CN114117953A - Hydrological model structure diagnosis method, runoff forecasting method and device - Google Patents
Hydrological model structure diagnosis method, runoff forecasting method and device Download PDFInfo
- Publication number
- CN114117953A CN114117953A CN202111312143.2A CN202111312143A CN114117953A CN 114117953 A CN114117953 A CN 114117953A CN 202111312143 A CN202111312143 A CN 202111312143A CN 114117953 A CN114117953 A CN 114117953A
- Authority
- CN
- China
- Prior art keywords
- model
- hydrological
- parameters
- runoff
- time
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 66
- 238000003745 diagnosis Methods 0.000 title claims abstract description 29
- 238000013277 forecasting method Methods 0.000 title claims abstract description 7
- 238000012216 screening Methods 0.000 claims abstract description 41
- 238000012937 correction Methods 0.000 claims abstract description 28
- 238000004088 simulation Methods 0.000 claims abstract description 24
- 230000000694 effects Effects 0.000 claims abstract description 17
- 230000008569 process Effects 0.000 claims abstract description 11
- 230000007547 defect Effects 0.000 claims abstract description 9
- 238000004422 calculation algorithm Methods 0.000 claims description 15
- 238000010206 sensitivity analysis Methods 0.000 claims description 13
- 238000012360 testing method Methods 0.000 claims description 12
- 230000007847 structural defect Effects 0.000 claims description 9
- 230000002596 correlated effect Effects 0.000 claims description 8
- 238000001914 filtration Methods 0.000 claims description 8
- 238000005457 optimization Methods 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 claims description 7
- 230000008859 change Effects 0.000 claims description 6
- 238000004891 communication Methods 0.000 claims description 6
- 230000000875 corresponding effect Effects 0.000 claims description 6
- 238000011156 evaluation Methods 0.000 claims description 6
- 238000010220 Pearson correlation analysis Methods 0.000 claims description 4
- 238000001704 evaporation Methods 0.000 claims description 4
- 230000008020 evaporation Effects 0.000 claims description 4
- 230000008595 infiltration Effects 0.000 description 5
- 238000001764 infiltration Methods 0.000 description 5
- 238000001556 precipitation Methods 0.000 description 5
- 239000002689 soil Substances 0.000 description 3
- 238000010219 correlation analysis Methods 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000001747 exhibiting effect Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/28—Design optimisation, verification or simulation using fluid dynamics, e.g. using Navier-Stokes equations or computational fluid dynamics [CFD]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/14—Force analysis or force optimisation, e.g. static or dynamic forces
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Mathematical Physics (AREA)
- Fluid Mechanics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Computing Systems (AREA)
- Pure & Applied Mathematics (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- Algebra (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a hydrological model structure diagnosis method, a runoff forecasting method and a device based on time-varying parameters, which can construct a model structure capable of accurately reflecting the hydrological physical process of a basin of a target area, thereby more accurately simulating and forecasting runoff. The hydrological model structure diagnosis method comprises the following steps: step 1, collecting hydrological data of a target drainage basin; step 2, screening sensitive parameters; step 3, calibrating parameters of the hydrological model to determine non-sensitive parameters, and assimilating data to identify time-varying sequences of the sensitive parameters; step 4, analyzing influence factors of the sensitive parameter time-varying sequence; step 5, diagnosing modules with possible defects of the model according to the influence factors, and selecting other generalized structures of the modules to form a model to be tested; step 6, identifying a time-varying sequence of sensitive parameters of the model to be tested; step 7, when the time variation of the parameters is weakened and the simulation effect is improved, determining that the corresponding model to be tested is a correction model with a better structure; and 8, determining a final model.
Description
Technical Field
The invention belongs to the technical field of hydrological models, and particularly relates to a hydrological model structure diagnosis method based on time-varying parameters, a runoff forecasting method and a runoff forecasting device.
Background
The runoff forecasting is basic data for developing, managing and utilizing water resources in a drainage basin, a drainage basin hydrological model is used as an effective tool for describing rainfall-runoff relation, and the reduction of uncertainty of the model is of great importance for improving runoff forecasting capability.
Data, model parameters, and model structure are three major sources of uncertainty for hydrological models, with less uncertainty being studied for model structures that are difficult to quantify. In the hydrological model building process, the complex hydrological process in the nature is generalized, different hydrological meteorological conditions of different watersheds are different, and hydrological laws are different, so that a generalized model structure is difficult to find and is suitable for all watershed conditions, and the uncertainty of the model structure is generated.
As the hydrological and meteorological conditions of different watersheds are different, theoretically, the watershed characteristics are combined to establish a model for each watershed independently. The problems of large workload, lack of theoretical guidance and the like exist in the process of establishing a new model aiming at a specific watershed. At present, a commonly used method is to construct a time-varying function of a hydrological model parameter according to factors such as basin hydrological weather and the like, so as to achieve the effects of expanding the hydrological model structure and compensating the defects of the model structure. However, the problem that the physical mechanism is unclear in constructing the time-varying function formula of the hydrological model parameters exists, the conventional function form cannot perfect the hydrological physical process, and the space for improving the simulation precision still exists; furthermore, time-varying parameters are also often not readily available in production practice.
Disclosure of Invention
The present invention is made to solve the above problems, and an object of the present invention is to provide a method for diagnosing a hydrological model structure based on time-varying parameters, a method for forecasting runoff, and a device thereof, which are capable of constructing a model structure that accurately reflects a target basin hydrological physical process, so as to more accurately simulate and forecast runoff and obtain runoff forecast data closer to actual conditions.
In order to achieve the purpose, the invention adopts the following scheme:
< method for diagnosing model Structure >
As shown in fig. 1, the present invention provides a method for diagnosing a hydrological model structure, which is characterized by comprising the following steps:
step 1, selecting a hydrological model M as an initial model, and collecting hydrological data of a target basin and the same-period hydrological meteorological factor data;
step 3, using an optimization algorithm to rate all parameters of the hydrological model M at a rate period to obtain constant parameters, taking values of non-sensitive parameters as constant parameter values, using a data assimilation method to identify time-varying sequences of the sensitive parameters, and assimilating a runoff observation value into the hydrological model by using a Kalman filtering data assimilation algorithm to identify the model parameters at each calculation step;
step 4, carrying out Pearson correlation analysis on the time variation sequence of the sensitive parameter of the hydrological model M and the hydrological meteorological factor data, and screening factors highly correlated with the time variation sequence of the sensitive parameter as influence factors;
step 5, analyzing hydrological modules possibly having defects in the model M according to the screened influence factors, wherein the hydrological modules have various different generalization methods in other models, and other generalization methods are selected as alternative structures to form the hydrological model to be tested { M }1,M2,…,Mm}；
and 7, analyzing the applicability of each hydrological model from two points, namely analyzing the initial model M and the test model { M }1,M2,…,MmThe parameters of the model M are time-varyingiThe time variation degree of the parameters is weakened, and the model M is illustratediThe compensation effect of the medium-time-varying parameters on the structural defects is weakened, and the evaluation indexes NSE and NSE adopting hydrologic predictionlog、KGE、VE、KGESRMTo evaluateEffect of model simulation, MiSimulation index promotion explanation model MiThe structure of (2) has stronger applicability in the drainage basin; selecting a model M with reduced parameter time variation degree and improved simulation indexiAs a correction model;
step 8, judging the correction model MiWhether the sensitive parameter is stable or not, and if so, determining MiThe final model of the model structure diagnosis is obtained; otherwise, returning to the step 4, and taking the corrected model Mi as the hydrological model M to be executed in sequence.
Preferably, the hydrological model structure diagnosis method provided by the invention can also have the following characteristics: in step 2, the sensitivity analysis method adopts MORIS sensitivity analysis, each parameter changes with the same relative variation, the more sensitive parameter has larger influence on the output of the model, the sum of squares of errors is taken as a target function to screen the sensitive parameter,
in the formula, QoiIs the observed runoff at the ith moment, QsiThe simulated runoff at the ith moment, and n is the runoff sequence length.
Preferably, the hydrological model structure diagnosis method provided by the invention can also have the following characteristics: the hydrological data collected in step 1 include rainfall, evaporation, runoff data;
the hydrometeorology factors include rainfall (P), potential evapotranspiration (E), relative humidity (Wet), sunshine hours (Sun), average temperature (T), maximum temperature (T)max) Minimum temperature (T)min) A wind speed (Vw) factor.
Preferably, the hydrological model structure diagnosis method provided by the invention can also have the following characteristics: in step 7, the indices NSE and NSE are evaluatedlog、KGE、VE、KGESRMRespectively adopting the following formula to calculate:
in the formula, QoiIs the observed runoff at the ith moment, QsiIs the simulated runoff at the ith moment,
in order to observe the average value of the runoff,
taking the sequence average value after logarithm of observed runoff, wherein n is the runoff sequence length, k is the number of model parameters, r is the correlation coefficient of the observed runoff and the simulated runoff, and muoAnd musMean values, σ, of observed runoff and simulated runoff, respectivelyoAnd σsStandard deviations of observed runoff and simulated runoff are respectively.
Preferably, the hydrological model structure diagnosis method provided by the invention can also have the following characteristics: in step 7, when there are a plurality of models with reduced time variation degree of parameter, the model with the minimum time variation degree of parameter is selected as the correction model.
Preferably, the hydrological model structure diagnosis method provided by the invention can also have the following characteristics: in step 8, smooth means that the parameter sequence trend, periodicity and change points are not significant.
< method for predicting runoff >
The invention also provides a runoff forecasting method based on the hydrological model structure diagnosis, which is characterized by comprising the following steps: step 1 to step 8 described in < model structure diagnosis method > above; and 9, inputting the actually measured hydrological data into the final model to carry out runoff forecasting.
< apparatus >
Further, the present invention provides a runoff forecasting device based on hydrologic model structure diagnosis, which is characterized by comprising:
a data acquisition unit for acquiring hydrological data of a target basin and contemporaneous hydrological meteorological factor data;
the sensitive parameter screening part selects the hydrological model M as an initial model, performs sensitivity analysis on the parameters of the hydrological model M by using regular data of the utilization rate, and screens the sensitive parameters by using the sum of squares of errors as a target function;
the identification part is used for calibrating all parameters of the hydrological model M at a calibration period by using an optimization algorithm to obtain constant parameters, the non-sensitive parameters are taken as constant parameter values, the sensitive parameters are identified by using a data assimilation method to carry out time-varying sequence identification, and a set Kalman filtering data assimilation algorithm is used for assimilating the runoff observation values into the hydrological model to identify the model parameters at each calculation step;
an influence factor screening part which analyzes the Pearson correlation between the time-varying sequence of the sensitive parameter of the hydrological model M and the hydrological meteorological factor data and screens a factor highly correlated with the time-varying sequence of the sensitive parameter as an influence factor;
and a model selection part to be tested analyzes the hydrological module possibly having defects in the model M according to the screened influence factors, wherein the hydrological module has a plurality of different generalization methods in other models, and selects other generalization methods as alternative structures to form the hydrological model { M to be tested1,M2,…,Mm}；
A time-varying sequence recognition unit for recognizing the hydrological model { M }to be tested1,M2,…,MmTime-varying sequence identification of sensitive parameters is carried out;
a correction model determination part forThe applicability of each hydrological model is analyzed from two points, namely an initial model M and a test model { M }1,M2,…,MmThe parameters of the model M are time-varyingiThe time variation degree of the parameters is weakened, and the model M is illustratediThe compensation effect of the medium-time-varying parameters on the structural defects is weakened, and the evaluation indexes NSE and NSE adopting hydrologic predictionlog、KGE、VE、KGESRMTo evaluate the model simulation effect, MiSimulation index promotion explanation model MiThe structure of (2) has stronger applicability in the drainage basin; selecting a model M with reduced parameter time variation degree and improved simulation indexiAs a correction model;
a final model determining part for determining the modified model MiWhether the sensitive parameter is stable or not, and if so, determining MiThe final model of the model structure diagnosis is obtained;
a circular execution part, taking the correction model Mi as a hydrological model M, and sequentially executing an influence factor screening part, a model selection part to be tested, a time-varying sequence identification part, a correction model and a final model determination part until the final model determination part determines that the model is positive;
the runoff forecasting part inputs the actually measured hydrological data into the final model to carry out runoff forecasting to obtain a runoff forecasting result; and
and the control part is in communication connection with the data acquisition part, the sensitive parameter screening part, the identification part, the influence factor screening part, the to-be-tested model selection part, the time-varying sequence identification part, the correction model determination part, the final model determination part, the circulation execution part and the runoff forecasting part and controls the operation of the data acquisition part, the sensitive parameter screening part, the identification part, the influence factor screening part, the to-be-tested model selection part and the time-varying sequence identification part.
Preferably, the present invention provides a runoff forecasting device based on the hydrologic model structure diagnosis, which may further include: and the input display part is in communication connection with the data acquisition part, the sensitive parameter screening part, the identification part, the influence factor screening part, the model selection part to be tested, the time-varying sequence identification part, the correction model determination part, the final model determination part, the cycle execution part and the control part, and displays corresponding information according to an operation instruction input by a user.
Preferably, the runoff forecasting device based on the hydrological model structure diagnosis provided by the invention can also have the following characteristics: and the input display part is in communication connection with the data acquisition part, the sensitive parameter screening part, the identification part, the influence factor screening part, the model selection part to be tested, the time-varying sequence identification part, the correction model determination part, the final model determination part, the cycle execution part and the control part, and displays corresponding information according to an operation instruction input by a user.
Action and Effect of the invention
1. The method extracts information from the time-varying parameters of the hydrological model, and a modified model structure capable of reflecting the hydrological physical process of the target basin better is constructed by diagnosing the model structure through the time-varying parameters, so that a method for diagnosing the hydrological model structure is provided for the target basins with different hydrological and meteorological conditions, more accurate runoff simulation can be realized, the accuracy of runoff prediction is improved, and runoff prediction data closer to actual conditions is obtained;
2. according to the invention, after the model structure more suitable for the target watershed is obtained, the model can be calibrated only by using an optimization algorithm, and the hydrological forecast can be carried out, so that the method is very convenient and efficient to use in the hydraulic engineering practice.
Drawings
FIG. 1 is a flow chart of a hydrological model structure diagnosis method according to the present invention;
FIG. 2 is a graph of time-varying parameter identification results for models in accordance with an embodiment of the present invention;
FIG. 3 is a diagram illustrating correlation analysis between time-varying parameters of the SIMHYD model and hydrometeorology factors according to an embodiment of the present invention;
fig. 4 is a comparison graph of simulated runoff and measured runoff of the SIMHYD _ INF model and the SIMHYD model at regular intervals in the embodiment of the invention.
Fig. 5 is a comparison graph of simulated runoff and measured runoff of the SIMHYD _ INF model and the SIMHYD model during the inspection period in the embodiment of the invention.
Detailed Description
The following describes in detail specific embodiments of the method for diagnosing a hydrological model structure, the method for forecasting runoff, and the apparatus according to the present invention with reference to the drawings.
< example >
The ten-day river drainage basin is located in Shaanxi province in China, and the drainage basin area is about 6448km2And the average altitude is 200-3000 m. The whole length of the river is about 218km, and the river is discharged into Hanjiang to a hydrological station on a home terrace. The annual average precipitation of the ten-day river basin is about 850mm, and the annual average runoff is about 357 mm.
The embodiment provides a runoff forecasting method based on hydrological model structure diagnosis, which comprises the following steps:
step 1, selecting a hydrological model SIMHYD as an initial model, wherein parameters contained in the model are shown in a table 1. And (4) settling the hydrological data of the drainage basin of the embodiment, including precipitation, runoff and evaporation. And collecting contemporaneous hydrometeorological factor data, the hydrometeorological factors including rainfall (P), potential evapotranspiration (E), relative humidity (Wet), sunshine hours (Sun), average temperature (T), maximum temperature (T)max) Minimum temperature (T)min) Wind speed (Vw). The data used in the examples are daily data from 1/1991 to 31/2001 and 12.
TABLE 1SIMHYD model parameters
in the formula, QoiIs the observed runoff at the ith moment, QsiThe simulated runoff at the ith moment, and n is the runoff sequence length.
The most sensitive parameter of the SIMHYD model in the examples is the infiltration loss index SQ.
And 3, using an optimization algorithm to rate all parameters of the SIMHYD model at a rate period to obtain constant parameters and model simulation results, taking values of non-sensitive parameters in the SIMHYD model as constant parameter values obtained by the rate determination, using an ensemble Kalman filtering method to identify time-varying sequences of the sensitive parameters, and using the ensemble Kalman filtering algorithm to assimilate the runoff observation values into the hydrological model so as to identify the model parameters at each calculation step. The annual time-varying trend of the sensitive parameter SQ of the SIMHYD model in the embodiment is shown in fig. 2, which has a relatively obvious seasonality, and shows that the SIMHYD model has structural defects in the embodiment, and the time-varying parameter compensates the structural defects.
And 4, performing correlation analysis on the hydrological meteorological factor data and the time variation sequence of the sensitive parameter, and screening factors highly correlated with the time variation sequence of the sensitive parameter as influence factors. Fig. 3 shows pearson correlation coefficients of various factors and time-varying parameters, and influence factors (i r i >0.4) in which rainfall and relative humidity are most correlated.
Step 5, analyzing hydrological modules possibly having defects in the model M according to the screened influence factors, wherein the hydrological modules have various different generalized structures in different models, and selecting other generalized structures as alternative structures to form a hydrological model to be tested { M }1,M2,…,Mm}. Rainfall and relative humidity, as well as temperature variables exhibiting weak correlations, are all related to the evaporation process, so test model 1 is designated as the SIMHYD _ E model: precipitation in the SIMHYD model directly enters soil after vegetation is intercepted, participates in evapotranspiration and runoff generation, precipitation in the SIMHYD _ E model does not directly enter soil, and potential evapotranspiration is met before the precipitation enters the soil (the same as the GR4J model). Rainfall intensity directly affects the infiltration process, so test model 2 is designated as the SIMHYD _ INF model: the infiltration equation of the SIMHYD model is in an exponential function form, and the infiltration equation of the SIMHYD _ INF model is modified into a power function form (the same as the HBV model).
And 6, the annual variation trend of the sensitive parameter time-varying sequences of the test model 1SIMHYD _ E and the test model 2SIMHYD _ INF is shown in figure 2.
Step 7, comparing and analyzing the parameter time varying degrees (figure 2) of the initial model and each test model and simulatingEffect (table 2). In the test model 2, namely the SMHYD _ INF model, the time variation degree of the sensitive parameters is weakened, which shows that the compensation effect of the time-varying parameters in the SMHYD _ INF model on the structural defects is weakened; compared with the SIMHYD model and the SIMHYD _ E model, the SMHYD _ INF model evaluates indexes NSE and NSElog、KGE、VE、KGESRMThe SMHYD _ INF model is selected as the correction model.
And 8, enabling parameters of the SMHYD _ INF model to be stable, and determining the SMHYD _ INF model as a final model for the structural diagnosis of the model.
In an embodiment, time-varying parameter identification is performed on an existing hydrological model, and information is extracted from the time-varying parameters to modify the model structure. The result shows that in the laetime river basin, the SIMHYD model infiltration equation should be modified into a power function form from an exponential function form to form the SMHYD _ INF model, so that the SIMHYD model should be modified into the SMHYD _ INF model in the embodiment.
TABLE 2
And 9, inputting the actually measured hydrological data into the final model to carry out runoff forecasting. In this embodiment, as for example shown in fig. 4 to 5, the runoff simulation effect of the final model SMHYD _ INF and the original model SIMHYD after inputting the actually measured historical hydrological data is closer to the actual observed runoff, and the overall effect is better than that of the SIMHYD model, which indicates that the final model obtained by the method of the present invention can actually realize more accurate runoff simulation, improve the accuracy of runoff prediction, and obtain runoff forecast data closer to the actual situation.
The invention further provides a runoff forecasting device for automatically forecasting runoff based on the method, and the device comprises a data acquisition part, a sensitive parameter screening part, an identification part, an influence factor screening part, a model selection part to be tested, a time-varying sequence identification part, a correction model determination part, a final model determination part, a cycle execution part, an input display part and a control part.
The data acquisition unit acquires hydrological data of a target basin and contemporaneous hydrological meteorological factor data.
The sensitive parameter screening part selects the hydrological model M as an initial model, sensitivity analysis is carried out on the parameters of the hydrological model M by using data with regular utilization rate, and the sum of squares of errors is used as a target function to screen the sensitive parameters.
The identification part uses an optimization algorithm to rate all parameters of the hydrological model M at a rate period to obtain constant parameters, the non-sensitive parameters are taken as constant parameter values, the sensitive parameters use a data assimilation method to identify time-varying sequences, and a set Kalman filtering data assimilation algorithm assimilates the runoff observation values into the hydrological model to identify the model parameters at each calculation step.
And the influence factor screening part is used for carrying out Pearson correlation analysis on the time change sequence of the sensitive parameter of the hydrological model M and the hydrological meteorological factor data and screening a factor highly correlated with the time change sequence of the sensitive parameter as an influence factor.
The model selection part to be tested analyzes the hydrological module possibly having defects in the model M according to the screened influence factors, the hydrological module has various different generalization methods in other models, and other generalization methods are selected as alternative structures to form the hydrological model { M to be tested1,M2,…,Mm}。
Time-varying sequence recognition part for hydrological model { M to be tested1,M2,…,MmTime-varying sequence identification of sensitive parameters is performed.
The applicability of each hydrological model is analyzed from two points, namely, the initial model M and the test model { M }1,M2,…,MmThe parameters of the model M are time-varyingiThe time variation degree of the parameters is weakened, and the model M is illustratediThe compensation effect of the medium-time-varying parameters on the structural defects is weakened, and the evaluation indexes NSE and NSE adopting hydrologic predictionlog、KGE、VE、KGESRMTo evaluate the model simulation effect, MiSimulation index promotion explanation model MiThe structure of (2) has stronger applicability in the drainage basin; selecting time variation of parametersModel M with weakened degree and improved simulation indexesiAs a correction model.
The final model determining part corrects the model MiJudging whether the middle sensitive parameter is stable or not, and if so, determining MiIs the final model of the structure diagnosis of the model.
And the cycle execution part takes the correction model Mi as a hydrological model M and sequentially executes the influence factor screening part, the model selection part to be tested, the time-varying sequence identification part, the correction model and the final model determination part until the final model determination part determines that the model is positive under the condition that the final model determination part determines that the model is negative.
And the runoff forecasting part inputs the actually measured hydrological data into the final model to carry out runoff forecasting to obtain a runoff forecasting result.
The input display part is communicated with the data acquisition part, the sensitive parameter screening part, the identification part, the influence factor screening part, the model selection part to be tested, the time-varying sequence identification part, the correction model determination part, the final model determination part, the cycle execution part and the control part, and displays corresponding information according to an operation instruction input by a user. For example, the input display portion may display the final model determined by the final model determining portion according to the operation instruction, and may also display the runoff forecast result data of the runoff forecasting portion in a list manner according to the operation instruction, or display the runoff forecast result data on a topographic map of a corresponding target basin, or display a dynamic runoff change process of the target basin in a period of time in a dynamic evolution manner.
The control part is in communication connection with the data acquisition part, the sensitive parameter screening part, the identification part, the influence factor screening part, the model to be tested selection part, the time-varying sequence identification part, the correction model determination part, the final model determination part, the circulation execution part and the runoff forecasting part to control the operation of the data acquisition part, the sensitive parameter screening part, the identification part, the influence factor screening part, the model to be tested selection part and the time-varying sequence identification part.
The above embodiments are merely illustrative of the technical solutions of the present invention. The method for diagnosing the structure of the hydrological model, the method for forecasting the runoff and the device thereof are not limited to the contents described in the above embodiments, but are subject to the scope defined by the claims. Any modification or supplement or equivalent replacement made by a person skilled in the art on the basis of this embodiment is within the scope of the invention as claimed in the claims.
Claims (10)
1. A hydrological model structure diagnosis method is characterized by comprising the following steps:
step 1, selecting a hydrological model M as an initial model, and collecting hydrological data of a target basin and the same-period hydrological meteorological factor data;
step 2, carrying out sensitivity analysis on the parameter of the hydrological model M by using data with regular utilization rate, and screening the sensitive parameter by taking the sum of squares of errors as a target function;
step 3, using an optimization algorithm to rate all parameters of the hydrological model M at a rate period to obtain constant parameters, taking values of non-sensitive parameters as constant parameter values, using a data assimilation method to identify time-varying sequences of the sensitive parameters, and assimilating a runoff observation value into the hydrological model by using a Kalman filtering data assimilation algorithm to identify the model parameters at each calculation step;
step 4, carrying out Pearson correlation analysis on the time variation sequence of the sensitive parameter of the hydrological model M and the hydrological meteorological factor data, and screening factors highly correlated with the time variation sequence of the sensitive parameter as influence factors;
step 5, analyzing hydrological modules possibly having defects in the model M according to the screened influence factors, wherein the hydrological modules have various different generalization methods in other models, and other generalization methods are selected as alternative structures to form the hydrological model to be tested { M }1,M2,…,Mm}；
Step 6, carrying out comparison on the hydrological model to be tested { M1,M2,…,MmTime-varying sequence identification of sensitive parameters is carried out;
and 7, analyzing the applicability of each hydrological model from two points, namely analyzing the initial model M and the test model { M }1,M2,…,MmThe parameters of the model M are time-varyingiThe time variation degree of the parameters is weakened, and the model M is illustratediMiddle time-varying parameter pairThe compensation effect of the structural defect is weakened, and the evaluation indexes NSE and NSE of hydrologic prediction are adoptedlog、KGE、VE、KGESRMTo evaluate the model simulation effect, MiSimulation index promotion explanation model MiThe structure of (2) has stronger applicability in the drainage basin; selecting a model M with reduced parameter time variation degree and improved simulation indexiAs a correction model;
step 8, judging the correction model MiWhether the sensitive parameter is stable or not, and if so, determining MiThe final model of the model structure diagnosis is obtained; otherwise, returning to the step 4, and taking the corrected model Mi as the hydrological model M to be executed in sequence.
2. The method for diagnosing a structure of a hydrological model according to claim 1, wherein:
wherein, in the step 2, the sensitivity analysis method adopts MORIS sensitivity analysis, each parameter changes with the same relative variation, the more sensitive parameter has larger influence on the output of the model, the error square sum is used as a target function to screen the sensitive parameter,
in the formula, QoiIs the observed runoff at the ith moment, QsiThe simulated runoff at the ith moment, and n is the runoff sequence length.
3. The method for diagnosing a structure of a hydrological model according to claim 1, wherein:
wherein, the hydrological data collected in the step 1 comprises rainfall, evaporation and runoff data;
the hydrological meteorological factors include rainfall P, potential evapotranspiration E, relative humidity Wet, sunshine hours Sun, average temperature T, maximum temperature TmaxMinimum temperature TminAnd a wind speed factor Vw.
4. The method for diagnosing a structure of a hydrological model according to claim 1, wherein:
wherein, in step 7, the indexes NSE and NSE are evaluatedlog、KGE、VE、KGESRMRespectively adopting the following formula to calculate:
in the formula, QoiIs the observed runoff at the ith moment, QsiIs the simulated runoff at the ith moment,
in order to observe the average value of the runoff,
taking the sequence average value after logarithm of observed runoff, wherein n is the runoff sequence length, k is the number of model parameters, r is the correlation coefficient of the observed runoff and the simulated runoff, and muoAnd musMean values, σ, of observed runoff and simulated runoff, respectivelyoAnd σsStandard deviations of observed runoff and simulated runoff are respectively.
5. The method for diagnosing a structure of a hydrological model according to claim 1, wherein:
in step 7, when there are a plurality of models with reduced time variation degrees of the parameters, the model with the minimum time variation degree of the parameters is selected as the correction model.
6. The method for diagnosing a structure of a hydrological model according to claim 1, wherein:
in step 8, the term "smooth" means that the trend, periodicity and change points of the parameter sequence are not significant.
7. The runoff forecasting method based on the hydrological model structure diagnosis is characterized by comprising the following steps of:
step 1, selecting a hydrological model M as an initial model, and collecting hydrological data of a target basin and the same-period hydrological meteorological factor data;
step 2, carrying out sensitivity analysis on the parameter of the hydrological model M by using data with regular utilization rate, and screening the sensitive parameter by taking the sum of squares of errors as a target function;
step 3, using an optimization algorithm to rate all parameters of the hydrological model M at a rate period to obtain constant parameters, taking values of non-sensitive parameters as constant parameter values, using a data assimilation method to identify time-varying sequences of the sensitive parameters, and assimilating a runoff observation value into the hydrological model by using a Kalman filtering data assimilation algorithm to identify the model parameters at each calculation step;
step 4, carrying out Pearson correlation analysis on the time variation sequence of the sensitive parameter of the hydrological model M and the hydrological meteorological factor data, and screening factors highly correlated with the time variation sequence of the sensitive parameter as influence factors;
step 5, analyzing hydrological modules possibly having defects in the model M according to the screened influence factors, wherein the hydrological modules have various different generalization methods in other models, and other generalization methods are selected as alternative structures to form the hydrological model to be tested { M }1,M2,…,Mm}；
Step 6, carrying out comparison on the hydrological model to be tested { M1,M2,…,MmTime-varying sequence identification of sensitive parameters is carried out;
and 7, analyzing the applicability of each hydrological model from two points, namely analyzing the initial model M and the test model { M }1,M2,…,MmThe parameters of the model M are time-varyingiThe time variation degree of the parameters is weakened, and the model M is illustratediThe compensation effect of the medium-time-varying parameters on the structural defects is weakened, and the evaluation indexes NSE and NSE adopting hydrologic predictionlog、KGE、VE、KGESRMTo evaluate the model simulation effect, MiSimulation index promotion explanation model MiThe structure of (2) has stronger applicability in the drainage basin; selecting a model M with reduced parameter time variation degree and improved simulation indexiAs a correction model;
step 8, judging the correction model MiWhether the sensitive parameter is stable or not, and if so, determining MiEntering step 9 for the final model of the model structure diagnosis; otherwise, returning to the step 4, and taking the corrected model Mi as a hydrological model M to be sequentially executed;
and 9, inputting the actually measured hydrological data into the final model to carry out runoff forecasting.
8. Runoff forecasting device based on hydrological model structure diagnosis, characterized by includes:
a data acquisition unit for acquiring hydrological data of a target basin and contemporaneous hydrological meteorological factor data;
the sensitive parameter screening part selects the hydrological model M as an initial model, performs sensitivity analysis on the parameters of the hydrological model M by using regular data of the utilization rate, and screens the sensitive parameters by using the sum of squares of errors as a target function;
the identification part is used for calibrating all parameters of the hydrological model M at a calibration period by using an optimization algorithm to obtain constant parameters, the non-sensitive parameters are taken as constant parameter values, the sensitive parameters are identified by using a data assimilation method to carry out time-varying sequence identification, and a set Kalman filtering data assimilation algorithm is used for assimilating the runoff observation values into the hydrological model to identify the model parameters at each calculation step;
an influence factor screening part which analyzes the Pearson correlation between the time-varying sequence of the sensitive parameter of the hydrological model M and the hydrological meteorological factor data and screens a factor highly correlated with the time-varying sequence of the sensitive parameter as an influence factor;
and a model selection part to be tested analyzes the hydrological module possibly having defects in the model M according to the screened influence factors, wherein the hydrological module has a plurality of different generalization methods in other models, and selects other generalization methods as alternative structures to form the hydrological model { M to be tested1,M2,…,Mm}；
A time-varying sequence recognition unit for recognizing the hydrological model { M }to be tested1,M2,…,MmTime-varying sequence identification of sensitive parameters is carried out;
a corrected model determining part for analyzing the applicability of each hydrological model from two points, wherein the first point is to analyze the initial model M and the test model { M }1,M2,…,MmThe parameters of the model M are time-varyingiThe time variation degree of the parameters is weakened, and the model M is illustratediThe compensation effect of the medium-time-varying parameters on the structural defects is weakened, and the evaluation indexes NSE and NSE adopting hydrologic predictionlog、KGE、VE、KGESRMTo evaluate the model simulation effect, MiSimulation index promotion explanation model MiThe structure of (2) has stronger applicability in the drainage basin; selecting a model M with reduced parameter time variation degree and improved simulation indexiAs a correction model;
a final model determining part for determining the modified model MiWhether the sensitive parameter is stable or not, and if so, determining MiThe final model of the model structure diagnosis is obtained;
a cycle execution unit configured to, when the final model determination unit determines that the model is a true model, sequentially execute the influence factor screening unit, the model to be tested selection unit, the time-varying sequence recognition unit, the correction model, and the final model determination unit, using the correction model Mi as a hydrological model M, until the final model determination unit determines that the model is a true model;
the runoff forecasting part inputs the actually measured hydrological data into the final model to carry out runoff forecasting to obtain a runoff forecasting result; and
and the control part is in communication connection with the data acquisition part, the sensitive parameter screening part, the identification part, the influence factor screening part, the model selection part to be tested, the time-varying sequence identification part, the correction model determination part, the final model determination part, the cycle execution part and the runoff forecasting part and controls the operation of the data acquisition part, the sensitive parameter screening part, the identification part, the influence factor screening part, the model selection part to be tested and the time-varying sequence identification part.
9. A runoff forecasting apparatus based on a hydrological model structure diagnosis as claimed in claim 8, further comprising:
and the input display part is in communication connection with the data acquisition part, the sensitive parameter screening part, the identification part, the influence factor screening part, the model selection part to be tested and the time-varying sequence identification part, the correction model determination part, the final model determination part, the cycle execution part and the control part, and displays corresponding information according to an operation instruction input by a user.
10. A runoff forecasting device based on a hydrological model structure diagnosis as claimed in claim 9, wherein:
the input display part can display the final model determined by the final model determining part according to an operation instruction, and can also display the runoff forecast result data of the runoff forecasting part in a list mode according to the operation instruction, or display the runoff forecast result data on a topographic map of a corresponding target basin or display a runoff dynamic change process of the target basin in a period of time in a dynamic evolution mode.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111312143.2A CN114117953B (en) | 2021-11-08 | Hydrologic model structure diagnosis method, runoff forecasting method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111312143.2A CN114117953B (en) | 2021-11-08 | Hydrologic model structure diagnosis method, runoff forecasting method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114117953A true CN114117953A (en) | 2022-03-01 |
| CN114117953B CN114117953B (en) | 2024-06-28 |
Family
ID=
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114970171A (en) * | 2022-05-31 | 2022-08-30 | 水利部交通运输部国家能源局南京水利科学研究院 | Hydrological model considering uncertainty of runoff generating structure and method for quantifying influence on surface and underground hydrological process |
| CN115000471A (en) * | 2022-05-18 | 2022-09-02 | 西安交通大学 | Fuel cell catalyst layer prediction-analysis-optimization method based on machine learning |
| CN116561988A (en) * | 2023-04-18 | 2023-08-08 | 武汉大学 | Simulation method for water balance of river basin on daily scale |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106599571A (en) * | 2016-12-12 | 2017-04-26 | 武汉大学 | Basin hydrological model stage-based rating method considering flow and evaporation |
| CN108897977A (en) * | 2018-08-24 | 2018-11-27 | 河海大学 | A kind of uncertain attribution method of Runoff Evolution based on the simulation of big Regional Hydrologic |
| CN112883558A (en) * | 2021-01-27 | 2021-06-01 | 长江水利委员会水文局 | Hydrological model parameter time-varying form construction method |
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106599571A (en) * | 2016-12-12 | 2017-04-26 | 武汉大学 | Basin hydrological model stage-based rating method considering flow and evaporation |
| CN108897977A (en) * | 2018-08-24 | 2018-11-27 | 河海大学 | A kind of uncertain attribution method of Runoff Evolution based on the simulation of big Regional Hydrologic |
| CN112883558A (en) * | 2021-01-27 | 2021-06-01 | 长江水利委员会水文局 | Hydrological model parameter time-varying form construction method |
Non-Patent Citations (1)
| Title |
|---|
| 熊立华 等: "考虑植被和人类活动影响的水文模型参数时变特征分析", 《水科学进展》, 12 September 2018 (2018-09-12) * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115000471A (en) * | 2022-05-18 | 2022-09-02 | 西安交通大学 | Fuel cell catalyst layer prediction-analysis-optimization method based on machine learning |
| CN114970171A (en) * | 2022-05-31 | 2022-08-30 | 水利部交通运输部国家能源局南京水利科学研究院 | Hydrological model considering uncertainty of runoff generating structure and method for quantifying influence on surface and underground hydrological process |
| CN114970171B (en) * | 2022-05-31 | 2023-03-14 | 水利部交通运输部国家能源局南京水利科学研究院 | Hydrological model considering uncertainty of runoff generating structure and method for quantifying influence on surface and underground hydrological process |
| CN116561988A (en) * | 2023-04-18 | 2023-08-08 | 武汉大学 | Simulation method for water balance of river basin on daily scale |
| CN116561988B (en) * | 2023-04-18 | 2024-04-02 | 武汉大学 | Simulation method for water balance of river basin on daily scale |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111665575B (en) | Medium-and-long-term rainfall grading coupling forecasting method and system based on statistical power | |
| CN111310968A (en) | LSTM neural network circulation hydrological forecasting method based on mutual information | |
| CN111259481B (en) | Method and device for evaluating design of indoor light environment of building by integrating space information | |
| CN108876021B (en) | Medium-and-long-term runoff forecasting method and system | |
| CN105740991A (en) | Climate change prediction method and system for fitting various climate modes based on modified BP neural network | |
| CN111898258A (en) | Two-dimensional drought disaster assessment method driven by hydrologic cycle variation | |
| CN115423163A (en) | Method and device for predicting short-term flood events of drainage basin and terminal equipment | |
| CN115759488A (en) | Carbon emission monitoring and early warning analysis system and method based on edge calculation | |
| CN113484813B (en) | Intelligent ammeter fault rate prediction method and system under multi-environment stress | |
| CN113901384A (en) | Ground PM2.5 concentration modeling method considering global spatial autocorrelation and local heterogeneity | |
| CN115495991A (en) | Rainfall interval prediction method based on time convolution network | |
| CN115374995A (en) | Distributed photovoltaic and small wind power station power prediction method | |
| CN112287299A (en) | River health change quantitative attribution method, device and system | |
| CN115759445A (en) | Machine learning and cloud model-based classified flood random forecasting method | |
| CN114372631A (en) | Data-lacking area runoff prediction method based on small sample learning and LSTM | |
| CN110188922A (en) | A kind of long-term Runoff Forecast method in the RBF neural based on runoff mechanism | |
| CN116341391B (en) | Precipitation prediction method based on STPM-XGBoost model | |
| CN117370766A (en) | Satellite mission planning scheme evaluation method based on deep learning | |
| CN117520784A (en) | Groundwater level multi-step prediction method based on convolution attention long-short-term neural network | |
| Lu et al. | Uncertainty quantification of machine learning models to improve streamflow prediction under changing climate and environmental conditions | |
| CN114117953B (en) | Hydrologic model structure diagnosis method, runoff forecasting method and device | |
| CN111914487A (en) | Data-free regional hydrological parameter calibration method based on antagonistic neural network | |
| CN114117953A (en) | Hydrological model structure diagnosis method, runoff forecasting method and device | |
| CN115936545A (en) | Two-parameter monthly scale hydrological model building method based on hydrothermal complementary equation | |
| CN110009132A (en) | A kind of short-term electric load fining prediction technique based on LSTM deep neural network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant |