CN108920737A - A kind of particle filter assimilation method of hydrodynamic model, device and calculate equipment - Google Patents
A kind of particle filter assimilation method of hydrodynamic model, device and calculate equipment Download PDFInfo
- Publication number
- CN108920737A CN108920737A CN201810372803.8A CN201810372803A CN108920737A CN 108920737 A CN108920737 A CN 108920737A CN 201810372803 A CN201810372803 A CN 201810372803A CN 108920737 A CN108920737 A CN 108920737A
- Authority
- CN
- China
- Prior art keywords
- particle
- water
- depth
- water level
- hydrodynamic model
- 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
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
Abstract
The present invention provides a kind of particle filter assimilation method of hydrodynamic model, device and equipment is calculated, specially the calculating parameter in acquisition research area, and sets the size of the grid in research area and go out the boundary condition that becomes a mandarin；The particle of multiple equal weights is generated according to the parameter at current time；It is utilized respectively the streamflow regime and model roughness coefficient that each particle is included, and successively drives hydrodynamic model using the boundary condition that becomes a mandarin out simultaneously, realizes that hydrodynamic model exports the update from current time to subsequent time；Calculate the likelihood function value of each particle under subsequent time, and the weight of more new particle；The optimal estimation of calculating simulation water level, analogue flow rate and default roughness coefficien.Due to considering the Spatial-Temporal Variability of roughness coefficien in this programme, therefore the accuracy and reliability of hydrodynamic model output result can be made higher, the optimal estimation that state variable and roughness coefficien at different spaces can accurately be obtained improves the accuracy and reliability of hydrodynamic model output result.
Description
Technical field
The present invention relates to hydrotechnics fields, particle filter assimilation method, dress more particularly to a kind of hydrodynamic model
Set and calculate equipment.
Background technique
Hydrodynamic model is the important tool for simulating water sports state in the waters such as lake, river and seashore, can be effective
Support and management decision is widely used in the fields such as dam-break water flow simulation, flood forecasting and simulation of water quality prediction.In practice, water
Waters landform needed for dynamic model modeling and the data such as flow that become a mandarin out are often difficult to accurately measure, further, since hydrodynamic force
The important parameter roughness coefficien that degree of roughness is characterized in model can not be measured directly and grid is discrete and model structure etc. causes
Error, cause hydrodynamic model output result full of uncertainty, reduce hydrodynamic model output result accuracy and
Reliability.
Present inventor is in practice, it has been found that can reasonably incorporate water for observation data using data assimilation
Dynamic model, and using system mode and parameter is constantly updated in hydrodynamic model simulation process, model can be reduced not
Certainty, so as to finally improve physical process simulations or forecast precision.Particle filter algorithm is as a kind of sequence Bayes
Filtering data assimilates algorithm, approximate can obtain using state variable and parameter as the mathematics phase of the arbitrary function form of independent variable
It hopes, the posterior probability Density Distribution of state variable and parameter in model, Neng Gouying is indicated using a certain number of random particles
For any non-linear stochastic model.
Wherein, roughness coefficien by the degree of roughness of riverbed and bulkhead wall, streamflow regime, water plant and its floodage etc. because
The influence of element has significant Spatial-Temporal Variability, only considers that the data assimilation method of roughness coefficien time frame coefficient can not capture
Influence of its Spatial Variability to model simulation results, to be difficult to accurately obtain state variable and roughness coefficien at different spaces
Optimal estimation, and then cause hydrodynamic model output result accuracy and reliability it is poor.
Summary of the invention
In view of this, the present invention provides a kind of particle filter assimilation method of hydrodynamic model, device and equipment is calculated,
Accuracy and reliability to solve the problems, such as current hydrodynamic model output result is poor.
To solve the above-mentioned problems, the invention discloses a kind of particle filter assimilation methods of hydrodynamic model, including step
Suddenly：
The calculating parameter in acquisition research area, and set the size of the grid in the research area and go out the boundary condition that becomes a mandarin；
The grain of multiple equal weights is generated according to water level, flow and the default roughness coefficien of all grids at current time
Son obtains particle assembly；
It is utilized respectively the streamflow regime and model roughness coefficient that each particle is included in the particle assembly, and sharp simultaneously
Hydrodynamic model is successively driven with the boundary condition that becomes a mandarin out, realizes mimic water-depth and the simulation of the hydrodynamic model output
Update of the flow from current time to subsequent time；
Judge to whether there is water level observation in the mimic water-depth under current time, if it does not exist the water level observation,
Then directly execute described the step of realizing the hydrodynamic model output mimic water-depth and analogue flow rate；
The water level observation if it exists then calculates the likelihood function value of each particle under the subsequent time, and
The weight of the particle is updated using the likelihood function value；
Calculate the optimal estimation of the mimic water-depth, the analogue flow rate and default roughness coefficien；
Multinomial resampling is carried out to the particle, obtains new particle set；
The new particle set is replaced the particle assembly, while using the subsequent time as described current
Moment executes described the step of realizing the hydrodynamic model output mimic water-depth and analogue flow rate.Optionally, the calculating ginseng
Number includes that the research Bottom Altitudes in area, boundary become a mandarin some or all of flow, initial water level and initial flow.
Optionally, the default roughness coefficien is roughness coefficien priori value.
Optionally, described to include to particle progress multinomial resampling：
Multiple random numbers are generated in being uniformly distributed from [0,1] at random；
It is adopted using particle corresponding to the random number for meeting preset condition in the multiple random number as new sample point
Sample obtains the new particle set.
A kind of particle filter assimilation device of hydrodynamic model is additionally provided, including：
Parameter collection module, for acquire research area calculating parameter, and set it is described research area grid size and
Become a mandarin boundary condition out；
First particle sampler module, for water level, flow and the default roughness according to all grids at current time
Coefficient generates the particle of multiple equal weights, obtains particle assembly；
Model output module, the streamflow regime and model for being included for being utilized respectively each particle in the particle assembly
Roughness coefficien, and hydrodynamic model is successively driven using the boundary condition that becomes a mandarin out simultaneously, realize that the hydrodynamic model is defeated
The update of mimic water-depth and analogue flow rate from current time to subsequent time out；
Condition judgment module, for judging with the presence or absence of water level observation in the mimic water-depth under current time, if not depositing
In the water level observation, then the realization hydrodynamic model output mimic water-depth and analogue flow rate are directly executed；
First computing module then calculates each grain under the subsequent time for the water level observation if it exists
The likelihood function value of son, and update using the likelihood function value weight of the particle；
Second computing module is estimated for calculating the mimic water-depth, the analogue flow rate and the optimal of default roughness coefficien
Value；
Second particle sampler module obtains new particle set for carrying out multinomial resampling to the particle；
Particle collection replacement module, for the new particle set to be replaced the particle assembly, while will be described
Subsequent time executes the step of the output mimic water-depth and analogue flow rate of hydrodynamic model described in the sight as the current time
Suddenly.
Optionally, the calculating parameter includes that the research Bottom Altitude in area, boundary become a mandarin flow, initial water level and just
Some or all of beginning flow.
Optionally, the default roughness coefficien is roughness coefficien priori value.
Optionally, the second particle sampler module includes：
Random number generation unit, for generating multiple random numbers at random in being uniformly distributed from [0,1]；
Execution unit is sampled, is made for particle corresponding to the random number of preset condition will to be met in the multiple random number
It is sampled for new sample point, obtains the new particle set.
In addition, also a kind of calculating equipment, is provided with particle filter assimilation device as described above.
Additionally provide it is another calculate equipment, including at least one processor, memory and for will it is described at least one
The data/address bus that processor is connect with the memory signals, the memory are handled for storing computer program or instruction
Device is for executing the computer program or instruction, to realize following steps：
The calculating parameter in acquisition research area, and set the size of the grid in the research area and go out the boundary condition that becomes a mandarin；
The grain of multiple equal weights is generated according to water level, flow and the default roughness coefficien of all grids at current time
Son obtains particle assembly；
It is utilized respectively the streamflow regime and model roughness coefficient that each particle is included in the particle assembly, and sharp simultaneously
Hydrodynamic model is successively driven with the boundary condition that becomes a mandarin out, realizes mimic water-depth and the simulation of the hydrodynamic model output
Update of the flow from current time to subsequent time；
Judge in the mimic water-depth under current time with the presence or absence of water level observation；The water level observation if it does not exist,
Then directly execute described the step of realizing the hydrodynamic model output mimic water-depth and analogue flow rate；
The water level observation if it exists then calculates the likelihood function value of each particle under the subsequent time, and
The weight of the particle is updated using the likelihood function value；
Calculate the optimal estimation of the mimic water-depth, the analogue flow rate and default roughness coefficien；
Multinomial resampling is carried out to the particle, obtains new particle set；
The new particle set is replaced the particle assembly, while using the subsequent time as described current
Moment executes the step of hydrodynamic model described in the sight exports mimic water-depth and analogue flow rate.It can from above-mentioned technical proposal
To find out, the present invention provides a kind of particle filter assimilation method of hydrodynamic model, device and equipment is calculated, is specially acquired
The calculating parameter in area is studied, and sets the size of the grid in research area and goes out the boundary condition that becomes a mandarin；According to owning for current time
Water level, flow and the default roughness coefficien of grid generate the particle of multiple equal weights, obtain particle assembly；It is utilized respectively particle collection
The streamflow regime and model roughness coefficient that each particle is included in conjunction, and water is successively driven using the boundary condition that becomes a mandarin out simultaneously
Dynamic model realizes the update of the mimic water-depth and analogue flow rate of hydrodynamic model output from current time to subsequent time；Sentence
It whether there is water level observation in mimic water-depth under disconnected current time；The water level observation if it does not exist then directly executes
It is described to realize the hydrodynamic model output mimic water-depth and analogue flow rate；Water level observation if it exists then calculates subsequent time
Under each particle likelihood function value, and using likelihood function value more new particle weight；Calculating simulation water level, analogue flow rate and
The optimal estimation of default roughness coefficien；Multinomial resampling is carried out to particle, obtains new particle set；By new particle set to grain
Subclass is replaced, while using subsequent time as current time, being executed and being realized hydrodynamic model output mimic water-depth and mould
The step of quasi- flow.Due to considering the Spatial-Temporal Variability of roughness coefficien in this programme, i.e., not only consider its time frame coefficient,
Also consider its Spatial Variability simultaneously, therefore the accuracy and reliability of hydrodynamic model output result can be made higher, it can
The accurate different spaces that obtain do well the optimal estimation of variable and roughness coefficien, to solve hydrodynamic model output result
The poor problem of accuracy and reliability.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
It obtains other drawings based on these drawings.
Fig. 1 is a kind of step flow chart of the particle filter assimilation method of hydrodynamic model provided in an embodiment of the present invention；
Fig. 2 is the landform and observation point position view of the river Toce provided in an embodiment of the present invention physical model；
Fig. 3 is that the physical model flow that becomes a mandarin in the river Toce provided in an embodiment of the present invention changes over time figure；
Fig. 4 is that water level assimilates root-mean-square error at 10 observation points of the river Toce provided in an embodiment of the present invention physical model
Figure；
Fig. 5 be the river Toce physical model P1, S6D, P8 and P21 observation point provided in an embodiment of the present invention at assimilate water level and
Mimic water-depth comparison diagram；
Fig. 6 is Manning roughness coefficien assimilation at 10 observation points of the river Toce provided in an embodiment of the present invention physical model
Result figure；
Fig. 7 is that a kind of particle filter of hydrodynamic model provided in an embodiment of the present invention assimilates the structural block diagram of device.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
Embodiment one
Fig. 1 is a kind of step flow chart of the particle filter assimilation method of hydrodynamic model provided in an embodiment of the present invention.
Shown in referring to Fig.1, particle filter assimilation method provided in this embodiment has for optimizing to hydrodynamic model
Body includes the following steps：
The calculating parameter in area is studied in S1, acquisition, and is set the size of the grid in research area and gone out the boundary condition that becomes a mandarin.
Here calculating parameter include study the Bottom Altitude in area, boundary becomes a mandarin the number such as flow, initial water level, initial flow
According to.
By taking Toce river water dynamic model as an example.Toce river water dynamic model is by 1:100 scale bar is to Milan, ITA
The hydrodynamic model established within the scope of the upstream about 5km of the city river Toce, is usually used in examining various dam break mathematical modulos as master pattern
The precision and stability of type.The hydrodynamic model is about 50m, wide about 11m.The spatial resolution of model elevation is 5cm, more smart
True describes the real terrain in the river Toce.
The landform of Toce river water dynamic model and observation point position are as shown in Figure 2.There are an empty reservoir, reservoir in korneforos portion
There is a opening close to river side, but gate is in close state.Due to embankment barrier effect, flood can be made when dam break starts
Water level harmony is high, and as water level rises, unrestrained dike is entered reservoir by flood.Toce river water dynamic model utilizes the prominent of top water level in water pool
So increase to simulate dam break and become a mandarin process, the boundary flow that becomes a mandarin changes over time as shown in Figure 3.The initial depth of water of model is 0m, mould
The outlet of type sets a free discharge boundary.Modeling sizing grid is 0.1m × 0.1m, and roughness takes ENEL to recommend
0.0162s/m1/3。
S2, multiple equal weights are generated according to water level, flow and the default roughness coefficien of all grids at current time
Particle, obtain particle assembly.
Or according to the priori of water level z and roughness coefficien n at all grids of Toce river water dynamic model initial time point
Cloth generates the particle of N number of equal weight at random respectively, to obtain above-mentioned particle assembly.
In formula：WithRespectively indicate mimic water-depth of i-th of the particle of t moment at j-th of grid, flow and
Roughness coefficien, i are particle number, and j is grid cell number；Ncell indicates total calculating grid number.
Influence of the primary condition to assimilation effect is smaller, and Toce river water dynamic model initial time water depth flow speed is 0,
Therefore when initialization particle, all particle water levels are disposed as 0, only initialize to Manning roughness coefficien.Particle is set
Number N=100, from Manning roughness coefficien section [0.062,0.0262] s/m1/3According to being uniformly distributed at random in all grids
Place generates 100 Manning roughness coefficiens respectively.
S3, streamflow regime representated by each particle and model roughness coefficient are utilized respectively as primary condition, using out
The boundary condition that becomes a mandarin successively drives hydrodynamic model, realizes the mimic water-depth of hydrodynamic model output from current time to lower a period of time
The update at quarter.
Hydrodynamic model is transported using water flow is described based on average two-dimensional shallow water governing equation (9)~(11) of the depth of water
Dynamic, the model is discrete to equation progress based on structured grid, and it is logical to calculate grid interface using the Approximate Riemann Solution of HLLC format
Amount, is constantly integrated forward using MUSCL-Hancock method, and model is made to have second order accuracy on space-time；Source item is carried out discrete
Processing ensures the stability of model；Model introduces effective wet-dry boundaries and the irregular terrain profiles boundary based on B-spline method simultaneously
Processing method, the dynamic of accurate simulation dry and wet unit alternately and the flow characteristics on complex boundary, realize all network analogs
The recursion at water level moment from t to t+1.
In formula, ζ is table relative altitude；H (=ζ+hs) it is total depth of water；hsFor the benchmark depth of water；U and v represent x, the side y
To flow velocity；G is acceleration of gravity；ρ is water density；τbxAnd τbyFor x, the bed surface friction stree in the direction y.
S4, judge whether there is water level observation in mimic water-depth under current time.
Here water level observation refers to the water level value that actual observation arrives in model actual motion.If it is not, then directly holding
Row is described to realize the step of hydrodynamic model exports mimic water-depth and analogue flow rate.
S5, if so, then calculate the likelihood function value of each particle under subsequent time, and the weight of more new particle.
Even there are the water level observations in mimic water-depth, then calculate (current time at t+1 moment according to formula (12)~(13)
Subsequent time) each particle likelihood function valueAnd particle weights are updated using the likelihood function value
In formula,WithRespectively water-level simulation value and j-th net of i-th of the particle of t+1 moment at j-th of grid
The observation of water level at lattice,For the standard deviation of observed stage at j-th of grid.
The water level observation at 10 observation points (Fig. 2) is selected, grid particle where 10 observation points at t+1 moment is calculated
Likelihood function value, update particle weights.Due to becoming a mandarin essentially 0 before 20s, assimilate since 20s.Observation error
It is a key parameter for influencing particle filter effect,It is arranged excessive, it is insensitive to observation to will lead to particle weights, drop
Low observation to the assimilation effect of the analogue value,It is arranged too small, it is too sensitive to observation to will lead to particle weights, makes close
The particle of observed stage obtains the weight close to 1, after resampling particle can serious dilution, influence filter effect.Comprehensively consider,
Setting
S6, the mimic water-depth for calculating hydrodynamic model, analogue flow rate and the optimal estimation for predicting roughness coefficien.
Specifically calculation formula is：
S7, multinomial resampling is carried out to all particles, obtains new particle set.
Corresponding water-level observation observation is not present i.e. in mimic water-depth, then multinomial resampling is carried out to particle, obtained
New particle set with equal weight.
One major defect of particle filter is the increase with the number of iterations, the very little that the weight of many particles becomes,
Only a small number of particles obtain biggish weight, sample degeneracy phenomenon occur, and a large amount of computing resources are close for updating these weights
In 0 particle.The main path for solving sample degeneracy at present is to carry out resampling to particle.The resampling side being widely used at present
Formula mainly includes：Multinomial resampling, layering resampling, system resampling and residual error resampling, multinomial resampling is other
The basis of three kinds of method for resampling can solve the problems, such as sample degeneracy during resampling substantially.Its concrete measure is as follows：
Firstly, N number of random number u is generated in being uniformly distributed from [0,1] at randomi。
Then, if wherein some random number uiMeet condition (14), then replicating m-th of particle is new sample point, to new
Sample point is sampled, and the new particle set is obtained.
Although resampling can solve sample degeneracy problem, but replicate the biggish particle of weight and will lead to a large amount of particles complete one
It causes, causes " particle dilution ", for " particle dilution " problem, most straightforward approach is parameter increase white noise, but with iteration
The variance of the increase of number, parameter prior distribution can constantly increase, and core smoothing method (formula (15)) can to avoid this problem,
Therefore the recursion at Fe coatings moment from t to t+1 is realized using core smoothing method.
In formula,For the Manning roughness coefficien of i-th of particle at jth grid,For particle N number of at j-th of grid institute
The mean value of the Manning roughness coefficien of expression, h are core smoothing parameter, VtVariance is disturbed for roughness coefficien.Core smoothing parameter h takes
Value 0.2, VtValue 3 × 10-4。
S8, using new particle set as original particle assembly, while using subsequent time as current time, execute hydrodynamic(al)
The operation of power model output mimic water-depth and analogue flow rate.
Even t=t+1, return step S3 carries out loop iteration, until the operation of all moment is completed.Calculate separately 10 sights
Assimilate water level and mimic water-depth root-mean-square error (RMSE, formula (16) are schemed (4)) at measuring point, assimilates water level and mimic water-depth comparison
As shown in Figure 5 (by taking P1, S6D, P8 and P21 as an example), it can be seen that water level and observed stage after particle filter assimilates are more
Close, root-mean-square error significantly reduces at each observation point.Since the hydrodynamic model does not have flow measured value at each observation point,
Therefore flow assimilation result is not verified.In addition, Manning roughness coefficien updates result as shown in fig. 6, can be with by Fig. 6
Find out, there are significant Spatial-Temporal Variabilities for Manning roughness coefficien.It should be pointed out that the Manning roughness coefficien of estimation is simultaneously
Actual degree of roughness is not indicated.Because only considered parameter uncertainty in the assimilation system, Manning roughness coefficien is estimated
Meter result contains that landform, boundary condition, grid be discrete and the brings such as model structure are uncertain.Such as observation point
Manning roughness coefficien estimated value at S6D is gradually increased, and Manning roughness coefficien maximum value is greater than 0.1s/m1/3, significant big
In the 0.0162s/m of recommendation1/3.Other uncertain sources can be balanced by parameter to miss to caused by modeling water level
Difference.The assimilation method considers the Spatial-Temporal Variability of roughness coefficien, passes through the roughness coefficien of adjust automatically different location, Ke Yizhun
The really mimic water-depth at estimation different location.
In formula, RMSEkFor the assimilation or mimic water-depth root-mean-square error at k-th of observation point, T is total simulated time, T=
180,For t moment water level observation at k-th of observation point,T moment water level assimilation number or simulation at k-th of observation point
Value.
It can be seen from the above technical proposal that present embodiments providing a kind of particle filter assimilation side of hydrodynamic model
Method, the specially calculating parameter in acquisition research area, and set the size of the grid in research area and go out the boundary condition that becomes a mandarin；According to working as
Water level, flow and the default roughness coefficien of all grids at preceding moment generate the particle of multiple equal weights, obtain particle assembly；Point
Not Li Yong each particle is included in particle assembly streamflow regime and model roughness coefficient, and utilize the perimeter strip that becomes a mandarin simultaneously
Part successively drives hydrodynamic model, and the mimic water-depth and analogue flow rate for realizing hydrodynamic model output are from current time to lower a period of time
The update at quarter；Judge in the mimic water-depth under current time with the presence or absence of water level observation；Water level observation if it exists then calculates
The likelihood function value of each particle under subsequent time, and utilize the weight of likelihood function value more new particle；Calculating simulation water level, mould
The optimal estimation of quasi- flow and default roughness coefficien；Water level observation if it does not exist then carries out multinomial resampling to particle, obtains
To new particle set；New particle set is replaced particle assembly, while using subsequent time as current time, executing view
Line hydrodynamic model exports the step of mimic water-depth and analogue flow rate.Due to considering the when space-variant of roughness coefficien in this programme
The opposite sex not only considers its time frame coefficient, also while considering its Spatial Variability, therefore can make hydrodynamic model output knot
The accuracy and reliability of fruit is higher, can accurately obtain different spaces and do well the optimal estimation of variable and roughness coefficien, from
And solve the problems, such as that the accuracy and reliability of hydrodynamic model output result is poor.
It should be noted that for simple description, therefore, it is stated as a series of action groups for embodiment of the method
It closes, but those skilled in the art should understand that, embodiment of that present invention are not limited by the describe sequence of actions, because according to
According to the embodiment of the present invention, some steps may be performed in other sequences or simultaneously.Secondly, those skilled in the art also should
Know, the embodiments described in the specification are all preferred embodiments, and the related movement not necessarily present invention is implemented
Necessary to example.
Embodiment two
Fig. 7 is that a kind of particle filter of hydrodynamic model provided in an embodiment of the present invention assimilates the structural block diagram of device.
Referring to shown in Fig. 7, particle filter assimilation device provided in this embodiment has for optimizing to hydrodynamic model
Body includes parameter collection module 10, the first particle sampler module 20, model output module 30, the meter of condition judgment module 40, first
Calculate module 50, the second computing module 60, the second particle sampler module 70 and particle collection replacement module 80.
Parameter collection module is used to acquire the calculating parameter in research area, and sets the size of the grid in research area and go out to become a mandarin
Boundary condition.
Here calculating parameter include study the Bottom Altitude in area, boundary becomes a mandarin the number such as flow, initial water level, initial flow
According to.
By taking Toce hydrodynamic model as an example.Toce river water dynamic model is by 1:100 scale bar is to Milan, ITA city
The hydrodynamic model established within the scope of the about 5km of the river Toce upstream, is usually used in examining various dam break mathematical models as master pattern
Precision and stability.The hydrodynamic model is about 50m, wide about 11m.The spatial resolution of model elevation is 5cm, more accurate
Describe the real terrain in the river Toce.
The landform of Toce river water dynamic model and observation point position are as shown in Figure 2.There are an empty reservoir, reservoir in korneforos portion
There is a opening close to river side, can make flood level harmony high due to embankment barrier effect, when dam break starts, on water level
It rises, unrestrained dike is entered reservoir by flood.Toce hydrodynamic model is become a mandarin using the unexpected increase of top water level in water pool to simulate dam break
Process, the boundary flow that becomes a mandarin change over time as shown in Figure 3.The initial depth of water of model is 0m, and the outlet of model sets a freedom
Outlet boundary.Modeling sizing grid is 0.1m × 0.1m, the 0.0162s/m that roughness takes ENEL to recommend1/3。
First particle sampler module is used for water level, flow and the default roughness system of all grids according to current time
Number generates the particle of multiple equal weights, obtains particle assembly.
Or according to the priori of water level z and roughness coefficien n at all grids of Toce river water dynamic model initial time point
Implantation generates the particle of N number of equal weight at random respectively, to obtain above-mentioned particle assembly.
In formula：WithRespectively indicate mimic water-depth of i-th of the particle of t moment at j-th of grid, flow and
Roughness coefficien, i are particle number, and j is grid cell number；Ncell indicates total calculating grid number.
Influence of the primary condition to assimilation effect is smaller, and Toce river water dynamic model initial time water depth flow speed is 0,
Therefore when initialization particle, all particle water levels are disposed as 0, only initialize to Manning roughness coefficien.Particle is set
Number N=100, from Manning roughness table coefficient section [0.062,0.0262] s/m1/3According to being uniformly distributed at random in institute in section
Have and generates 100 Manning roughness table coefficients at grid respectively.
Model output module is for being utilized respectively streamflow regime representated by each particle and model roughness coefficient as just
Beginning condition successively drives hydrodynamic model using the boundary condition that becomes a mandarin is gone out, and realizes the mimic water-depth of hydrodynamic model output from working as
Update of the preceding moment to subsequent time.
Hydrodynamic model is transported using water flow is described based on average two-dimensional shallow water governing equation (9)~(11) of the depth of water
Dynamic, the model is discrete to equation progress based on structured grid, and it is logical to calculate grid interface using the Approximate Riemann Solution of HLLC format
Amount, is constantly integrated forward using MUSCL-Hancock method, and model is made to have second order accuracy on space-time；Source item is carried out discrete
Processing ensures the stability of model；Model introduces effective wet-dry boundaries and the irregular terrain profiles boundary based on B-spline method simultaneously
Processing method, the dynamic of accurate simulation dry and wet unit alternately and the flow characteristics on complex boundary, realize all network analogs
The recursion at water level moment from t to t+1.
In formula, ζ is table relative altitude；H (=ζ+hs) it is total depth of water；hsFor the benchmark depth of water；U and v represent x, the side y
To flow velocity；G is acceleration of gravity；ρ is water density；τbxAnd τbyFor x, the bed surface friction stree in the direction y.
Condition judgment module is used to judge whether have water level observation in the mimic water-depth under current time.
Here water level observation refers to the water level value that actual observation arrives in model actual motion.If it is not, then directly holding
Row is described to realize the hydrodynamic model output mimic water-depth and analogue flow rate.
If the first computing module is used to have water level observation, the likelihood function value of each particle under subsequent time is calculated,
And the weight of more new particle.
Even there are the water level observations in mimic water-depth, then calculate (current time at t+1 moment according to formula (12)~(13)
Subsequent time) each particle likelihood function valueAnd particle weights are updated using the likelihood function value
In formula,WithRespectively water-level simulation value and j-th net of i-th of the particle of t+1 moment at j-th of grid
The observation of water level at lattice,For the standard deviation of observed stage at j-th of grid.
The water level observation at 10 observation points (Fig. 2) is selected, grid particle where 10 observation points at t+1 moment is calculated
Likelihood function value, update particle weights.Due to becoming a mandarin essentially 0 before 20s, assimilate since 20s.Observation error
It is a key parameter for influencing particle filter effect,It is arranged excessive, it is insensitive to observation to will lead to particle weights, drop
Low observation to the assimilation effect of the analogue value,It is arranged too small, it is too sensitive to observation to will lead to particle weights, makes close
The particle of observed stage obtains the weight close to 1, after resampling particle can serious dilution, influence filter effect.Comprehensively consider,
Setting
Second computing module is used to calculate the optimal of the mimic water-depth of hydrodynamic model, analogue flow rate and prediction roughness coefficien
Valuation.
Specifically calculation formula is：
Second particle sampler module is used to carry out multinomial resampling to all particles, obtains new particle set.
Multinomial resampling is carried out to the particle of update, obtains the new particle set with equal weight.
One major defect of particle filter is the increase with the number of iterations, the very little that the weight of many particles becomes,
Only a small number of particles obtain biggish weight, sample degeneracy phenomenon occur, and a large amount of computing resources are close for updating these weights
In 0 particle.The main path for solving sample degeneracy at present is to carry out resampling to particle.The resampling side being widely used at present
Formula mainly includes：Multinomial resampling, layering resampling, system resampling and residual error resampling, multinomial resampling is other
The basis of three kinds of method for resampling can solve the problems, such as sample degeneracy during resampling substantially.The second particle sampler module
Specifically include random number generation unit and sampling execution unit.
Random number generation unit for generating N number of random number u in being uniformly distributed from [0,1] at randomi。
Execution unit is sampled then if wherein some random number uiMeet condition (14), then replicating m-th of particle is new sample
This point samples new sample point, obtains the new particle set.
Although resampling can solve sample degeneracy problem, but replicate the biggish particle of weight and will lead to a large amount of particles complete one
Cause, cause " particle dilution ", for " particle dilution " problem, most straightforward approach be to parameter increase white noise, but with repeatedly
The variance of the increase of generation number, parameter prior distribution can constantly increase, and core smoothing method (formula (15)) can ask to avoid this
Topic, therefore using the recursion at core smoothing method realization Fe coatings moment from t to t+1.
In formula,For the Manning roughness coefficien of i-th of particle at jth grid,For particle N number of at j-th of grid institute
The mean value of the Manning roughness coefficien of expression, h are core smoothing parameter, VtVariance is disturbed for roughness coefficien.Core smoothing parameter h takes
Value 0.2, VtValue 3 × 10-4。
Particle collection replacement module is used for using new particle set as original particle assembly, while using subsequent time as working as
The preceding moment executes the operation of hydrodynamic model output mimic water-depth and analogue flow rate.
Even t=t+1, return step S3 carries out loop iteration, until the operation of all moment is completed.Calculate separately 10 sights
Assimilate water level and mimic water-depth root-mean-square error (RMSE, formula (16) are schemed (4)) at measuring point, assimilates water level and mimic water-depth comparison
As shown in Figure 5 (by taking P1, S6D, P8 and P21 as an example), it can be seen that water level and observed stage after particle filter assimilates are more
Close, root-mean-square error significantly reduces at each observation point.Since the hydrodynamic model does not have flow measured value at each observation point,
Therefore flow assimilation result is not verified.In addition, Manning roughness coefficien updates result as shown in fig. 6, can be with by Fig. 6
Find out, there are significant Spatial-Temporal Variabilities for Manning roughness coefficien.It should be pointed out that the Manning roughness coefficien of estimation is simultaneously
Actual degree of roughness is not indicated.Because only considered parameter uncertainty in the assimilation system, Manning roughness coefficien is estimated
Meter result contains that landform, boundary condition, grid be discrete and the brings such as model structure are uncertain.Such as observation point
Manning roughness coefficien estimated value at S6D is gradually increased, and Manning roughness coefficien maximum value is greater than 0.1s/m1/3, significant big
In the 0.0162s/m of recommendation1/3.Other uncertain sources can be balanced by parameter to miss to caused by modeling water level
Difference.The assimilation method considers the Spatial-Temporal Variability of roughness coefficien, passes through the roughness coefficien of adjust automatically different location, Ke Yizhun
The really mimic water-depth at estimation different location.
In formula, RMSEkFor the assimilation or mimic water-depth root-mean-square error at k-th of observation point, T is total simulated time, T=
180,For t moment water level observation at k-th of observation point,T moment water level assimilation number or the analogue value at k-th of observation point.
It can be seen from the above technical proposal that present embodiments providing a kind of particle filter of hydrodynamic model with makeup
It sets, specially the calculating parameter in acquisition research area, and sets the size of the grid in research area and go out the boundary condition that becomes a mandarin；According to working as
Water level, flow and the default roughness coefficien of all grids at preceding moment generate the particle of multiple equal weights, obtain particle assembly；Point
Not Li Yong each particle is included in particle assembly streamflow regime and model roughness coefficient, and utilize the perimeter strip that becomes a mandarin simultaneously
Part successively drives hydrodynamic model, and the mimic water-depth and analogue flow rate for realizing hydrodynamic model output are from current time to lower a period of time
The update at quarter；Judge in the mimic water-depth under current time with the presence or absence of water level observation；Water level observation if it does not exist, then directly
Connect the step of hydrodynamic model output mimic water-depth and analogue flow rate are realized in execution；Water level observation if it exists then calculates next
When inscribe the likelihood function value of each particle, and utilize the weight of likelihood function value more new particle；Calculating simulation water level, analog stream
The optimal estimation of amount and default roughness coefficien；Multinomial resampling is carried out to particle, obtains new particle set；By new particle set
Particle assembly is replaced, while using subsequent time as current time, executing and realizing that hydrodynamic model exports mimic water-depth
And the step of analogue flow rate.Due to considering the Spatial-Temporal Variability of roughness coefficien in this programme, i.e., not only consider that its time becomes
The opposite sex also while considering its Spatial Variability, therefore the accuracy and reliability of hydrodynamic model output result can be made higher,
Different spaces can accurately be obtained to do well the optimal estimation of variable and roughness coefficien, to solve hydrodynamic model output knot
The poor problem of the accuracy and reliability of fruit.
Embodiment three
A kind of calculating equipment is present embodiments provided, which is provided with the particle filter provided in an embodiment
Assimilate device, which is used to acquire the calculating parameter in research area, and sets the size of the grid in research area and go out the boundary that becomes a mandarin
Condition；The particle that multiple equal weights are generated according to the water level, flow and default roughness coefficien of all grids at current time, obtains
Particle assembly；It is utilized respectively the streamflow regime and model roughness coefficient that each particle is included in particle assembly, and is utilized simultaneously
The boundary condition that becomes a mandarin out successively drives hydrodynamic model, and the mimic water-depth and analogue flow rate for realizing hydrodynamic model output are from current
Update of the moment to subsequent time；Judge in the mimic water-depth under current time with the presence or absence of water level observation；Water if it does not exist
Position observation then directly executes the step of realizing hydrodynamic model output mimic water-depth and analogue flow rate；Water-level observation if it exists
Value then calculates the likelihood function value of each particle under subsequent time, and utilizes the weight of likelihood function value more new particle；Calculate mould
The optimal estimation of quasi- water level, analogue flow rate and default roughness coefficien；Multinomial resampling is carried out to particle, obtains new particle collection
It closes；New particle set is replaced particle assembly, while using subsequent time as current time, executing and realizing hydrodynamic force mould
Type exports the step of mimic water-depth and analogue flow rate.Since the particle filter assimilation device of the calculating equipment in this programme is considered
The Spatial-Temporal Variability of roughness coefficien, i.e., not only consider its time frame coefficient, also while considering its Spatial Variability, therefore can
Keep the accuracy and reliability of hydrodynamic model output result higher, can accurately obtain different spaces and do well variable and roughness
The optimal estimation of coefficient, to solve the problems, such as that the accuracy and reliability of hydrodynamic model output result is poor.
Example IV
A kind of calculating equipment is present embodiments provided, which includes at least one processor and memory, processing
Device is connected with memory by data/address bus.
For storing computer program or instruction, processor is then used to execute the computer program or instruction memory, from
And the calculating equipment is made to realize following operation：
The calculating parameter in acquisition research area, and set the size of the grid in research area and go out the boundary condition that becomes a mandarin；
The particle of multiple equal weights is generated according to the water level, flow and default roughness coefficien of all grids at current time,
Obtain particle assembly；
It is utilized respectively the streamflow regime and model roughness coefficient that each particle is included in particle assembly, and simultaneously using going out
The boundary condition that becomes a mandarin successively drives hydrodynamic model, realize hydrodynamic model output mimic water-depth and analogue flow rate from it is current when
It is carved into the update of subsequent time；
Judge in the mimic water-depth under current time with the presence or absence of water level observation；
Water level observation if it does not exist then directly executes the step for realizing hydrodynamic model output mimic water-depth and analogue flow rate
Suddenly；
Water level observation if it exists then calculates the likelihood function value of each particle under subsequent time, and utilizes likelihood function
It is worth the weight of more new particle；
The optimal estimation of calculating simulation water level, analogue flow rate and default roughness coefficien；
Multinomial resampling is carried out to particle, obtains new particle set；
New particle set is replaced particle assembly, while using subsequent time as current time, executing and realizing water
Dynamic model exports the step of mimic water-depth and analogue flow rate.
It has been specifically contemplated that the Spatial-Temporal Variability of roughness coefficien in above scheme, i.e., has not only considered its time frame coefficient, also
Consider its Spatial Variability simultaneously, therefore the accuracy and reliability of hydrodynamic model output result can be made higher, Neng Gouzhun
It really obtains different spaces to do well the optimal estimation of variable and roughness coefficien, to solve the standard of hydrodynamic model output result
True property and the poor problem of reliability.
For device embodiment, since it is basically similar to the method embodiment, related so being described relatively simple
Place illustrates referring to the part of embodiment of the method.
All the embodiments in this specification are described in a progressive manner, the highlights of each of the examples are with
The difference of other embodiments, the same or similar parts between the embodiments can be referred to each other.
It should be understood by those skilled in the art that, the embodiment of the embodiment of the present invention can provide as method, apparatus or calculate
Machine program product.Therefore, the embodiment of the present invention can be used complete hardware embodiment, complete software embodiment or combine software and
The form of the embodiment of hardware aspect.Moreover, the embodiment of the present invention can be used one or more wherein include computer can
With in the computer-usable storage medium (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.) of program code
The form of the computer program product of implementation.
The embodiment of the present invention be referring to according to the method for the embodiment of the present invention, terminal device (system) and computer program
The flowchart and/or the block diagram of product describes.It should be understood that flowchart and/or the block diagram can be realized by computer program instructions
In each flow and/or block and flowchart and/or the block diagram in process and/or box combination.It can provide these
Computer program instructions are set to general purpose computer, special purpose computer, Embedded Processor or other programmable data processing terminals
Standby processor is to generate a machine, so that being held by the processor of computer or other programmable data processing terminal devices
Capable instruction generates for realizing in one or more flows of the flowchart and/or one or more blocks of the block diagram
The device of specified function.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing terminal devices
In computer-readable memory operate in a specific manner, so that instruction stored in the computer readable memory generates packet
The manufacture of command device is included, which realizes in one side of one or more flows of the flowchart and/or block diagram
The function of being specified in frame or multiple boxes.
These computer program instructions can also be loaded into computer or other programmable data processing terminal devices, so that
Series of operation steps are executed on computer or other programmable terminal equipments to generate computer implemented processing, thus
The instruction executed on computer or other programmable terminal equipments is provided for realizing in one or more flows of the flowchart
And/or in one or more blocks of the block diagram specify function the step of.
Although the preferred embodiment of the embodiment of the present invention has been described, once a person skilled in the art knows bases
This creative concept, then additional changes and modifications can be made to these embodiments.So the following claims are intended to be interpreted as
Including preferred embodiment and fall into all change and modification of range of embodiment of the invention.
Finally, it is to be noted that, herein, relational terms such as first and second and the like be used merely to by
One entity or operation are distinguished with another entity or operation, without necessarily requiring or implying these entities or operation
Between there are any actual relationship or orders.Moreover, the terms "include", "comprise" or its any other variant meaning
Covering non-exclusive inclusion, so that process, method, article or terminal device including a series of elements not only wrap
Those elements are included, but also including other elements that are not explicitly listed, or further includes for this process, method, article
Or the element that terminal device is intrinsic.In the absence of more restrictions, being wanted by what sentence "including a ..." limited
Element, it is not excluded that there is also other identical elements in process, method, article or the terminal device for including the element.
Technical solution provided by the present invention is described in detail above, specific case used herein is to this hair
Bright principle and embodiment is expounded, method of the invention that the above embodiments are only used to help understand and its
Core concept；At the same time, for those skilled in the art, according to the thought of the present invention, in specific embodiment and application
There will be changes in range, in conclusion the contents of this specification are not to be construed as limiting the invention.
Claims (10)
1. a kind of particle filter assimilation method of hydrodynamic model, which is characterized in that including step：
The calculating parameter in acquisition research area, and set the size of the grid in the research area and go out the boundary condition that becomes a mandarin；
The particle of multiple equal weights is generated according to water level, flow and the default roughness coefficien of all grids at current time,
Obtain particle assembly；
It is utilized respectively the streamflow regime and model roughness coefficient that each particle is included in the particle assembly, and utilizes institute simultaneously
It states out the boundary condition that becomes a mandarin and successively drives hydrodynamic model, realize the mimic water-depth and analogue flow rate of the hydrodynamic model output
Update from current time to subsequent time；
Judge to whether there is water level observation in the mimic water-depth under current time, if it does not exist the water level observation, then directly
It connects and executes described the step of realizing the hydrodynamic model output mimic water-depth and analogue flow rate；
The water level observation if it exists, then calculate the likelihood function value of each particle under the subsequent time, and utilizes
The likelihood function value updates the weight of the particle；
Calculate the optimal estimation of the mimic water-depth, the analogue flow rate and roughness coefficien；
Multinomial resampling is carried out to the particle, obtains new particle set；
The new particle set is replaced the particle assembly, at the same using the subsequent time as it is described current when
It carves, executes described the step of realizing the hydrodynamic model output mimic water-depth and analogue flow rate.
2. particle filter assimilation method as described in claim 1, which is characterized in that the calculating parameter includes the research area
Bottom Altitude, boundary becomes a mandarin some or all of flow, initial water level and initial flow.
3. particle filter assimilation method as described in claim 1, which is characterized in that the default roughness coefficien is roughness coefficien
Priori value.
4. particle filter assimilation method as described in claim 1, which is characterized in that described to carry out multinomial weight to the particle
Sampling includes：
Multiple random numbers are generated in being uniformly distributed from [0,1] at random；
It samples, obtains using particle corresponding to the random number for meeting preset condition in the multiple random number as new sample point
To the new particle set.
5. a kind of particle filter of hydrodynamic model assimilates device, which is characterized in that including：
Parameter collection module for acquiring the calculating parameter in research area, and sets the size and discrepancy of the grid in the research area
Flow boundary condition；
First particle sampler module, for water level, flow and the default roughness coefficien according to all grids at current time
The particle for generating multiple equal weights, obtains particle assembly；
Model output module, the streamflow regime and model roughness for being included for being utilized respectively each particle in the particle assembly
Coefficient, and hydrodynamic model is successively driven using the boundary condition that becomes a mandarin out simultaneously, realize the hydrodynamic model output
The update of mimic water-depth and analogue flow rate from current time to subsequent time；
Condition judgment module, for judging in the mimic water-depth under current time with the presence or absence of water level observation；
First computing module then calculates each particle under the subsequent time for the water level observation if it exists
Likelihood function value, and update using the likelihood function value weight of the particle；
Second computing module, for calculating the optimal estimation of the mimic water-depth, the analogue flow rate and default roughness coefficien；
Second particle sampler module obtains new particle set for carrying out multinomial resampling to the particle；
Particle collection replacement module, for the new particle set to be replaced the particle assembly, while will be described next
Moment as the current time, executes described the step of realizing the hydrodynamic model output mimic water-depth and analogue flow rate.
6. particle filter as claimed in claim 5 assimilates device, which is characterized in that the calculating parameter includes the research area
Bottom Altitude, boundary becomes a mandarin some or all of flow, initial water level and initial flow.
7. particle filter as claimed in claim 5 assimilates device, which is characterized in that the default roughness coefficien is roughness coefficien
Priori value.
8. particle filter as claimed in claim 5 assimilates device, which is characterized in that the second particle sampler module includes：
Random number generation unit, for generating multiple random numbers at random in being uniformly distributed from [0,1]；
Execution unit is sampled, particle corresponding to the random number for will meet preset condition in the multiple random number is as new
Sample point is sampled, and the new particle set is obtained.
9. a kind of calculating equipment, which is characterized in that setting is just like the described in any item particle filters of claim 5~8 with makeup
It sets.
10. a kind of calculating equipment, which is characterized in that including at least one processor, memory and be used at least one described in general
The data/address bus that processor is connect with the memory signals, the memory are handled for storing computer program or instruction
Device is for executing the computer program or instruction, to realize following steps：
The calculating parameter in acquisition research area, and set the size of the grid in the research area and go out the boundary condition that becomes a mandarin；
The particle of multiple equal weights is generated according to water level, flow and the default roughness coefficien of all grids at current time,
Obtain particle assembly；
It is utilized respectively the streamflow regime and model roughness coefficient that each particle is included in the particle assembly, and utilizes institute simultaneously
It states out the boundary condition that becomes a mandarin and successively drives hydrodynamic model, realize the mimic water-depth and analogue flow rate of the hydrodynamic model output
Update from current time to subsequent time；
Judge in the mimic water-depth under current time with the presence or absence of water level observation；
The water level observation if it does not exist then directly executes the output mimic water-depth of hydrodynamic model described in the sight and simulation
The step of flow；
The water level observation if it exists, then calculate the likelihood function value of each particle under the subsequent time, and utilizes
The likelihood function value updates the weight of the particle；
Calculate the optimal estimation of the mimic water-depth, the analogue flow rate and default roughness coefficien；
Multinomial resampling is carried out to the particle, obtains new particle set；
The new particle set is replaced the particle assembly, at the same using the subsequent time as it is described current when
The step of carving, executing the output mimic water-depth of hydrodynamic model described in the sight and analogue flow rate.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810372803.8A CN108920737B (en) | 2018-04-24 | 2018-04-24 | Particle filtering assimilation method and device of hydrodynamic model and computing equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810372803.8A CN108920737B (en) | 2018-04-24 | 2018-04-24 | Particle filtering assimilation method and device of hydrodynamic model and computing equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108920737A true CN108920737A (en) | 2018-11-30 |
| CN108920737B CN108920737B (en) | 2021-04-27 |
Family
ID=64403843
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810372803.8A Active CN108920737B (en) | 2018-04-24 | 2018-04-24 | Particle filtering assimilation method and device of hydrodynamic model and computing equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108920737B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110119590A (en) * | 2019-05-22 | 2019-08-13 | 中国水利水电科学研究院 | A kind of water quality model particle filter assimilation method based on multi-source observation data |
| CN111859820A (en) * | 2020-06-23 | 2020-10-30 | 河海大学 | Roughness partition calibration method and system based on posterior distribution |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20160117423A1 (en) * | 2013-03-13 | 2016-04-28 | The Government Of The United States Of America, As Represented By The Secretary Of The Navy | System and method for correcting a model-derived vertical structure of ocean temperature and ocean salinity based on sea surface height observations |
| CN107678050A (en) * | 2017-09-05 | 2018-02-09 | 西南交通大学 | GLONASS phase inter-frequency deviation real-time tracings and precise estimation method based on particle filter |
-
2018
- 2018-04-24 CN CN201810372803.8A patent/CN108920737B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20160117423A1 (en) * | 2013-03-13 | 2016-04-28 | The Government Of The United States Of America, As Represented By The Secretary Of The Navy | System and method for correcting a model-derived vertical structure of ocean temperature and ocean salinity based on sea surface height observations |
| CN107678050A (en) * | 2017-09-05 | 2018-02-09 | 西南交通大学 | GLONASS phase inter-frequency deviation real-time tracings and precise estimation method based on particle filter |
Non-Patent Citations (1)
| Title |
|---|
| 徐兴亚: "河道洪水实时概率预报模型与应用", 《水科学进展》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110119590A (en) * | 2019-05-22 | 2019-08-13 | 中国水利水电科学研究院 | A kind of water quality model particle filter assimilation method based on multi-source observation data |
| CN110119590B (en) * | 2019-05-22 | 2020-08-11 | 中国水利水电科学研究院 | Water quality model particle filtering assimilation method based on multi-source observation data |
| CN111859820A (en) * | 2020-06-23 | 2020-10-30 | 河海大学 | Roughness partition calibration method and system based on posterior distribution |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108920737B (en) | 2021-04-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108334660B (en) | Water-sand prediction method and system of strong alluvial river based on data assimilation | |
| Dee | Simplification of the Kalman filter for meteorological data assimilation | |
| CN105302974B (en) | A kind of real-time cutting simulation method of flexible article analyzed based on finite element and time-varying modal | |
| CN108256177A (en) | A kind of parameter optimization method and system of river Water-sand model | |
| Avolio et al. | SCIDDICA-SS 3: a new version of cellular automata model for simulating fast moving landslides | |
| Cañizares | On the application of data assimilation in regional coastal models | |
| CN108563837A (en) | A kind of the model parameter real-time correction method and system of alluvial stream Water-sand model | |
| CN110232471A (en) | A kind of precipitation sensing net node layout optimization method and device | |
| Lafdani et al. | Research article daily rainfall-runoff prediction and simulation using ANN, ANFIS and conceptual hydrological MIKE11/NAM models | |
| CN106651007A (en) | Method and device for GRU-based medium and long-term prediction of irradiance of photovoltaic power station | |
| Mohandes et al. | Short term wind speed estimation in Saudi Arabia | |
| CN108920737A (en) | A kind of particle filter assimilation method of hydrodynamic model, device and calculate equipment | |
| Üneş et al. | Determination of groundwater level fluctuations by artificial neural networks | |
| JP6819692B2 (en) | Simulation equipment, simulation methods and programs | |
| Tayyab et al. | Application of artificial intelligence method coupled with discrete wavelet transform method | |
| CN116502775B (en) | Hydrologic sequence enhancement and prediction method | |
| Khalaj et al. | Investigation about climate change and human activity effects on groundwater level and groundwater quality in semiarid region | |
| CN115952753B (en) | Chip flow simulation method, system and equipment combining CA and LBM | |
| CN115034159A (en) | Power prediction method, device, storage medium and system for offshore wind farm | |
| CN110442836A (en) | Short-term wind speed forecasting method of wind farm and system | |
| Jahromi et al. | Application of the HBMO approach to predict the total sediment discharge | |
| Kuok et al. | Particle swarm optimization for calibrating and optimizing Xinanjiang model parameters | |
| Ali et al. | Statistical downscaling of precipitation and temperature using gene expression programming | |
| Demirci et al. | Monthly groundwater level modeling using data mining approaches | |
| Azami Babani et al. | Prioritization flood risk of hydrologic units (Case Study: Pole Shah catchment) |
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 | ||
| GR01 | Patent grant |