72 results about "Flood forecast" patented technology

The invention discloses a design method for a distributed hydrological model by using a grid as an analog unit, which is called as ESSI for short. The design method comprises the following steps of: obtainment of a distributed parameter: converting vector data into grid data; generalization of a watershed hydrological process: establishing a universal runoff generating type for the grid; design of a runoff generating process: respectively computing the water-quantity distributing condition of each part according to different prior-period soil water conditions; design of a flow collecting process: respectively designing a Muskingum-Cunge method, a delay algorithm and a riverway segmentation Muskingum method for the flow collecting process computation of the model under different conditions; and model development and integration. The invention not only can finish the watershed hydrological process simulation at an arid region and a humid region, but also can realize the short-term flood forecast and the long-term rainfall-runoff process simulation and prediction of a watershed by using modularization and integration ideas as means, thereby providing scientific reference bases for deeply learning about the physical mechanism of water circulation by people, reducing the drought and water-logging disasters and reasonably developing and utilizing water resources.

The invention discloses a burst disaster warning system establishing method for a barrier lake in a data-lacking-area, and belongs to the technical field of disaster warning. The burst disaster warning system establishing method integrates four main processes, flood forecast, discharged flow drafting, dam bursting flood routing and disaster warning. The whole system is constructed as a complete platform, so that the process for inputting basic data and outputting warning signals can be realized, and flood forecast, dam bursting warning and flood routing can be carried out in the system. For calculation of the dam breach peak point flow, by means of deep research of a Xie Renzhi formula and integration with a physical model experiment, the soil coefficients and the flow coefficients for different ingredient barrier bodies can be obtained and the applicability of the formula to a barrier lake can be improved.

The invention provides an ensemble rainfall forecast and real-time flood probability forecast coupling method. The method comprises the steps of collecting hydro-meteorological basic materials of drainage basins and ensemble forecast rainfall data and building a hydrological model to obtain ensemble flood certainty forecasts; on the basis of real-time correction of error, converting corrected ensemble flood certainty forecasts into ensemble flood probability forecasts; performing weight fusion on the obtained ensemble flood probability forecasts to obtain comprehensive real-time flood probability forecasts. The method couples ensemble rainfall forecast information and Copula-BFS and is capable of considering the comprehensive influence of uncertainties such as rainfall forecast input, hydrological model parameters and model structures and estimating the total uncertainty of flood forecasts quantitatively. By performing error real-time correction on corresponding flood certainty forecasts before flood probability forecasts, the flood forecast precision can be increased and the performance of flood uncertainty forecast intervals is improved.

The invention discloses a multi-model meta-synthesis flood forecasting system which comprises a data integration module, component integration module, model integration module, scheme integration module and a result publishing module, wherein the data integration module, the component integration module, the model integration module, the scheme integration module and the result publishing module are sequentially connected. The invention also discloses a method for flood forecast by utilizing the multi-model meta-synthesis flood forecasting system, comprising the following steps of unifiedly treating historical or real-time hydrological data through the data integration to enable the hydrological data to accord with the use specifications and the requirements of the component integration; packaging each set of obtained data into a plurality of components according to the needs of a constructed framework; obtaining a plurality of flood forecasting models by respectively setting up the plurality of components on the basis of the structure requirements of the models and obtaining corresponding flood forecasting results; obtaining a final forecasting result and a final forecasting scheme by treating the plurality of flood forecasting results in a unified way; and publishing the final result. The method realizes rapid setting up of various flood forecasting models and can provide a plurality of schemes and scheme optimizations.

The invention provides a land-atmosphere coupling-based method and a system for flood forecast of minor watersheds, which are used for solving the problem of low accuracy for the flood forecast of the minor watersheds. The method comprises the following steps: setting model parameters; inputting initial data; judging a runoff-generation manner through analyzing the initial data, calculating surface runoff if the runoff-generating manner is infiltration excess runoff, and taking the surface runoff as simulation runoff; calculating the infiltration amount of soil if the runoff-generating manner is saturation excess runoff, and calculating base runoff and subsurface runoff by using the soil infiltration amount; calculating to obtain the simulation runoff according to the base runoff and the subsurface runoff when the soil is unsaturated; and further calculating the surface runoff and calculating to obtain the simulation runoff according to the base runoff, the subsurface runoff and the surface runoff when the soil is saturated. The land-atmosphere coupling-based method and the system provided by the invention are suitable for calculating the simulation runoff of the minor watersheds, and the calculating result is more accurate through the combination of theory and practice, thus the land-atmosphere coupling-based method and the system are suitable for short-term or ultra-short-term runoff forecast and have high forecast accuracy.

The invention relates to an optimal selecting method of real-time correction models in a flood forecast system, which belongs to the technical field of flood forecast. The optimal selecting method comprises the following steps of: firstly, configuring at least two existing real-time correction models and selecting a precision evaluating indicator for the flood forecast system; extracting a time period just before the forecast time of the flood forecast system as an attempting forecast time period after system initialization; respectively applying the real-time correction models to the attempting forecast time period for attempting real-time correction forecast so as to obtain attempting real-time correction forecast runoffs; respectively counting the precision evaluating indicators of the attempting real-time correction forecast runoffs, and selecting the real-time correction model corresponding to the attempting real-time correction forecast runoff with the optimal precision evaluating indicator as the optimal real-time correction model. The optimal selecting method adopts multiple kinds of the existing real-time correction models for multiple times of attempting real-time correction forecasts so as to select the optimal real-time correction model applicable to floods in the current valley in the flood forecast system, and thus, the precision of the flood forecast system in actual forecasts is increased.

The invention discloses a data-driven small watershed real-time flood forecast method comprising the following steps: (1) constructing a delay time sequence and a mean time sequence based on rainfall and river flow data; (2) calculating the grey relevance between the time sequences, and selecting the higher-correlation time sequence as modeling sample data; (3) modeling input-output data by use of a machine learning algorithm to get forecast models and a fitting value; (4) calculating the weight of each forecast model based on the information entropy theory; and (5) forecasting the river flow by use of the forecasting models, and performing weighted calculation on single-model forecast results to get a forecasted value of river flow. According to the invention, the complex nonlinear mapping relationship of small watershed flood is simulated by making full use of rainfall and river flow data, the information in the data is mined, and therefore, the accuracy of real-time small watershed flood forecast is improved.

The invention discloses a kernel extreme learning machine flood forecast method based on sparse self-encoding. The method comprises the following steps that (1) medium and small river flood data are selected, and the data are collated and cleaned; (2) the appropriate forecast factors are selected and samples are collated, and the sample data are preprocessed; (3) unsupervised learning is performedon the original sample data through a multilayer sparse self-encoder, and the optimal network layer parameters are trained; and (4) the sample data through multilayer sparse self-encoding learning act as the input of a KELM model, an SAE_KELM model is constructed and the corresponding result is predicted and evaluated. The SAE method is fused on the basis of the KELM model, a deep network model is constructed, and the abundant intrinsic information between the complex data can be learnt by increasing the number of layers of the model. The "essential" characteristics of the original hydrological data can be learnt by the SAE_KELM model, the learnt characteristics have more essential depiction for the data, and the mapping relationship between the characteristic value and the target value can be better fit by the KELM model.

The invention discloses a method for forecasting flood based on a Boosting algorithm and a support vector machine, which comprises following steps: use the correlation coefficient method to determine the forecast factors; utilize kernel principal component analysis to process the forecast factors with dimension reduction; utilize the Boosting algorithm to select a sample and establish a plurality of support vector machine prediction models, introduce loss function and the correlation coefficient to adjust sample weight, and finally combine the plurality of prediction models as a total prediction model; and utilize the total prediction model to predict a test sample. In the invention, the previous steps are about data pre-processing, which aims to extract useful information in flood datum and eliminate disturbance of redundant information to the forecast; in the third step, the Boosting algorithm is introduced into the flood forecast so as to try to extract a sample of one model that can't learn well for training the next model; in this way, the accuracy of flood forecast can be improved effectively by the combined model; and the last step is used for testing the model effect. According to the experimental datum, the forecast accuracy can be improved effectively by the technical solution.

The invention relates to a river channel water and sediment real-time prediction method based on data assimilation, and belongs to the technical field of water conservancy projects. The method includes the steps of firstly, collecting topographic data of a water channel to be predicated, upstream boundary condition data, downstream boundary condition data and fracture surface data of the river channel, setting up a one-dimensional non-steady-flow and non-balance sediment transport model, and solving the model; secondly, conducting water and sediment model assimilation on real-time observation data while receiving the real-time observation data, and enabling the assimilation value to serve as an initial field for calculation; thirdly, calculating the changes of the future water level, the further flow and the future sediment concentration according to information of the initial field and the boundary condition data. According to the method, the water level and flow speed prediction function, the flood forecast function and other functions are achieved on the basis of data such as the assimilation real-time water level, the flow speed and the sediment concentration, and the method has the advantages of being strong in pertinency, comprehensive in function, convenient to use, practical and the like, can be applied to river channel flood real-time forecasting of great rivers, and provides the decision-making support for the practical flood prevention command work.

The invention discloses a medium and small river integrated forecasting method based on negative correlation learning, and the medium and small river integrated forecasting method comprises the steps:firstly carrying out the specific analysis according to different watershed features and forecasting requirements, determining the research content, and carrying out the analysis of data; performingdata preprocessing, and selecting the data with the highest correlation with the prediction result to construct model input and output data; based on the idea of ensemble learning, combining the characteristics of a target drainage basin and the complexity of a sample data set to select sub-networks forming an integrated neural network and determine the structure of the sub-networks; constructingan integrated forecasting model by using a negative correlation learning method, and selecting an optimization algorithm and a loss function to train and optimize the model under different hyper-parameter conditions; and performing flood forecasting by using the model, calculating a corresponding flood process evaluation index to evaluate the forecasting effect of the model, and performing corresponding real-time forecasting by using the preprocessed hydrological historical data as the input of the integrated forecasting model and the basin outlet section flow corresponding to the forecast period moment as the output of the integrated forecasting model when the model is applied to an actual scene.

The invention relates to an ensemble forecast flood forecasting method for improving forecast rainfall based on data assimilation. The ensemble forecast flood forecasting method comprises the following steps: step 1, data preparation; step 2, rainfall ensemble forecast based on the data assimilation; step 3, rainfall data post-processing; step 4, ensemble flood forecast; step 5, flood forecast result publishing and evaluation. The ensemble flood forecasting method provided by the invention has the benefits that not only is the forecast period of the flood forecast prolonged, but also the forecast accuracy of the flood forecast is improved; compared with traditional flood forecast performed by selecting through fall, a system selects a data assimilation technology to correct a numerical atmosphere mode, so that the high-resolution accurate forecast rainfall is obtained in advance and is taken as an input to a hydrological model for the flood forecast; in addition, in order to reduce the forecast uncertainty, an idea of the ensemble forecast is introduced for the rainfall forecast and the flood forecast.

The invention belongs to the technical field of flood forecasting, and relates to a method for inquiring the stage design flood of a reservoir. The method comprises the following steps: flood information with randomness and regularity is converted into directional data by applying a round fractional step method, and determining the stage time intervals of the reservoir by calculating a vector average angle, average radius and concentration ratio of the directional data; establishing a trivariate combined distribution model based on a three-dimensional asymmetric Copula function for constructing edge distribution to be P-III type distribution, and analyzing the correlation among stage flood by applying the trivariate combined distribution model; determining a deterministic relationship between the reservoir stage design flood frequency and a flood control standard by applying the trivariate combined distribution model; and finally determining the stage design flood frequency and a design value. According to the method, the defect of a conventional design flood inquiry technology can be overcome, so that the result can fulfill the flood control standard and can reflect the seasonal regulation of the flood.

The invention discloses a reservoir flood regulation multi-dimensional uncertainty risk analysis method based on a Bayesian network. The reservoir flood regulation multi-dimensional uncertainty risk analysis method comprises the following steps: identifying risk factors of regulation starting water level uncertainty and flood forecasting uncertainty; carrying out bayesian network structure learning based on an expert experience method; carrying out bayesian network parameter learning to acquire a conditional probability table (CPT) of each node; performing Bayesian network probability reasoning; carrying out risk calculation and analysis. In order to couple the influence of regulation starting uncertainty and reservoir flood forecasting errors on reservoir flood control scheduling risks, areservoir flood control risk analysis model based on the Bayesian network is established, and coordinated conversion of reservoir interest benefits and flood control risks in the flood season can beachieved; the bidirectional reasoning of the Bayesian network can establish a reservoir scheduling risk bidirectional analysis and evaluation mode, and has a good application prospect.

The invention belongs to the technical field of information processing, and discloses a BiGRU multi-step prediction method and system applied to flood prediction and a storage medium. In an Attentionmechanism layer, a hidden layer state sequence vector is input into a learnable function to generate a probability vector, and a subsequent hidden layer intermediate vector is generated by vector weighted averaging; the Attention circularly calculates the adaptive weighted average of the hidden layer state sequence of each time step to generate the intermediate vector; the important information ofeach time step is output backwards according to a certain weight; and the information integration capability is realized as time goes on. According to the BiGRU multi-step prediction method, the information can be stored in the memory as time goes on, and the BiGRU multi-step prediction method has great advantages in processing time sequence problems; and in combination with an Attention mechanism, a test result shows that the BiGRU multi-step flood forecasting model based on the Attention mechanism can better forecast the flood peak arrival time and the flood peak value.

The invention discloses a flood forecasting method suitable for a runoff data lack basin based on machine learning. The flood forecasting method comprises the following steps: 1) extracting and parameterizing sample basin features; 2) carrying out basin flood response characteristic analysis; 3) generating a drainage basin feature sample set; 4) generating a classification tree based on the basinfeature sample set; 5) generating a training data set based on the tree nodes; 6) carrying out flood forecasting based on the classification tree and the data driving model; and 7) updating the classification tree and the training set. Flood response characteristics of the drainage basin are analyzed by utilizing a machine learning algorithm; and based on the watershed characteristics and the flood response characteristics, an association relationship between watersheds is established. According to the method, the sample data set is generated on the basis of the basin characteristics and the flood response similarity, then the data driving model is trained according to the sample data set, the rainfall and flood response relation of the medium and small rivers is simulated, and therefore real-time forecasting of the flood of the medium and small rivers is achieved. According to the method provided by the invention, the data driving model can be applied to flood forecasting of runoff data lack drainage basins, and the dependence of a previous parameter transplanting mode on a model structure and model parameters is changed, so that the flood forecasting precision is improved.

The invention discloses a flood forecasting method under influences of human activities. A conventional flood forecasting method is improved from three links: history data analysis, human activity influence analysis and standard determination, and forecasting model parameter calibration which influence flood forecasting work precision, and the purpose of improving the flood forecasting work precision is achieved. The influences of human activities are quantificationally assessed and analyzed on the base of qualitative analysis, and different strategies and methods are adopted for different influencing factors, so that the practicalness is relatively high, and the method system is complete, etc.

The embodiment of the invention provides a mountain torrent disaster early warning method and device, a server and a storage medium. By monitoring data acquisition, calibrating the initial hydrological forecasting model according to the standard water flow in the monitoring data; taking the calibrated model as a target hydrological forecasting model; carrying out flood forecasting calculation on the monitoring data according to the target hydrological forecasting model; and performing evolution overflow calculation on the prediction calculation result according to the evolution overflow modelto obtain an overflow flooding analysis result of the current monitoring area, and further determining whether to generate early warning information for alarming the current monitoring area accordingto the overflow flooding analysis result. Furthermore, the monitoring data is analyzed by combining the target hydrological forecasting model and the evolution overflow model; the purposes of integrated data collection, data analysis and alarming are achieved, the effects of improving mountain torrent disaster early warning and reducing the labor cost are achieved, meanwhile, the early warning precision and the early warning efficiency are improved, and the requirement for accurate and efficient early warning of mountain torrent disasters is met.

The invention discloses a multi-station linkage rating curve fitting method, and the method comprises the steps: selecting a test station; determining a flood forecast period l of the test station; constructing a rating curve (shown in the description) of the test station, wherein t is a sampling moment, N is the number of actual measurement stations of the upstream side of the test station, i isthe serial number of the actual measurement stations of the upstream side of the test station, Z<t+l>, Q<t+l> and a<t+l> are respectively are a water level predicted value, a flow predicted value anda flow coefficient of the test station in the forecast period, a<i, t>, Q<i, t>, Ab<i, t> and Z<i, t> are respectively a flow coefficient, a flow value, a water level coefficient and a water level value of the i-th actual measurement station at a moment t, c is an offset of the flow of the test station relative to the water level in the forecast period, Epsilon is a remainder term and approaches to an infinitesimal value, and l is nit greater than the time lag of a flood wave from the upstream side to a downstream side. According to the invention, the real-time stage-discharge relation of thetest station is represented as a function of the water level of the upstream station and the flow at a historical moment, and the method can obtain the more precise result in the real-time flood forecast applications in different forecast periods, and provides a real-time flood forecast level and the forecast precision.

The invention provides a flood forecasting method based on CS-LSTM, which is used for solving the technical problem of poor prediction performance of a model caused by inaccurate parameter selection in an existing prediction model. The method comprises the steps that firstly, collecting characteristic parameters of a target hydrological station to serve as sample data, performing normalization processing on the sample data, and dividing the normalized sample data into a training set and a verification set according to the time sequence; secondly, constructing a flood forecasting model based ona long-short-term memory neural network, and performing iterative optimization on the flood forecasting model by utilizing a cuckoo search algorithm, a training set and a verification set to obtain afinal flood forecasting model; and finally, inputting the to-be-predicted characteristic parameters into the final flood forecasting model to obtain a prediction result. According to the method, theflood process in the future prediction period can be accurately predicted, a certain basis can be provided for hydrology and water resource management and reservoir scheduling, and a reference is provided for application of deep learning in the hydrology field.

The invention discloses a stream big data processing mode-based rainfall runoff prediction calculation method. Medium and large-scale basins are logically divided into different sub-basins, so that the performance problem of an SCS model and an improved SCS model in real-time flood forecast of the medium and large-scale basins is solved through parallel computing of the sub-basins in combination with real-time monitored cross section flow of the sub-basins in a stream type big data processing mode.

The invention discloses a small river flood forecasting method. By using the open source GIS codes for the secondary development, the GIS requirement of flood forecasting can be met. Meanwhile, through the classification of the model calculation engines, the flexible configuration of the models, the automatic extraction of a drainage basin and the matching association of a node topological relation are achieved, the automatic and flexible construction of a scheme is achieved, and the workload of scheme operators can be saved to a large extent while the forecasting precision of medium and smallrivers is improved.

The invention discloses a forecast scheduling method for reducing reservoir flood regulation starting water level by considering forecast errors, and belongs to the technical field of flood preventionforecast scheduling. The method comprises the following steps: firstly, performing flood forecasting feasibility analysis on a reservoir control basin, and identifying forecasting error distributionby adopting a maximum entropy principle; secondly, formulating a dispatching rule framework for reducing the flood starting water level of the flood rising section by utilizing a pre-discharge thoughtfor the whole dispatching system; optimizing the forecast scheduling framework to obtain a forecast scheduling scheme optimization point set; thirdly, screening out all optimization point sets meeting the upstream and downstream flood control safety under the condition of the maximum forecast error from the forecast scheduling scheme optimization point set; and finally, comprehensively considering forecast errors and different preferences of a decision maker, and evaluating optimal forecast scheduling scheme points by utilizing a binary comparison method and a fuzzy optimization model to obtain a final forecast scheduling scheme. The method is simple and easy to operate, and the flood control benefit of the reservoir and the elasticity of a downstream protection point under the flood action are improved under the condition of keeping the interest benefit.

The invention discloses a reservoir real-time flood control multi-target robust optimization regulation and control method for resisting flood forecast error disturbance. The method comprises the steps: collecting and arranging flood forecast error samples in a reservoir system database, and randomly simulating to generate a real-time flood control scheduling flood error scene set; establishing a reservoir flood control dispatching multi-target robust optimization model: taking the upstream and downstream flood control risk rates and the conditional value-at-risk of the upstream and downstream flood control risk loss as optimization targets, and generating a multi-target robust optimization non-inferior scheme solution set; and carrying out multi-attribute risk decision-making under the uncertain condition: introducing different weight parameters for different decision-making preferences, and obtaining optimal equilibrium solutions under different preferences by adopting a TOPSIS method. According to the method, a prediction-scheduling-decision-making whole process decision-making support model set integrating error scene simulation, risk robust regulation and control and non-inferior solution set decision-making is created in the field of reservoir flood control scheduling, and the reliability and safety of reservoir flood control scheduling can be improved.

The invention discloses a flood prediction method based on an extreme learning machine, and relates to the technical field of disaster prediction. According to the invention, a flood prediction modelis established according to various flood causes, an extreme learning machine and a geographic information system (GIS); the efficiency and precision advantages of the extreme learning machine relative to an artificial neural network are verified by determining a coefficient r, a Wilter index WI, a Nash efficiency index Ens, a root mean square error RMSE, a mean absolute error MAE and a related error RE. Experimental results show that the learning speed of the extreme learning machine is 32 times of that of an artificial neural network, the noise processing capacity of the extreme learning machine is superior to that of the artificial neural network, and compared with the artificial neural network, extreme learning has great advantages in prediction capacity and efficiency and is a more appropriate choice for flood forecasting models.

The invention discloses a back-fitting algorithm-based flood forecast real-time correction method. According to the method, a back-fitting algorithm in a data mining technology is introduced based on a conventional error autoregression correction method to correct parameters of a hydrologic model, and stationary processing of a residual error sequence is considered, so that a new method for improving conventional flood forecast real-time correction is proposed. The method can be widely applied to real-time flood forecast, can effectively improve the flood forecast precision, and provides an important basis for a flood prevention dispatching decision.

The invention discloses a flood forecasting method based on a novel general input and output structure and a long-short term memory network. Firstly, confluence characteristics are analyzed according to historical session flood data of a research basin, and average confluence time of the research basin is calculated; secondly, determining the number of output time periods of the LSTM flood forecasting model according to the average flow concentration time of the drainage basin, and giving the number of hidden layers and the number of neuron nodes of the hidden layers of the LSTM flood forecasting model; thirdly, designing an input and output structure, an input training set and a verification set sample training model of the novel universal LSTM flood forecasting model, and obtaining flood forecasting models with different structures and parameters; and finally, comparing and analyzing the performance of the LSTM flood forecasting model in the training set and the verification set under different input lengths, determining a final better LSTM flood forecasting model, and evaluating and analyzing the forecasting effect of the LSTM flood forecasting model in the test set. The method is high in universality, the established LSTM flood forecasting model can achieve a good forecasting effect, and new technical support is provided for flood disaster defense work of the drainage basin.