoff producing mechanism of the whole watershed after a rainfall event occurs. Because the conceptual hydrological model does not have the capacity of calculating the drainage basin cells, when rainfall spatial distribution is uneven, the flood peak time predicted by the conceptual hydrological model and the actual flood peak time are prone to have large deviation, and particularly the deviation occurs in a research area where the drainage basin is long and narrow or the area of the drainage basin is large.
Disclosure of Invention
Aiming at the defects of the prior art, the invention aims to provide a flood forecasting method and system based on a support vector machine model.
The purpose of the invention can be realized by the following technical scheme:
a flood forecasting method based on a support vector machine model comprises the following steps:
s1, collecting hydrological meteorological data of a research area;
s2, determining a current generation module parameter CN value of the SCS-CN model according to the collected hydrological meteorological data;
s3, improving the SCS-CN model runoff producing module by utilizing the rainfall characteristic factor and the runoff producing module parameter CN value;
s4, establishing an SCS-CN three-water-source confluence model by coupling the improved flow generation module and the three-water-source confluence module in the S3;
s5, constructing a dynamic convergence time parameter through a support vector machine model to improve an SCS-CN three-water-source convergence model;
and S6, forecasting a flood process line of the outlet section of the drainage basin by using the actually measured rainfall data.
Further, the hydrometeorological data collected in S1 includes: rainfall intensity, runoff, soil type, and land use type.
Further, in S2, the step of determining parameters of the labor module is:
s21, determining a hydrological soil group through soil type data of a research area, and dividing according to infiltration capacity; resampling the land utilization type of the research area on an Arcgis platform, matching the land utilization type with grids under each soil type, calculating a CN value in each grid, carrying out weighted average on the CN value of each grid, and finally determining a basin comprehensive CN value;
s22, dividing the soil humidity condition into three grades of drought, normal and humidity according to the total rainfall of the previous 5 days; the interval of the soil wetting condition is divided again by adopting K-mean clustering, the total rainfall in the previous 5 days is the early-stage influence rainfall, and the threshold value is set to be 120mm;
s23, converting the CN value according to the actual soil humidity condition of the research areaTo CN 1 、CN 2 And CN 3 And carrying out interpolation processing on the mutation intervals to obtain CN values under different early-stage influence rainfall.
Further, in S3, the SCS-CN runoff yield model is improved by adding the influence of rainfall characteristic factors on the basis of the original SCS-CN model, and the concrete steps are as follows:
s31, in the runoff generating module, early rainfall needs to supplement the initial water shortage of the watershed, if the total rainfall does not meet the initial water shortage, the watershed does not generate runoff, and a runoff calculation formula formed by the rainfall in the field is as follows:
in the formula: p is total rainfall, mm; q is the surface runoff, mm; i is a Is the initial loss, mm; s is potential water storage capacity, mm;
s32, calculating the initial loss and the potential water storage capacity according to the formula:
I a ＝λS (2)
in the formula: lambda is the initial loss rate; CN is a dimensionless parameter;
s33, adding characteristic factors such as rainfall intensity and the like into the runoff producing module, wherein the revised runoff producing calculation formula is as follows:
in the formula: i is 60 The maximum rainfall intensity is 60min in the field rainfall process, and the rainfall intensity is mm/h;
the average rain intensity of field rainfall is mm/h; beta is a raininess correction parameter; />
The method is the comprehensive embodiment of rainfall characteristic factors.
Furthermore, in S4, a process curve of the runoff depth of the research area can be obtained through the SCS-CN runoff generating module, the runoff depth generated by the research area in each period can be obtained through differential processing, and then the runoff depth is input into the three-water-source confluence module for confluence calculation.
Further, the three water sources refer to: surface runoff, interflow and subsurface runoff, and are divided by a parabolic free water storage curve and a free water storage reservoir.
Furthermore, in the confluence calculation, a unit line method is adopted in surface water confluence calculation, and a basin is simulated into the series connection of n linear reservoirs by using a unit line; the confluence calculation of the interflow and the subsurface runoff adopts a linear reservoir method; the river network convergence adopts a hysteresis algorithm, and the calculation formula is as follows:
Q3(I)＝CS×Q3(I-1)+(1-CS)×QT3(I-L) (5)
QT3(I-L)＝QS(I-L)+QG(I-L)+QI(I-L) (6)
in the formula: q3 (I) is the unit area river network confluence of the I time period, m 3 S; CS is the river network water flow fading coefficient; l is a convergence time parameter h; QS surface runoff, m 3 S; QG subsurface runoff, m 3 S; QI interflow, m 3 S; QT3 (I-L) is the sum of surface runoff, subsurface runoff and interflow in the I-L periods, m 3 /s。
Further, the air conditioner is provided with a fan,in the step S5, the first step is carried out,the regression variables of the dynamic confluence time parameter include: the length of a rainfall center to a drainage basin section in a research area, the gradient of the rainfall center, total rainfall, early-stage influence rainfall and total river length; the regression function f (x) of the dynamic convergence time parameter is:
in the formula: (alpha.) of i ,α i * ) Is a lagrange multiplier; b is an offset; k (x, x) i ) The radial basis function from the input space to the high-order feature space is a kernel function; σ is the expansion constant of the radial basis function, representing the radial range of action of the function.
Further, in S6, firstly, the measured rainfall data and the production convergence parameters of S2, S3, and S4 are input into the SCS-CN three-water-source convergence model in S4 to obtain the runoff rate of each computational grid in the research area, and then the runoff rate is input into the improved three-water-source convergence model in S5, so as to obtain the flood process line of the watershed outlet section.
A flood forecasting system based on a support vector machine model comprises:
a data acquisition unit: collecting hydrological meteorological data of a research area;
a parameter solving unit of the runoff generating module: determining a production flow module parameter CN value of the SCS-CN model according to the collected hydrological meteorological data;
the birth flow module improves the unit: improving an SCS-CN model runoff generating module by utilizing rainfall characteristic factors and a runoff generating module parameter CN value;
the SCS-CN three-water-source confluence model building unit comprises: establishing an SCS-CN three-water-source confluence model by coupling an improved flow production module and a three-water-source confluence module;
SCS-CN three water source confluence model improvement unit: constructing a dynamic convergence time parameter through a support vector machine model to improve an SCS-CN three-water-source convergence model;
flood forecasting unit: forecasting a flood process line of the cross section of the drainage basin outlet by using the measured rainfall data;
the invention has the beneficial effects that:
1. according to the method, partition intervals of soil humidity conditions in the previous period are subdivided by adopting K-mean clustering analysis, and CN values under different previous-period influence rainfall are obtained by adopting three-time linear interpolation. The influence of rainfall characteristic factors such as rainfall intensity, average rainfall intensity and the like is considered, the problem that the rainfall intensity is not considered in the original production flow module is solved, and the prediction capability of the model in a region with a relatively dry climate is improved;
2. according to the method, a dynamic watershed convergence time lag parameter is constructed by using a support vector machine model, factors such as a field rainfall spatial distribution condition, early rainfall and underlying surface conditions are considered, so that the risk of forecasting deviation of a hydrological model caused by uneven rainfall spatial distribution is reduced, the flood peak time forecasting capability of the hydrological model is improved, meanwhile, the defects and shortcomings of a conceptual hydrological model in a forecasting process are relieved to a certain extent, and a solution can be provided for the conceptual hydrological model when the conceptual hydrological model is applied to a complex research watershed;
3. according to the invention, the SCS-CN current generation module and the three-water-source convergence module are coupled to obtain the flood process line of the outlet section of the drainage basin, and the SCS-CN three-water-source convergence model has higher precision in terms of simulating the peak flow and the peak time, so that the application range of the SCS-CN model is expanded.
Drawings
In order to more clearly illustrate the embodiments or technical solutions in the prior art of the present invention, the drawings used in the description of the embodiments or prior art will be briefly described below, and it is obvious for those skilled in the art that other drawings can be obtained based on these drawings without creative efforts.
Fig. 1 is a flow chart of steps of a flood forecasting method;
FIG. 2 is a schematic diagram of distribution of sites in a drainage basin of a certain village;
FIG. 3 is a plot of land utilization (left) and soil type (right) for a village basin;
FIG. 4 is a diagram of a dynamic convergence time parameter simulation based on a support vector machine model;
FIG. 5 shows the partial simulation results of SCS-CN three-water-source confluence model.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
As shown in fig. 1, a flood forecasting method based on a support vector machine model includes the following steps:
s1, collecting hydrological meteorological data of a research area; including rainfall intensity, runoff, soil type, land utilization type, and the like.
S2, determining parameters of a runoff generating module of the SCS-CN model according to relevant data such as soil types, land utilization types and the like of a research area; the method comprises the following specific steps:
and S21, determining a hydrological soil group through parameters such as soil texture, soil infiltration rate and the like of the research area, wherein the hydrological soil group can be divided into four types of A, B, C and D, and the infiltration capacity of the hydrological soil group is reduced in sequence. Resampling the land utilization types of the researched area on an Arcgis platform, enabling the land utilization types to be matched with grids under each soil type, calculating CN (Curve Number) values in each grid, carrying out weighted average on the CN values of each grid, and finally determining the comprehensive CN value of the drainage basin (the soil humidity is normal).
S22, dividing soil humidity condition (AMC) into three levels according to the total rainfall of the previous 5 days by an SCS-CN model, namely drought (AMC 1), normal (AMC 2) and humidity (AMC 3);
the accuracy of determination of the early soil humidity condition directly influences the prediction precision and stability of the hydrological model, the relation between the total rainfall 5 days before flood and the appropriate CN value is analyzed through statistics, the K-mean value clustering is adopted to subdivide the interval of the soil humidity condition, the total rainfall 5 days before the flood is the early influence rainfall, and the threshold value is set to be 120mm.
S23, converting the CN value according to the actual soil humidity condition of the research area; since the method for defining soil humidity of SCS-CN model has mutation points, CN is subjected to 1 、CN 2 And CN 3 The mutation interval between them is interpolated, CN 1 The soil is compared with the dry CN value, CN 2 The CN value is the CN value (the CN value obtained by S21) when the soil is normal 3 The CN value of the soil when the soil is wet is shown.
S3, improving an SCS-CN model runoff generating module by adopting characteristic factors such as rainfall intensity, average rainfall intensity and the like;
the improvement of the SCS-CN runoff generating model is to add the influence of rainfall characteristic factors on the basis of the original SCS-CN model;
the SCS-CN model is an empirical model established based on a water balance equation and two basic assumptions; the first assumption is that: the ratio of the surface runoff to the possible maximum runoff is equal to the actual infiltration capacity and the potential water storage capacity; the second assumption is that: a certain proportional relation exists between the initial loss and the potential water storage capacity;
in the runoff generating module, early rainfall needs to supplement the initial water shortage of the watershed, if the total rainfall does not meet the initial water shortage, the watershed does not generate runoff, and a runoff calculation formula formed by the rainfall in the field is as follows:
in the formula: p is total rainfall; q is the surface runoff; I.C. A a The initial loss is the initial loss; s is potential water storage capacity; the initial loss and the potential water storage capacity are calculated according to the formula:
I a ＝λS (2)
in the formula: lambda is the initial loss rate; CN is a dimensionless parameter and can be obtained from S2, the larger CN is, the larger the runoff yield capability of the watershed is, the value of the CN is related to factors such as the land utilization type, the vegetation coverage type and the early rainfall of a research area, and meanwhile, the CN value is a more sensitive parameter in the runoff yield module, and the change of the CN value has a larger influence on the prediction result of the research watershed.
In semiarid regions, the water level of soil is low, the water shortage of the soil is large, after a rain fall, the water holding capacity of a river basin can be hardly reached, and the rain fall intensity exceeds the infiltration intensity to generate the water flow, so the rain fall intensity in the regions plays a main role in the water flow generation process. In contrast, characteristic factors such as rainfall intensity and the like are added into the runoff generating module to improve the prediction accuracy of the model in the semiarid region, and the revised runoff generating calculation formula is as follows:
in the formula: I.C. A 60 The maximum 60min rain intensity in the rainfall process of a field is mm/h;
is the average rain intensity of the rainfall in a field, mm/h; beta is a rainfall intensity correction parameter and needs to be calibrated through the rainfall process of the flood field in the early stage; />
The rainfall characteristic factors such as total rainfall, rainfall intensity and the like are comprehensively embodied.
S4, establishing an SCS-CN three-water-source confluence model by coupling the improved flow generation module and the three-water-source confluence module in the S3;
the SCS-CN runoff generating module can obtain a process curve of the runoff depth of the research area, the runoff depth generated by the research area in each period can be obtained by carrying out differential processing on the process curve, and then the runoff depth is input into the three-water-source confluence module for confluence calculation; the three water sources refer to surface runoff, interflow and subsurface runoff and are divided by a parabolic free water storage curve and a free water storage reservoir.
In the confluence calculation, a unit line method is adopted in the surface water confluence calculation, a drainage basin is simulated as the series connection of n linear reservoirs by using the unit line, the confluence calculation of the interflow and the subsurface runoff adopts the linear reservoir method, a delay algorithm is adopted in the river network confluence, and the calculation formula is as follows:
Q3(I)＝CS×Q3(I-1)+(1-CS)×QT3(I-L) (5)
QT3(I-L)＝QS(I-L)+QG(I-L)+QI(I-L) (6)
in the formula: q3 (I) is the unit area river network confluence of the I time period, m 3 S; CS is the river network water flow fading coefficient; l is a convergence time parameter h; QS surface runoff, m 3 S; QG subsurface runoff, m 3 S; QI interflow, m 3 S; QT3 (I-L) is the sum of surface runoff, subsurface runoff and interflow in the I-L periods, m 3 /s。
S5, on the basis of the SCS-CN three-water-source convergence model obtained in the S4, constructing a dynamic convergence time lag parameter through a support vector machine model to improve the SCS-CN three-water-source convergence model;
when the rainfall space is uneven or the basin shape is narrow in the research basin, the basin confluence time is changed to a certain extent frequently, in the past, the confluence time of most hydrological models is kept unchanged, and the great deviation between the flood peak time predicted by the hydrological models and the actual flood peak time is easily caused;
based on the phenomenon, a dynamic convergence time parameter (L) is constructed by using a support vector machine model, factors which possibly influence the convergence time of the drainage basin are screened out firstly, then principal component analysis is carried out, main influence factors are determined, and final regression variables comprise the length of the rainfall center of a research area from the section of the drainage basin, the gradient of the rainfall center, total rainfall, early influence rainfall and the length of a total river channel;
the basic idea of the support vector machine implementation is to map a low-dimensional nonlinear space to a high-dimensional feature space by using a kernel function, so that the low-dimensional nonlinear hydrological data can be divided in the high-dimensional space, and through a series of derivation, a regression function f (x) of the low-dimensional nonlinear hydrological data can be finally expressed as:
in the formula: (alpha i ,α i * ) Is a lagrange multiplier; b is an offset; k (x, x) i ) The radial basis function from the input space to the high-order feature space is a kernel function; σ is the expansion constant of the radial basis function, representing the radial range of action of the function.
After a functional relation between the convergence time parameter and the regression variable is established through a support vector machine model, the convergence time parameter of the flood in the field is predicted according to the actually measured station data and underlying surface data (including gradient, river length and the like), and then the convergence time parameter is substituted into the formulas (5) and (6), so that an improved SCS-CN three-water-source convergence model can be obtained.
S6, forecasting a flood process line of the outlet section of the drainage basin by using the actually measured rainfall data;
firstly, inputting measured rainfall data and production confluence parameters of S2, S3 and S4 into an SCS-CN three-water-source confluence model in S4 to obtain the runoff rate of each computational grid in a research area, and then inputting the runoff rate into the three-water-source confluence model improved in S5 to obtain a flood process line of the outlet section of the drainage basin.
A flood forecasting system based on a support vector machine model, comprising: the system comprises a data acquisition unit, a runoff generating module parameter solving unit, a runoff generating module improving unit, an SCS-CN three-water-source confluence model building unit, an SCS-CN three-water-source confluence model improving unit and a flood forecasting unit;
wherein:
a data acquisition unit: collecting hydrological meteorological data of a research area;
a parameter solving unit of the runoff generating module: determining a production flow module parameter CN value of the SCS-CN model according to the collected hydrological meteorological data;
the birth flow module improves the unit: improving an SCS-CN model runoff generating module by utilizing rainfall characteristic factors and a runoff generating module parameter CN value;
SCS-CN three water source confluence model construction unit: establishing an SCS-CN three-water-source confluence model by coupling an improved production flow module and a three-water-source confluence module;
SCS-CN three water source confluence model improvement unit: constructing a dynamic convergence time parameter through a support vector machine model to improve an SCS-CN three-water-source convergence model;
flood forecasting unit: forecasting a flood process line of the outlet section of the drainage basin by using the actually measured rainfall data;
example (b):
the flood forecasting method is applied by taking a certain place as a specific embodiment:
the research area is located in a village basin, and the schematic diagram of the research basin is shown in figure 2; the hydrological station is positioned at the downstream of a water area, the water collecting area is 745km2, and the annual average rainfall is about 500 mm. The village drainage basin has 5 rainfall stations in total, the density is about 149km 2/station, the average gradient of the drainage basin is 9.19 percent, the village drainage basin belongs to temperate monsoon climate and semiarid regions, and the land utilization type is mainly cultivated land and forest land.
Acquiring hydrological meteorological data of the village basin in 1957-2004 (in 6-9 month flood season), including rainfall, runoff, evaporation, land utilization type and other data, originating from the provincial hydrological bureau, and finally screening 25 floods after three-property examination of the rainfall runoff data, wherein the 25 floods include the historical maximum floods of the village hydrological station (the peak flow is 4520m & lt 3 & gt/s, and the peak water level is 57.78 m).
And downloading DEM data of the village drainage basin from the geospatial data cloud, and performing operations such as projection, resampling, hole filling, flow direction extraction, flow rate extraction and the like in Arcgis software to further obtain a geospatial distribution map of the research area. The land use types within the area under study are then resampled to match the grids under each soil type.
And determining the village basin as a B-type hydrological soil group according to the data such as the soil saturation infiltration rate, the minimum infiltration rate, the soil texture and the like of the village basin, and comprehensively determining the initial CN value (the soil humidity is normal) of the research area through the land utilization type of the research area. The soil data was derived from the FAO world soil type database with a spatial resolution of 1000m × 1000m, the land use data was derived from global surface cover products (GlobeLand 30) with a spatial resolution of 30m × 30m, and the distribution is shown in fig. 3.
The SCS-CN model determines soil wetting conditions according to the total rainfall of the previous 5 days and places where mutation is likely to occur, so that the soil wetting condition division regions of the original model are divided again by adopting K-mean clustering, and CN is subjected to CN mean clustering 1 、CN 2 And CN 3 The mutation interval between them is interpolated, CN 1 The soil was compared for the CN value at drought, and CN 3 The CN values of the research area under different early-stage influences on rainfall are shown in table 1:
TABLE 1 SCS-CN model CN value
After determining the CN values for different sessions, the session radial flow may be calculated by:
in the formula: p is total rainfall, mm; q is the surface runoff, mm; i is a Is the initial loss, mm; s is potential water storage capacity, mm.
And the initial loss and the potential water storage capacity in the flow production module can be obtained by adopting the following formula:
I a ＝λS (2)
in the formula: and lambda is the initial loss rate.
In semiarid regions, the water level of soil is low, the water shortage of soil is large, after a rain fall, the drainage basin can hardly reach the field water capacity, the rainfall intensity exceeds the infiltration intensity, and then the drainage can be produced, so that characteristic factors such as the rainfall intensity and the like need to be added into a drainage module to improve the prediction precision of the model in the drainage basin in the village, and the revised drainage calculation formula is as follows:
in the formula: I.C. A 60 The maximum 60min rain intensity in the rainfall process of a field is mm/h;
the average rain intensity of field rainfall is mm/h; beta is a raininess correction parameter; />
The rainfall characteristic factors such as total rainfall, rainfall intensity and the like are comprehensively embodied.
The process curve of the runoff depth of the research area can be obtained through the steps, the runoff depth generated by the village drainage basin in each time period can be obtained through differential processing, and then the runoff depth is input into the three-water-source confluence module for confluence calculation.
The constant term (B) and the initial loss rate of the empirical relation of the potential water storage capacity and the CN value have a certain influence on the runoff producing module, so in this example, the constant term and the initial loss rate of the empirical relation of the potential water storage capacity and the CN value in the runoff producing module are calibrated, meanwhile, the sensitive parameters KI, KG, CS, CI, CG and SM in the confluence module are calibrated, 16-1979-1957 floods are used for calibrating the model parameters, the rest 9-1979 floods are used for verifying the model, the parameter calibration result is shown in table 2, and the field simulation result is shown in table 3:
TABLE 2 SCS-CN model parameter values
Table 3. Statistics index of simulation result of SCS-CN model in village basin
And dividing the total runoff into three types of surface runoff, interflow and subsurface runoff through a parabolic free water storage curve, and respectively calculating by a confluence module. Meanwhile, according to the concept of full runoff production, runoff is possibly generated only in the runoff production area, the rainfall in the runoff production area generates runoff completely and enters the reservoir, and the rainfall becomes the supply amount of the free water reservoir.
In the confluence calculation, a unit line method is adopted in the surface water confluence calculation, a drainage basin is simulated as the series connection of n linear reservoirs by using the unit line, the confluence calculation of the interflow and the subsurface runoff adopts the linear reservoir method, a delay algorithm is adopted in the river network confluence, and the calculation formula is as follows:
Q3(I)＝CS×Q3(I-1)+(1-CS)×QT3(I-L) (5)
QT3(I-L)＝QS(I-L)+QG(I-L)+QI(I-L) (6)
in the formula: q3 (I) is the unit area river network convergence quantity m of the I-th time period 3 S; CS is the river network water flow fading coefficient; l is a convergence time parameter, h; QS surface runoff, m 3 S; QG subsurface runoff, m 3 S; QI interflow, m 3 S; QT3 (I-L) is the sum of surface runoff, subsurface runoff and interflow in the I-L periods, m 3 /s。
If the research area is arranged in a narrow watershed or the watershed area is large, the spatial distribution condition of rainfall is often required to be considered, in the example, a support vector machine model is used for constructing dynamic convergence time lag parameters, and relevant variables comprise the length of a rainfall center from a watershed section, the gradient of the rainfall center, the total river length, the total rainfall, early-stage influence rainfall and other hydrometeorology characteristic parameters. After a series of derivation, the regression function can be finally expressed as:
in the formula: (alpha i ,α i * ) Is a lagrange multiplier; b is an offset; k (x, x) i ) The radial basis function from the input space to the high-order feature space is a kernel function; σ is the expansion constant of the radial basis function, representing the radial range of action of the function.
The support vector machine model is commonly used for solving the machine learning problem under the condition of small samples, and can simplify the problems of classification, regression and the like. Because the support vector machine model is a supervised machine learning method and requires certain early-stage data to construct a database, a model is established by adopting partial early-stage flood data, the actual value of the model is a convergence time parameter when the simulated flood peak time of the hydrological model is the same as the actual flood peak time, and the predicted value is closer to 1:1 line, the better the model prediction result, and the prediction result is shown in fig. 4.
In conclusion, the SCS-CN three-water source flow model can be constructed through the above steps, and the simulation results of part of the runs are shown in fig. 5.
In the description herein, references to the description of "one embodiment," "an example," "a specific example" or the like are intended to mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
The foregoing shows and describes the general principles, principal features, and advantages of the invention. It will be understood by those skilled in the art that the present invention is not limited to the embodiments described above, which are given by way of illustration of the principles of the present invention, but that various changes and modifications may be made without departing from the spirit and scope of the invention, and such changes and modifications are within the scope of the invention as claimed.
Claims (10)
1. A flood forecasting method based on a support vector machine model is characterized by comprising the following steps:
s1, collecting hydrological meteorological data of a research area;
s2, determining a current generation module parameter CN value of the SCS-CN model according to the collected hydrological meteorological data;
s3, improving the SCS-CN model runoff producing module by utilizing the rainfall characteristic factor and the runoff producing module parameter CN value;
s4, establishing an SCS-CN three-water-source confluence model by coupling the improved flow generation module and the three-water-source confluence module in the S3;
s5, constructing a dynamic convergence time parameter through a support vector machine model to improve an SCS-CN three-water-source convergence model;
and S6, forecasting a flood process line of the cross section of the drainage basin outlet by using the actually measured rainfall data.
2. A flood forecasting method based on a support vector machine model according to claim 1, characterized in that the hydrometeorology data collected in S1 includes: rainfall intensity, runoff, soil type, and land use type.
3. A flood forecasting method based on a support vector machine model according to claim 2, characterized in that in S2, the step of determining parameters of a runoff generating module is:
s21, determining a hydrological soil group through soil type data of a research area, and dividing according to infiltration capacity; resampling the land utilization type of the research area on an Arcgis platform, matching the land utilization type with grids under each soil type, calculating a CN value in each grid, carrying out weighted average on the CN value of each grid, and finally determining a basin comprehensive CN value;
s22, dividing the soil humidity condition into three grades of drought, normal and humidity according to the total rainfall of the previous 5 days; the interval of the soil wetting condition is divided again by adopting K-mean clustering, the total rainfall in the previous 5 days is the early-stage influence rainfall, and the threshold value is set to be 120mm;
s23, converting the CN value according to the actual soil humidity condition of the research area, and converting the CN value 1 、CN 2 And CN 3 And carrying out interpolation processing on the mutation intervals to obtain CN values under different early-stage influences on rainfall.
4. The flood forecasting method based on the support vector machine model according to claim 3, wherein in S3, the SCS-CN runoff generating model is improved by adding the influence of rainfall characteristic factors on the basis of the original SCS-CN model, and the specific steps are as follows:
s31, in the runoff generating module, early rainfall needs to supplement the initial water shortage of the watershed, if the total rainfall does not meet the initial water shortage, the watershed does not generate runoff, and a runoff calculation formula formed by rainfall in a field is as follows:
in the formula: p is total rainfall, mm; q is the surface runoff, mm; i is a Is the initial loss, mm; s is potential water storage capacity, mm;
s32, calculating the initial loss and the potential water storage capacity according to the formula:
I a ＝λS (2)
in the formula: lambda is the initial loss rate; CN is a dimensionless parameter;
s33, adding characteristic factors such as rainfall intensity and the like into the runoff producing module, wherein the revised runoff producing calculation formula is as follows:
in the formula: I.C. A 60 The maximum 60min rain intensity in the rainfall process of a field is mm/h;
is the average rain intensity of the rainfall in a field, mm/h; beta is a raininess correction parameter; />
The rainfall characteristic factor comprehensive embodiment.
5. The flood forecasting method based on the support vector machine model as claimed in claim 1, wherein in S4, a process curve of runoff depth of the research area is obtained through the SCS-CN runoff generating module, the runoff depth generated in the research area at each time period is obtained through differential processing, and then the runoff depth is input to the three-water-source confluence module for confluence calculation.
6. A flood forecasting method based on a support vector machine model according to claim 5, characterized in that the three water sources refer to: surface runoff, interflow runoff and subsurface runoff, and is divided by a parabolic free water storage curve and a free water storage reservoir.
7. A flood forecasting method based on a support vector machine model according to claim 5, characterized in that in the confluence calculation, a surface water confluence calculation adopts a unit line method, and a watershed is simulated as the series connection of n linear reservoirs by using a unit line; the confluence calculation of the interflow and the subsurface runoff adopts a linear reservoir method; the river network convergence adopts a hysteresis algorithm, and the calculation formula is as follows:
Q3(I)＝CS×Q3(I-1)+(1-CS)×QT3(I-L) (5)
QT3(I-L)＝QS(I-L)+QG(I-L)+QI(I-L) (6)
in the formula: q3 (I) is the unit area river network convergence quantity m of the I-th time period 3 S; CS is the river network water flow fading coefficient; l is a convergence time parameter, h; QS surface runoff, m 3 S; QG subsurface runoff, m 3 S; QI interflow, m 3 S; QT3 (I-L) is the sum of surface runoff, groundwater runoff and interflow in the I-L periods, m 3 /s。
8. A flood forecasting method based on a support vector machine model according to claim 1, characterized in that the regression variables of the dynamic confluence time parameter comprise: the length of the distance between a rainfall center and a watershed section in a research area, the gradient of the rainfall center, total rainfall, early-stage influence rainfall and total river length; the regression function f (x) of the dynamic convergence time parameter is:
in the formula: (alpha.) of i ,α i * ) Is a lagrange multiplier; b is an offset; k (x, x) i ) The radial basis function from the input space to the high-order feature space is a kernel function; σ is the spreading constant of the radial basis function, representing the radial range of action of the function.
9. The flood forecasting method based on the support vector machine model of claim 1, wherein in S6, firstly, the measured rainfall data and the production convergence parameters of S2, S3 and S4 are input into the SCS-CN three-water-source convergence model in S4 to obtain the runoff of each computational grid in the research area, and then the runoff is input into the three-water-source convergence model improved in S5 to obtain the flood process line of the outlet section of the basin.
10. A flood forecasting system based on a support vector machine model, comprising:
a data acquisition unit: collecting hydrological meteorological data of a research area;
a parameter solving unit of the runoff generating module: determining a production flow module parameter CN value of the SCS-CN model according to the collected hydrological meteorological data;
the runoff producing module improving unit: improving an SCS-CN model runoff generating module by utilizing rainfall characteristic factors and a runoff generating module parameter CN value;
SCS-CN three water source confluence model construction unit: establishing an SCS-CN three-water-source confluence model by coupling an improved production flow module and a three-water-source confluence module;
SCS-CN three water source confluence model improvement unit: constructing a dynamic convergence time parameter through a support vector machine model to improve an SCS-CN three-water-source convergence model;
flood forecasting unit: and forecasting a flood process line of the outlet section of the drainage basin by using the actually measured rainfall data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211363780.7A CN115905963A (en) | 2022-11-02 | 2022-11-02 | Flood forecasting method and system based on support vector machine model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211363780.7A CN115905963A (en) | 2022-11-02 | 2022-11-02 | Flood forecasting method and system based on support vector machine model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115905963A true CN115905963A (en) | 2023-04-04 |
Family
ID=86477060
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211363780.7A Pending CN115905963A (en) | 2022-11-02 | 2022-11-02 | Flood forecasting method and system based on support vector machine model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115905963A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117036099A (en) * | 2023-08-14 | 2023-11-10 | 上海勘测设计研究院有限公司 | Spatially fine accounting method and spatially fine accounting system suitable for vegetation flood regulation |
-
2022
- 2022-11-02 CN CN202211363780.7A patent/CN115905963A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117036099A (en) * | 2023-08-14 | 2023-11-10 | 上海勘测设计研究院有限公司 | Spatially fine accounting method and spatially fine accounting system suitable for vegetation flood regulation |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Li et al. | Effects of urbanization on direct runoff characteristics in urban functional zones | |
| Jia et al. | Development of the WEP-L distributed hydrological model and dynamic assessment of water resources in the Yellow River basin | |
| Hou et al. | Rapid forecasting of urban flood inundation using multiple machine learning models | |
| Notter et al. | Impacts of environmental change on water resources in the Mt. Kenya region | |
| Schumann et al. | Application of a geographic information system for conceptual rainfall–runoff modeling | |
| CN113610264B (en) | Refined power grid typhoon flood disaster prediction system | |
| Wu et al. | Simulation of soil loss processes based on rainfall runoff and the time factor of governance in the Jialing River Watershed, China | |
| Decharme et al. | Impact of an exponential profile of saturated hydraulic conductivity within the ISBA LSM: simulations over the Rhône basin | |
| CN113742910A (en) | Reservoir water inflow early warning and forecasting method and system based on flood forecasting of medium and small watershed | |
| Guan et al. | Past variations and future projection of runoff in typical basins in 10 water zones, China | |
| Jin et al. | Effects of land-use data resolution on hydrologic modelling, a case study in the upper reach of the Heihe River, Northwest China | |
| CN105893770A (en) | Method for quantifying influence on basin water resources by climate change and human activities | |
| Kalma et al. | Predicting catchment‐scale soil moisture status with limited field measurements | |
| Yan et al. | Modified palmer drought severity index based on distributed hydrological simulation | |
| CN115130396A (en) | Distributed hydrological model modeling method for riverway type reservoir area | |
| Kittel et al. | Informing hydrological models of poorly gauged river catchments–a parameter regionalization and calibration approach | |
| Nakayama et al. | Evaluation of uneven water resource and relation between anthropogenic water withdrawal and ecosystem degradation in Changjiang and Yellow River basins | |
| Canuto et al. | Influence of reservoir management on Guadiana streamflow regime | |
| CN115905963A (en) | Flood forecasting method and system based on support vector machine model | |
| Hassan et al. | Studying the impact of climate change on the average temperature using CanESM2 and HadCM3 modelling in Iraq | |
| Leta et al. | Parameter uncertainty analysis for streamflow simulation using SWAT model in Nashe Watershed, Blue Nile River Basin, Ethiopia | |
| Kaur et al. | Impact of climate change on groundwater levels in Sirhind Canal Tract of Punjab, India | |
| Wu et al. | The impact of human activities on blue-green water resources and quantification of water resource scarcity in the Yangtze River Basin | |
| Wu et al. | Impact of surface and underground water uses on streamflow in the upper-middle of the Weihe River basin using a modified WetSpa model | |
| CN115965243A (en) | Model chain for dynamic evaluation of flood risk under land use change |
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 |