CN-121977605-A - Hydrologic prediction error correction method and device, electronic equipment and storage medium

CN121977605ACN 121977605 ACN121977605 ACN 121977605ACN-121977605-A

Abstract

The application provides a hydrological forecast error correction method, a hydrological forecast error correction device, electronic equipment and a storage medium, wherein the method comprises the steps of obtaining a hydrological forecast error time sequence of each monitoring site in a target flow field; the method comprises the steps of establishing an error propagation diagram based on a communication relation among monitoring stations in a target flow, carrying out joint processing on the propagation process of errors in a space-time dimension based on time evolution characteristics represented by an error time sequence and space propagation characteristics represented by the error propagation diagram, outputting an error correction interval of each monitoring station in a prediction period, and correcting original hydrological forecast data of each monitoring station in the prediction period based on the error correction interval to obtain corrected hydrological forecast results. The method provided by the embodiment of the application is beneficial to solving the technical problems of insufficient correction precision and lower reliability of correction results of hydrological prediction errors in the prior art.

Inventors

WANG GUOMIAO
ZHAO HAORAN
WU JIANMING

Assignees

浙江远算科技有限公司

Dates

Publication Date: 20260505
Application Date: 20260407

Claims (10)

1. A method for correcting hydrologic forecast errors, the method comprising: acquiring a hydrological forecast error time sequence of each monitoring site in a target flow; Constructing an error propagation diagram based on the communication relation among all monitoring stations in the target flow field; based on the time evolution characteristic represented by the error time sequence and the space propagation characteristic represented by the error propagation diagram, carrying out joint processing on the propagation process of the error in the space-time dimension, and outputting an error correction interval of each monitoring station in the foreseeing period; And correcting the original hydrologic forecast data of each monitoring site in the foreseeing period based on the error correction interval to obtain corrected hydrologic forecast results.
2. The method of claim 1, wherein after the constructing an error propagation map, the method further comprises: And aiming at each monitoring station, carrying out weighted aggregation on hydrologic prediction errors of upstream monitoring stations according to the communication relation, and superposing random disturbance of the monitoring stations to obtain an error estimation value for quantifying spatial propagation characteristics of the monitoring stations, wherein the error estimation value is used as one item of input data of the joint processing.
3. The method of claim 1, wherein constructing an error propagation map based on a connectivity relationship between monitoring stations within the target flow stream comprises: determining the upstream and downstream hydraulic communication relation among all monitoring stations according to the digital elevation model of the target river basin and the river network topology information; Constructing a watershed directed graph structure for representing a watershed water flow path by taking the monitoring station as a node and the upstream-downstream hydraulic communication relationship as a directed edge, wherein the direction of the directed edge is consistent with the water flow direction and is used for representing the hydraulic communication relationship between the nodes; And introducing an error propagation constraint rule on the basis of the watershed directed graph structure to obtain the error propagation graph, wherein the error propagation constraint rule is provided with a maximum acting range threshold of error propagation and is used for limiting the hydrologic prediction error to propagate from an upstream node to a downstream node along the directed edge only within the maximum acting range threshold.
4. A method according to claim 3, wherein the joint processing of the propagation of the error in the space-time dimension and outputting the error correction interval of each monitoring station in the foreseeable period comprises: Taking the error time sequence and the error propagation diagram as input data, and inputting the input data into a pre-trained space-time diagram neural network model; utilizing the space-time diagram neural network model to carry out joint modeling on a space propagation process and a time evolution process of errors in the watershed directed graph structure, and outputting probability distribution parameters of hydrologic prediction errors of each monitoring station in the prediction period; and constructing an error correction interval of each monitoring station in the foreseeing period according to the probability distribution parameters.
5. The method of claim 4, wherein the modeling the spatial propagation process and the time evolution process of the error in the watershed directed graph structure by using the space-time graph neural network model, and outputting the probability distribution parameters of the hydrologic prediction error of each monitoring station in the foreseeing period, comprises: utilizing a graph annotating force mechanism of the space-time graph neural network model to adaptively learn the influence weight of the error of the upstream monitoring site on the downstream monitoring site according to the hydraulic communication relation represented by the directed edge; According to the influence weight, carrying out weighted aggregation on errors of upstream monitoring stations to obtain error characteristics of each monitoring station after spatial aggregation; modeling an evolution process of the spatially aggregated error features in a time dimension by using a time modeling structure of the space-time diagram neural network model, and extracting a time sequence state representation representing error hysteresis and cumulative effect; And generating probability distribution parameters of hydrologic forecast errors of all monitoring stations in the forecast period through a probability output unit of the space-time diagram neural network model based on the time sequence state representation.
6. The method of claim 4, wherein the error correction interval comprises an expected value derived based on the probability distribution parameter and a float range determined based on an uncertainty strength of the expected value in the probability distribution parameter; The correcting the original hydrologic forecast data of each monitoring site in the foreseeing period based on the error correction interval comprises the following steps: Correcting the original hydrological forecast data based on the expected value to obtain a corrected forecast value; Determining a floating interval of the corrected forecast value based on the floating range; and taking the corrected forecast value and the floating interval as the corrected hydrologic forecast result.
7. The method of claim 6, wherein correcting the original hydrologic forecast data based on the expected value results in a corrected forecast value, comprising: Correcting the original hydrological forecast data based on the expected value to obtain an initial correction value; And introducing upstream and downstream consistency constraints based on the watershed directed graph structure, and adjusting the initial correction value to obtain a corrected forecast value, wherein the upstream and downstream consistency constraints are used for limiting the deviation between the correction result of the downstream monitoring station and the correction result of the upstream monitoring station not to exceed a preset threshold so as to ensure that the corrected forecast value accords with an upstream and downstream physical association rule of the hydrologic process.
8. A hydrological forecast error correction apparatus, the apparatus comprising: The acquisition module is used for acquiring the hydrological forecast error time sequence of each monitoring site in the target flow; The diagram construction module is used for constructing an error propagation diagram based on the communication relation among all monitoring stations in the target flow domain; The processing module is used for carrying out joint processing on the propagation process of the error in the space-time dimension based on the time evolution characteristic represented by the error time sequence and the space propagation characteristic represented by the error propagation diagram, and outputting an error correction interval of each monitoring station in the foreseeing period; and the correction module is used for correcting the original hydrologic forecast data of each monitoring site in the forecast period based on the error correction interval to obtain corrected hydrologic forecast results.
9. An electronic device comprising a processor, a storage medium and a bus, the storage medium storing machine-readable instructions executable by the processor, the processor in communication with the storage medium via the bus when the electronic device is in operation, the processor executing the machine-readable instructions to perform the steps of the hydrological forecast error correction method of any of claims 1 to 7.
10. A computer readable storage medium, characterized in that it has stored thereon a computer program which, when executed by a processor, performs the steps of the hydrological forecast error correction method according to any of claims 1 to 7.

Description

Hydrologic prediction error correction method and device, electronic equipment and storage medium Technical Field The application relates to the technical field of hydrologic forecasting, in particular to a hydrologic forecasting error correction method, a hydrologic forecasting error correction device, electronic equipment and a storage medium. Background Hydrologic forecasting is a technology for forecasting future changes of hydrologic elements such as runoff in a flow area, water level and the like, and is an important foundation for water resource scheduling and hydraulic engineering operation. In practical application, due to the influence of factors such as uncertainty of hydrologic model parameters, weather forecast errors, observed data deviations and the like, systematic deviations and random errors often exist in hydrologic forecast results. Therefore, error correction is carried out on the hydrologic forecasting result, and the method is a key technical link for improving forecasting precision and reliability. At present, the common hydrologic prediction error correction method is mainly used for correcting the current prediction result according to the mapping relation between the historical prediction value and the actual measurement value. The method still has the problems of insufficient correction precision, poor stability and the like in practical application, and is difficult to fully meet the requirements of engineering scenes such as water resource scheduling, water resource allocation and the like on the reliability of the forecasting result. Disclosure of Invention In view of the above, the embodiments of the present application provide a method, an apparatus, an electronic device, and a storage medium for correcting a hydrological forecast error, so as to solve the technical problems of insufficient accuracy and low reliability of a correction result in the prior art. In a first aspect, an embodiment of the present application provides a method for correcting a hydrological forecast error, where the method includes: acquiring a hydrological forecast error time sequence of each monitoring site in a target flow; Constructing an error propagation diagram based on the communication relation among all monitoring stations in the target flow field; based on the time evolution characteristic represented by the error time sequence and the space propagation characteristic represented by the error propagation diagram, carrying out joint processing on the propagation process of the error in the space-time dimension, and outputting an error correction interval of each monitoring station in the foreseeing period; And correcting the original hydrologic forecast data of each monitoring site in the foreseeing period based on the error correction interval to obtain corrected hydrologic forecast results. In one possible embodiment, after the constructing the error propagation map, the method further comprises: And aiming at each monitoring station, carrying out weighted aggregation on hydrologic prediction errors of upstream monitoring stations according to the communication relation, and superposing random disturbance of the monitoring stations to obtain an error estimation value for quantifying spatial propagation characteristics of the monitoring stations, wherein the error estimation value is used as one item of input data of the joint processing. In one possible embodiment, the constructing an error propagation map based on a communication relationship between monitoring stations in the target flow field includes: determining the upstream and downstream hydraulic communication relation among all monitoring stations according to the digital elevation model of the target river basin and the river network topology information; Constructing a watershed directed graph structure for representing a watershed water flow path by taking the monitoring station as a node and the upstream-downstream hydraulic communication relationship as a directed edge, wherein the direction of the directed edge is consistent with the water flow direction and is used for representing the hydraulic communication relationship between the nodes; And introducing an error propagation constraint rule on the basis of the watershed directed graph structure to obtain the error propagation graph, wherein the error propagation constraint rule is provided with a maximum acting range threshold of error propagation and is used for limiting the hydrologic prediction error to propagate from an upstream node to a downstream node along the directed edge only within the maximum acting range threshold. In a possible embodiment, the joint processing of the propagation process of the error in the space-time dimension and outputting the error correction interval of each monitoring station in the foreseeing period include: Taking the error time sequence and the error propagation diagram as input data, and inputting the input data into a pre-trained