CN115422840A - Ridge-scale runoff estimation method based on physical model mixed deep learning model - Google Patents
Ridge-scale runoff estimation method based on physical model mixed deep learning model Download PDFInfo
- Publication number
- CN115422840A CN115422840A CN202211088365.5A CN202211088365A CN115422840A CN 115422840 A CN115422840 A CN 115422840A CN 202211088365 A CN202211088365 A CN 202211088365A CN 115422840 A CN115422840 A CN 115422840A
- Authority
- CN
- China
- Prior art keywords
- model
- runoff
- data
- physical
- substeps
- 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 42
- 238000013136 deep learning model Methods 0.000 title claims abstract description 27
- 238000013499 data model Methods 0.000 claims abstract description 20
- 238000012549 training Methods 0.000 claims abstract description 13
- 238000012360 testing method Methods 0.000 claims description 12
- 238000011156 evaluation Methods 0.000 claims description 7
- 238000005259 measurement Methods 0.000 claims description 7
- 238000012795 verification Methods 0.000 claims description 5
- 238000010606 normalization Methods 0.000 claims description 4
- 238000007781 pre-processing Methods 0.000 claims description 4
- 238000011160 research Methods 0.000 claims description 4
- 238000004088 simulation Methods 0.000 claims description 4
- 238000005457 optimization Methods 0.000 claims description 3
- 238000001556 precipitation Methods 0.000 claims description 3
- 238000006243 chemical reaction Methods 0.000 claims description 2
- 230000000737 periodic effect Effects 0.000 claims description 2
- 230000008901 benefit Effects 0.000 abstract description 3
- 230000009286 beneficial effect Effects 0.000 abstract description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 8
- 238000013135 deep learning Methods 0.000 description 7
- 230000008569 process Effects 0.000 description 4
- 238000011144 upstream manufacturing Methods 0.000 description 3
- 230000002159 abnormal effect Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 238000005293 physical law Methods 0.000 description 1
- 238000000053 physical method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 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
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2113/00—Details relating to the application field
- G06F2113/08—Fluids
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A10/00—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE at coastal zones; at river basins
- Y02A10/40—Controlling or monitoring, e.g. of flood or hurricane; Forecasting, e.g. risk assessment or mapping
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Geometry (AREA)
- Medical Informatics (AREA)
- Computer Hardware Design (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a method for estimating Japanese scale runoff based on a physical model mixed deep learning model, which comprises the following steps: the method comprises the following steps of (I) estimating preliminary runoff based on a HIMS hydrological model; and (II) based on the mixed physical data model HPD, taking the preliminary runoff data estimated by the hydrological model in the step (I) and other site observation data as a training data set, acquiring an optimal HPD model, and carrying out daily runoff estimation by using the optimal model. The invention has the following beneficial effects: the method combines the advantages of a physical model and a deep learning model, has good learning capacity of simulating the runoff, and can more accurately estimate the runoff; while the estimation of peak runoff is optimized.
Description
Technical Field
The invention relates to a method for simulating regional daily runoff, in particular to a daily runoff estimation method based on a mixed physical model and a deep learning network model.
Background
Runoff is an important link in water circulation movement, determines the water resource condition and the ecological environment quality in a certain area to a certain degree, and is particularly important for timely and accurately acquiring dynamic changes of river runoff on different time scales. Runoff prediction is one of key tasks of effective water resource management, and aims to obtain the runoff volume in a certain runoff area through certain priori knowledge and technical means, make relevant schemes and policies aiming at runoff prediction results, accurately predict runoff, enable people to take corresponding measures as soon as possible on the problems of drought resistance, flood control, water abandonment, water storage and the like, perform overall arrangement and better maximize comprehensive benefits on the premise of safety.
The current runoff prediction method mainly comprises a hydrological model and a data driving method, wherein the hydrological model shows an interpretable relation between input variables and output variables on the basis of a physical model, each parameter of the model has definite physical significance, runoff abnormal values caused by extreme rainfall can be well captured, but the model is complex and depends heavily on expert knowledge, the water circulation process is assumed and simplified in the modeling process, and a plurality of physical parameters are generally concentrated into one parameter to reduce the complexity of the model.
In recent years, deep learning techniques have enjoyed tremendous success in many computer vision and natural language processing applications. These techniques are becoming more popular in geoscience applications including hydrology, and data-driven runoff prediction methods based on deep learning techniques can better capture the nonlinear relationship between input features and runoff, and can achieve higher runoff prediction accuracy (Kratzert, klotz et al 2018, anh, loc et al 2020, gauch, kratzert et al 2021, frame, kratzert et al 2022).
Since deep learning models can only capture the correlation between variables, even though the performance prediction of data-driven models based on deep learning methods is high, it provides little physical explanation except its fitting ability, deep learning methods completely disregard the physical laws behind the data set, and do not see the fundamental laws of physics, and thus lack a generalized model that can be reliably used to simulate a hydrological process.
Physics-based models are well suited to represent conceptually well understood processes. Deep learning models may fit observed data well, but predictions may be physically inconsistent, resulting in large variations even with slight disturbances. Thus, a deep learning model of physical guidance is a possible approach to solving the current problem.
Disclosure of Invention
Aiming at the defects in the prior art, the invention aims to provide a daily-scale runoff estimation method based on a physical model mixed deep learning model, so that the runoff estimation precision is further improved, and particularly, the extreme runoff is estimated.
In order to solve the problems, the invention adopts the following technical scheme:
a method for estimating the daily-scale runoff based on a physical model mixed deep learning model is characterized by comprising the following steps:
carrying out calibration and verification through a distributed hydrological model HIMS by utilizing meteorological data and runoff data in a research area, and carrying out primary runoff estimation based on the calibrated and verified distributed hydrological model HIMS to obtain a primary runoff estimation Q';
step (two) is to input D and output Y of the physical model PHY Taken together as a deep learning model f PHY The input characteristics of the physical data model are used for constructing a mixed physical data model HPD; and (3) based on the constructed mixed physical data model HPD, taking the preliminary runoff data obtained by the hydrological model estimation in the step (I) and other site observation data as a training data set, obtaining an optimal mixed physical data model HPD, and carrying out daily runoff estimation by using the optimal mixed physical data model HPD.
Compared with the prior art, the invention has the following beneficial effects:
the traditional hydrological model has definite physical significance, but more required input parameters and more complex model forms, and the deep learning model can obtain higher prediction precision by using less input data, but the physical relation between the input parameters and runoff is unclear and has no interpretability.
The invention provides a mixed model runoff prediction method, which comprises the steps of preprocessing collected and sorted data sets, preprocessing time sequence data, filling missing values by using a linear interpolation method, dividing all the data sets into a calibration set (or a training set) and a test set according to time, calibrating a HIMS physical model by using the calibration data set, and predicting runoff in a time range of the test set by using the calibrated model.
And inputting the prediction result of the HIMS model, the rainfall, the maximum value of the daily air temperature, the minimum value of the daily air temperature and the average value of the daily air temperature which are main factors influencing the runoff into an LSTM deep learning network, and predicting the runoff by capturing time series characteristics between input data and the runoff by using the LSTM model, thereby reducing error accumulation caused by time series increase. And finally obtaining a runoff prediction method combining a physical model and an LSTM model.
The invention applies some new physical methods and deep learning combined methods to the upstream runoff prediction of the black river, ensures the physical consistency of the upstream runoff while improving the prediction precision, generates a result with physical significance, and provides an important decision basis for scientifically making a water plan, improving the utilization rate of water resources and relieving the contradiction between supply and demand.
To summarize, to overcome the drawbacks of the two prior art methods in the background art, the present invention proposes a runoff prediction method for a hydrophysical model-guided deep learning model, which combines a physical equation-based model (if available) with a data-driven deep learning model to achieve predictive modeling of solar runoff upstream of a black river.
Drawings
FIG. 1 is a flow chart of the overall method of the present invention
FIG. 2 is a schematic diagram of a hybrid physical data model according to the present invention
Detailed Description
The invention is described in further detail below with reference to the figures and examples.
In consideration of the respective advantages of the hydrological physical model and the deep learning model, the invention provides a mixed model runoff prediction method for coupling the HIMS hydrological model and the LSTM deep learning model. As shown in fig. 1, the present invention includes the following specific method steps.
Performing preliminary runoff estimation based on a distributed hydrological model HIMS:
the method comprises the following specific implementation steps:
substep 1-1: acquiring the lowest temperature, the highest temperature, precipitation data and hydrological runoff data of a meteorological station from 2000 to 2016 in a research area;
substeps 1-2: data preprocessing: missing value filling is carried out on the meteorological data and runoff data obtained in the sub-step 1-1 by using a linear interpolation method to obtain a data set with continuous time series, the data set is adjusted to a TXT input data format of an HIMS model, runoff data are 4 columns, the first 3 columns are respectively years, months and days, the last column of a runoff data file is runoff data of a date, an air temperature data file is 5 columns, the first 3 columns are respectively years, months and days, and the last two columns are respectively the highest temperature and the lowest temperature.
Substeps 1-3: selecting data of a time sequence from 2000 to 2009 as rate periodic data to rate the HIMS model; the invention discloses a parameter automatic optimization method provided by a parameter calibration selection system (namely, the optimized parameters adopt parameters provided by an HIMS system platform), wherein 3 calibration period evaluation indexes selected by the method are volume error, efficiency coefficient (Nash efficiency coefficient) and correlation coefficient (Pearson correlation coefficient), and the formulas are respectively as follows:
wherein Q is obs,i And Q sim,i Respectively actually measured and simulated runoff sequences;
and
actual measurement and simulation of average runoff for many years are respectively carried out; n is the number of the measured runoff.
When the volume error is less than +/-10%, the model is basically practical, the Nash efficiency coefficient is used for evaluating the closeness degree of the predicted value and the measured value of the model, the range is (-infinity, 1), the closer to 1, the higher the prediction accuracy of the model is, the correlation coefficient represents the correlation between the predicted value and the measured value, and the closer to 1, the absolute value of the correlation coefficient represents the better the linear correlation between the predicted runoff quantity and the measured runoff quantity of the model is.
Substeps 1-4: selecting data of a time sequence from 2010 to 2016 as verification period data, and verifying the HIMS model obtained by the HIMS model rate periodically by integrating 3 evaluation indexes of volume error, nash efficiency coefficient and correlation coefficient;
specifically, for the HIMS model obtained through the calibration in the substep 1-3, the corresponding runoff is simulated through the data in the verification period and is compared with the observed runoff to verify the HIMS model.
Substeps 1-5: estimating the runoff by using the HIMS model verified in the substeps 1-4 to obtain a preliminary runoff estimation Q';
(II) obtaining an estimation of daily runoff from 2013 to 2016 based on the HPD model:
as shown in FIG. 2, the present invention combines the input D and the output Y of the physical model PHY Taken together as a deep learning model f PHY The input characteristics of the physical data model are used for constructing a mixed physical data model HPD; and (3) based on the constructed mixed physical data model HPD, taking the preliminary runoff data obtained by the hydrological model estimation in the step (I) and other site observation data as a training data set, obtaining an optimal mixed physical data model HPD, and carrying out daily runoff estimation by using the optimal mixed physical data model HPD. The observation data of other sites are precipitation, lowest temperature, highest temperature, average temperature and the like in the ERA5 Daily Aggredates data set.
Substep 2-1: combining the complete data set subjected to missing value filling and abnormal value processing in the substep 1-2 with the preliminary runoff estimation Q' provided in the step (one) to form a new data set;
substep 2-2: dividing a training set and a testing set: taking data from 2000 to 2012 as training set data, and taking data from 2013 to 2016 as test set data;
substeps 2-3: data normalization: carrying out standardization processing on the divided training set and the test set by using a maximum and minimum normalization algorithm, wherein a Min-Max standardization formula is as follows:
wherein, max (x) i ) Represents the maximum value in the array, min (x) i ) Represents the minimum value in the array, x i Represents data before conversion, x' i Representing the normalized data;
substeps 2-4: constructing a Hybrid Physical Data (HPD) model framework, which is based on the principle shown in FIG. 2, and taking the input and the output of a physical model as the input characteristics of a deep learning model together to generate a final output; wherein, the physical model is a distributed hydrological model HIMS model, the runoff Q' acquired in substeps 1-5 is adopted in the physical model output, and the deep learning model is an LSTM model in the HPD model, and the specific principle steps are as follows:
f t ＝σ(W f ·[h t-1 ,x t ]+b f )
i t ＝σ(W i ·[h t-1 ,x t ]+b i )
o t ＝σ(W o [h t-1 ,x t ]+b o )
h t ＝o t *tanh(C t )
wherein, W f 、W i 、W C And W o Different weights from the input gate, the forgetting gate and the output gate respectively; b f 、b i 、b C And b o Is a corresponding deviation value, f t 、i t And o t Is the output of the activation function at time t.
Substeps 2-5: inputting the learning rate, the dropping rate and the iteration times of the hyper-parameters into the constructed HPD model framework, debugging the model by using the divided training set, and selecting the evaluation index as
Wherein Q is obs,i And Q sim,i Respectively actually measured and simulated runoff sequences;
and
actual measurement and simulation of average runoff for many years are respectively carried out; and N is the number of the actual measurement runoff. NSE is the nash efficiency coefficient; RMSE is the root mean square error; r is the Pearson correlation coefficient.
Substeps 2-6: testing the debugged model by using the divided test set, comprehensively considering 3 evaluation coefficients of decision coefficients, root-mean-square errors and efficiency coefficients to evaluate the model, and selecting an optimal HPD model;
the sense of the Nash efficiency coefficient NSE and the Pearson correlation coefficient r is the same as that in substeps 1-3, and is used for verifying the accuracy and correlation of the model prediction result, and RMSE is the root mean square error and is used for measuring the deviation between the predicted runoff and the measured value.
Substeps 2-7: acquiring daily runoff estimation Q from 2013 to 2016 through an optimal HPD model ＊ 。
Finally, it should be noted that the above embodiments are only used for illustrating the technical solutions of the present invention and not for limiting, and although the present invention is described in detail with reference to the preferred examples, it should be understood by those skilled in the art that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, which should be covered by the claims of the present invention.
Claims (6)
1. A method for estimating the daily-scale runoff based on a physical model mixed deep learning model is characterized by comprising the following steps:
carrying out calibration and verification through a distributed hydrological model HIMS by utilizing meteorological data and runoff data in a research area, and carrying out primary runoff estimation based on the calibrated and verified distributed hydrological model HIMS to obtain a primary runoff estimation Q';
step (two) is to input D and output Y of the physical model PHY Taken together as a deep learning model f PHY The input characteristics of the physical data model are used for constructing a mixed physical data model HPD; and (3) based on the constructed mixed physical data model HPD, taking the initial runoff data obtained by the hydrological model estimation in the step (I) and observation data of other sites as training data sets to obtain an optimal mixed physical data model HPD, and carrying out daily runoff estimation by using the optimal mixed physical data model HPD.
2. The method for estimating the daily runoff based on the physical model hybrid deep learning model as claimed in claim 1, wherein the step (one) comprises the following substeps:
substep 1-1: acquiring the lowest temperature, the highest temperature, precipitation data and hydrological runoff data of a meteorological station from 2000 to 2016 in a research area;
substeps 1-2: data preprocessing: missing value filling is carried out on the meteorological data and runoff data obtained in the sub-step 1-1 by using a linear interpolation method to obtain a data set with continuous time series, and the data set is adjusted into a TXT input data format of an HIMS model;
substeps 1-3: selecting data of a time sequence from 2000 to 2009 as rate periodic data, and carrying out parameter calibration on the HIMS model;
substeps 1-4: selecting data of a time sequence from 2010 to 2016 as verification period data, and verifying optimization parameters of the HIMS model rate period by integrating a volume error, a Nash efficiency coefficient and a Pearson correlation coefficient;
substeps 1-5: and (5) estimating the runoff by using the HIMS model verified in the substeps 1-4 to obtain a preliminary runoff estimation, which is denoted by Q'.
3. The method for estimating the daily runoff based on the physical model hybrid deep learning model according to claim 2, wherein the parameter calibration method in the substeps 1-3 is as follows: carrying out automatic optimization based on an HIMS system platform, wherein the evaluation coefficients adopted by the parameter calibration timing comprise: volume error, nash efficiency coefficient, and pearson correlation coefficient.
4. The method for estimating the daily-scale runoff based on the physical model hybrid deep learning model according to claim 2 or 3, wherein the volume error, the Nash efficiency coefficient and the correlation coefficient are expressed as follows:
wherein, V e Is the volume error; NSE is the efficiency coefficient; r is a correlation coefficient; q obs,i And Q sim,i Respectively actually measured and simulated runoff sequences;
and
are respectively provided withActual measurement and simulation of average runoff for many years; n is the number of actual measurement runoff.
5. The method for estimating the daily runoff based on the physical model hybrid deep learning model according to any one of claims 1 to 4, wherein the step (two) is specifically as follows:
substep 2-1: combining the complete data set subjected to missing value filling in the substep 1-2 with the preliminary runoff estimation Q' to form an updated data set;
substep 2-2: dividing a training set and a testing set: taking the updated data set from 2000 to 2012 as training set data, and taking the updated data set from 2013 to 2016 as test set data;
substeps 2-3: data normalization: carrying out standardization treatment on the divided training set and the test set by utilizing a maximum and minimum normalization algorithm:
wherein, max (x) i ) Represents the maximum value in the array, min (x) i ) Represents the minimum value in the array, x i Represents the data before conversion, x' i Representing the normalized data;
substeps 2-4: constructing a HPD framework of the mixed physical data model: input D and output Y of the physical model PHY Taken together as a deep learning model f PHY To produce a final output;
substeps 2-5: setting the learning rate, rejection rate and iteration times of the hyperparameter learin the HPD frame of the mixed physical data model constructed in the substeps 2-4, and debugging the model by using the divided training set, wherein the evaluation coefficients selected by debugging are as follows: NSE, RMSE and r, wherein the RMSE is expressed as follows:
wherein Q obs,i And Q sim,i Respectively actual measurement runoff sequence and simulation runoff sequence, wherein N is the number of actual measurement runoff;
substeps 2-6: testing the debugged model by using the divided test set, evaluating the HPD of the mixed physical data model by using a decision coefficient, a root-mean-square error and an efficiency coefficient as evaluation coefficients, and selecting an optimal HPD model;
substeps 2-7: acquiring daily runoff estimation Q from 2013 to 2016 through an optimal HPD model ＊ 。
6. The method for estimating daily runoff according to claim 5 based on the physical model hybrid deep learning model, wherein the deep learning model f is PHY The LSTM model is adopted for realization.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211088365.5A CN115422840B (en) | 2022-09-07 | 2022-09-07 | Daily-scale runoff estimation method based on physical model hybrid deep learning model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211088365.5A CN115422840B (en) | 2022-09-07 | 2022-09-07 | Daily-scale runoff estimation method based on physical model hybrid deep learning model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115422840A true CN115422840A (en) | 2022-12-02 |
| CN115422840B CN115422840B (en) | 2023-10-27 |
Family
ID=84201645
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211088365.5A Active CN115422840B (en) | 2022-09-07 | 2022-09-07 | Daily-scale runoff estimation method based on physical model hybrid deep learning model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115422840B (en) |
-
2022
- 2022-09-07 CN CN202211088365.5A patent/CN115422840B/en active Active
Non-Patent Citations (6)
| Title |
|---|
| CHENCHEN ZHANG 等: "RUNOFF ESTIMATION BASED ON HYBRID-PHYSICS-DATA MODEL", pages 3 * |
| 吴梦莹;王中根;党素珍;: "基于HIMS的黑河上游山区径流模拟与分析", 资源科学, no. 10 * |
| 秦隆宇;: "基于辽河流域的贝叶斯模型平均径流模拟及不确定性研究" * |
| 秦隆宇;: "基于辽河流域的贝叶斯模型平均径流模拟及不确定性研究", 水土保持应用技术, no. 05 * |
| 苏辉东;贾仰文;倪广恒;龚家国;曹雪健;张明曦;牛存稳;张迪;: "机器学习在径流预测中的应用研究" * |
| 苏辉东;贾仰文;倪广恒;龚家国;曹雪健;张明曦;牛存稳;张迪;: "机器学习在径流预测中的应用研究", 中国农村水利水电, no. 06 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115422840B (en) | 2023-10-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110766212B (en) | Ultra-short-term photovoltaic power prediction method for historical data missing electric field | |
| CN102183621B (en) | Aquaculture dissolved oxygen concentration online forecasting method and system | |
| CN111310968A (en) | LSTM neural network circulation hydrological forecasting method based on mutual information | |
| CN109711617B (en) | Medium-and-long-term runoff prediction method based on BLSTM deep learning | |
| CN105740991A (en) | Climate change prediction method and system for fitting various climate modes based on modified BP neural network | |
| US20210125200A1 (en) | Method and system for predicting medium-long term water demand of water supply network | |
| CN116150897A (en) | Machine tool spindle performance evaluation method and system based on digital twin | |
| CN111767517A (en) | BiGRU multi-step prediction method and system applied to flood prediction and storage medium | |
| CN110442911B (en) | High-dimensional complex system uncertainty analysis method based on statistical machine learning | |
| CN112330065A (en) | Runoff forecasting method based on basic flow segmentation and artificial neural network model | |
| CN111753965A (en) | Deep learning-based river flow automatic editing method and system | |
| CN114154716B (en) | Enterprise energy consumption prediction method and device based on graph neural network | |
| CN111311026A (en) | Runoff nonlinear prediction method considering data characteristics, model and correction | |
| CN114692981A (en) | Medium-and-long-term runoff forecasting method and system based on Seq2Seq model | |
| CN103279030B (en) | Dynamic soft measuring modeling method and device based on Bayesian frame | |
| CN114357737A (en) | Agent optimization calibration method for time-varying parameters of large-scale hydrological model | |
| CN113988210A (en) | Method and device for restoring distorted data of structure monitoring sensor network and storage medium | |
| CN112307536B (en) | Dam seepage parameter inversion method | |
| CN110852415B (en) | Vegetation index prediction method, system and equipment based on neural network algorithm | |
| CN116960962A (en) | Mid-long term area load prediction method for cross-area data fusion | |
| CN112581311B (en) | Method and system for predicting long-term output fluctuation characteristics of aggregated multiple wind power plants | |
| CN115422840A (en) | Ridge-scale runoff estimation method based on physical model mixed deep learning model | |
| CN113487069B (en) | Regional flood disaster risk assessment method based on GRACE daily degradation scale and novel DWSDI index | |
| CN115330085A (en) | Wind speed prediction method based on deep neural network and without future information leakage | |
| CN114372615A (en) | Short-term photovoltaic power prediction method and system considering spatial correlation |
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 |