Hydrological deduction and error tracing method and system under influence of multiple factors
By constructing a gridded influence field, the spatial location problem of error source tracing in traditional hydrological extrapolation is solved, achieving efficient and accurate error source tracing, supporting precise management decisions, and reducing model complexity and safety risks.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SICHUAN HUANENG JIALINGJIANG HYDROPOWER CO LTD
- Filing Date
- 2026-02-12
- Publication Date
- 2026-04-24
AI Technical Summary
Traditional hydrological extrapolation methods cannot effectively link macroscopic water quantity errors with microscopic geospatial events, resulting in error tracing remaining at the factor level and failing to pinpoint specific spatial locations. Furthermore, the errors increase under non-steady flow conditions, lacking an effective differentiation mechanism, which affects the pertinence and safety of management decisions.
By constructing a 'gridized impact field' bound to real geographic space, the area above the reservoir inlet section is divided into a computational grid. Static attribute vectors are established, multi-source data streams are acquired and mapped to the grid, the dynamic impact field is calculated, local disturbances are quantified in real time, and suspected events in specific geographic areas are located based on water balance and dynamic impact field source error.
It improves the accuracy of error tracing, enabling water imbalance to be directly mapped to specific geographic areas, achieving efficient and accurate management decision support, and reducing model complexity and safety risks.
Smart Images
Figure CN121920236A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the fields of data processing and hydrological scheduling technology, and in particular to a method and system for hydrological extrapolation and error tracing under the influence of multiple factors. Background Technology
[0002] Traditional hydrological simulations (water balance calculations) for hydropower stations typically treat the reservoir inflow section as a whole, employing lumped models or simple partitioned models. When abnormal errors occur in the simulation results, only the deviation in the "inflow" component can be identified.
[0003] Existing technologies lack a mechanism to effectively correlate and trace macroscopic water volume errors with microscopic geospatial events, resulting in error tracing remaining at the factor level and failing to pinpoint spatial locations, thus making management decisions less targeted.
[0004] Existing error analysis methods are typically limited to calibrating input data (such as the accuracy of rainfall data) or overall model parameters. Specifically, existing technologies have the following two main limitations: In terms of spatial dimension, the ability to trace the source of errors is severely inadequate. Traditional methods cannot pinpoint the specific geographic location within the watershed where the error originates. For example, when simulation results show that the inflow is systematically overestimated, current technology can only indicate that "there is a deviation in the inflow simulation process," but it cannot further determine whether this deviation is due to unmonitored heavy rainfall on a tributary, unrecorded water intake activities in a certain area, or inaccurate description of the topography and soil properties of a specific sub-region by the model. This "black box" error attribution makes it difficult for operation and management personnel to take precise and efficient countermeasures, often resulting in only global, empirical adjustments to model parameters, which are inefficient and lack specificity.
[0005] In terms of error identification, there is a lack of effective differentiation mechanisms, especially under conditions where hydropower stations participate in peak shaving and flood discharge, generating severe unsteady flow, the extrapolation error will increase significantly. Existing technologies usually attribute all errors to model imperfections and attempt to find solutions by complicating the hydraulic model. This not only leads to increasingly complex models and difficulties in calibration, but more seriously, it masks potential operational compliance and equipment control accuracy issues in actual production operations, creating hidden safety hazards. Summary of the Invention
[0006] This application provides a method and system for hydrological extrapolation and error tracing under the influence of multiple factors. In the error tracing process, the macroscopic water imbalance can be directly mapped and located to one or more specific geographic spatial areas and suspected events occurring in those areas, thereby improving the accuracy of the tracing.
[0007] This application proposes a method for hydrological extrapolation and error tracing under the influence of multiple factors, including: Using the inlet section of the target reservoir as the outlet, a digital elevation model (DEM) is created for the catchment area above it, and a computational grid is defined. A static attribute vector is established for each computing grid, and the static attribute vector includes parameters describing the association between the corresponding computing grid and hydrological scheduling; Acquire multi-source data streams and map the acquired multi-source data streams to the partitioned computational grid; In each hydrological extrapolation calculation step, the overall water balance and dynamic influence field of the reservoir are calculated. The dynamic influence field is calculated by calculating the dynamic influence weight of each calculation grid in the calculation step, and the calculation grids with dynamic influence weights exceeding a preset threshold are aggregated into an influence field for spatial correction based on the influence field. In any hydrological extrapolation calculation step, spatial source tracing is performed if the error obtained from the extrapolation continues to exceed the threshold, so as to determine the abnormal grid based on water balance and dynamic influence field.
[0008] This application proposes a hydrological extrapolation and error tracing system under the influence of multiple factors, including a processor and a memory. The memory stores a computer program, and when the computer program is executed by the processor, it implements the steps of the hydrological extrapolation and error tracing method under the influence of multiple factors as described above.
[0009] The method in this application constructs a "gridized influence field" that is strictly bound to the real geographic space and whose weights can be dynamically responded to as a parallel analysis. In the simulation, it acts as a spatial sensor to quantify the disturbances in various localities within the watershed in real time. In the process of error source tracing, it can directly map and locate the macroscopic water imbalance to one or more specific geographic spatial regions and the suspected events occurring in those regions, thereby improving the accuracy of source tracing.
[0010] The above description is merely an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, specific embodiments of the present invention are described below. Attached Figure Description
[0011] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 This is a schematic diagram of the basic process of the hydrological extrapolation and error tracing method under the influence of multiple factors in this embodiment. Detailed Implementation
[0012] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.
[0013] This application proposes a method for hydrological extrapolation and error tracing under the influence of multiple factors, such as... Figure 1 As shown, it includes the following steps: In step S101, taking the inlet section of the target reservoir as the outlet, a digital elevation model (DEM) is created for the catchment area upstream of it, and a computational grid is defined, for example, one of the defined computational grids is used as a hydrological response unit. In some embodiments, step S101 specifically includes calculating the topographic index of each raster cell within the catchment area based on the DEM. The topographic index includes normalized runoff accumulation, slope, and humidity index, wherein the runoff accumulation represents the upstream catchment area, and the topographic humidity index TWI is... ,in For unit width bus area, For the slope angle, in a specific example, basic weights can be assigned to the terrain indices to configure the magnitude of each index's impact on water level management.
[0014] All pixels within the catchment area are clustered unsupervised based on their attributes in the multidimensional feature space to create a computational grid.
[0015] In step S102, a static attribute vector is established for each computational grid. This static attribute vector includes parameters describing the correlation between the corresponding computational grid and hydrological scheduling. In a specific example, the static attributes originate from the topographic indices of the computational grid, such as grid slope and humidity index, and are the feature vectors of the aforementioned clustering. The static attribute vector is used to characterize the inherent properties of the computational grid. The hydrological response units generated through clustering have irregular geometric shapes, and their boundaries naturally conform to the transition zones of topography, land use, and soil. The unit size may form larger grids in flat and homogeneous areas, while smaller grids are automatically formed in areas with complex topography and fragmented land use, to adapt to the requirements of considering topographic trends in hydrological scheduling.
[0016] In step S103, multi-source data streams are acquired and mapped to the divided computing grids. This step maps the changing data streams to each computing grid in the watershed to facilitate hydrological extrapolation and simulation.
[0017] In step S104, in each hydrological extrapolation calculation step, the overall water balance and dynamic influence field of the reservoir are calculated. The dynamic influence field is calculated by assigning dynamic influence weights to each calculation grid in the calculation step and aggregating calculation grids with dynamic influence weights exceeding a preset threshold into an influence field for spatial correction.
[0018] In step S105, at any hydrological extrapolation calculation step, spatial source tracing is performed if the error obtained from the extrapolation continues to exceed a threshold, in order to determine the abnormal grid based on water balance and dynamic influence field. In a specific example, hydrological extrapolation can be performed through hydrodynamic simulation or other methods.
[0019] In some embodiments, mapping the acquired multi-source data streams to a partitioned computational grid includes: Based on the humidity index of the terrain index, the humidity index is classified to determine the water retention probability. The water retention probability is then mapped to the corresponding computational grid. In a specific example, the humidity index may vary over time. The humidity index of some computational grids can be adjusted according to different seasons. By further converting it into the water retention probability, the impact of the grid on the runoff can be intuitively shown.
[0020] The acquired rainfall data is converted into rainfall intensity sequences for each computational grid using spatial interpolation. Based on the monitoring logs, determine the scheduling events and their scope of influence, and convert the monitoring logs into a sequence of events for the computing grid within the scope of influence. In a specific example, the monitoring logs may include things like the opening of gates and the guide vanes of the generator units.
[0021] In some embodiments, calculating the overall water balance and dynamic influence field of the reservoir in each hydrological simulation calculation step includes: The event source grid is determined based on rainfall data or monitoring logs. The event source grid is the central grid corresponding to the core impact area of the event. Specifically, the rainfall center area can be determined based on rainfall data or scheduling events, and the gate opening area can be used as the central grid. From this, a portion of the computational grid is selected as the central grid.
[0022] Based on the central grid and associated rainfall or scheduling data, the dynamic impact radius is determined. For rainfall, the dynamic impact radius is positively correlated with rainfall intensity and can be fitted based on historical data. For scheduling events, it can be directly derived from the expected impact range recorded in the event log.
[0023] Based on the dynamic influence radius, a search is performed in the neighborhood of the event source grid to obtain the dynamic computation grid, for example, to search for the base grid whose boundary intersects with or is inside the center grid.
[0024] The dynamic computational grid and the event source grid are updated and combined in each computation step to construct a dynamic influence field. The constructed dynamic influence field may include several computational grids, thereby allowing for the determination of a comprehensive terrain index based on the covered grids. Dynamic influence weights are used to quantify the degree of influence of each computational grid (HRU) on the inflow at time t. It can be represented as: in, , , For the coefficients of each part, () represents the rainfall impact function. For the event effect function, Due to changes in soil moisture, This is a humidity change adjustment item. for Each operation event; This represents the rainfall data for the grid. In a specific example, after the dynamic influence weights of the dynamic influence field decay to a certain proportion of their initial peak value, the dynamic influence field can be decoupled to update the dynamic influence field.
[0025] In some embodiments, the following abnormal mode discrimination process is also included: Under a set period, if the dynamic influence weights of multiple computational grids with continuous relationships are consistently greater than a set threshold during periods without rainfall and monitoring events, or if the rainfall-runoff response relationship in the corresponding area consistently deviates from the original trend and its contribution to rainfall consistently exceeds the expected runoff volume (e.g., consistently exceeding the expected runoff volume for a specified duration), then it is determined that a new confluence path exists.
[0026] Based on multiple interconnected computational grids, and combined with existing runoff paths within the pre-defined grids, the system determines the nearest or newly added runoff path and updates existing runoff paths. In other words, newly added runoff paths are either merged into existing paths or new runoff paths are created directly.
[0027] In some embodiments, the errors obtained from the deduction include reservoir capacity cumulative error, flow process error, and water balance closure error, wherein... The cumulative reservoir capacity error is the difference between the simulated reservoir capacity calculated in the hydrological extrapolation calculation step and the actual reservoir capacity determined by the measured reservoir water level.
[0028] The flow process error is the Nash efficiency coefficient between the calculated flow process curve at the inlet section and the inlet flow obtained through water balance inversion.
[0029] The water balance closure error is the difference between the change in reservoir water storage and the total inflow, outflow, and loss during the same period within a time period Δt. In a specific example, single-error triggering or multi-error combination triggering can be set as needed. For example, if the reservoir capacity accumulation error is triggered under a single condition, and it exceeds the threshold continuously for M time periods, then a grid anomaly is judged. Multi-error triggering can set combined conditions.
[0030] In some embodiments, determining anomalous grids based on water balance and dynamic influence fields when spatial source tracing is triggered includes: Set the maximum backtracking time, such as 72 hours, to determine the dynamic influence field through which each confluence path flows, as the expected influence field.
[0031] For the expected impact field, the expected contribution value of each computational grid is calculated for each time period of the backtracking duration. That is, hydrological extrapolation is performed separately for each time period within the local expected impact field to determine the expected contribution value. Furthermore, based on the maximum backtracking duration, the measured values of the error term are decomposed along the associated confluence path. The specific decomposition can be proportionally allocated according to the flow rate of the confluence path.
[0032] Determine the deviation between the expected contribution value and the measured value of the decomposition for each computational grid to identify anomalous grids.
[0033] In some embodiments, the method further includes: sorting the deviations between the expected contribution value and the measured value of the decomposed computation grid; The first k sorted computational grids are subjected to spatial distance and terrain index similarity analysis. Clustering is achieved through similarity analysis to trace the main factors causing the error. For example, the static attributes of each grid in the first k computational grids, the category proportion of terrain index, and spatial characteristics, such as the upstream of a tributary or a specific land type, can be analyzed.
[0034] The computational grids that belong to the error sources are determined based on the results of similarity analysis.
[0035] The method in this application constructs a "gridized influence field" that is strictly bound to the real geographic space and whose weights can be dynamically responded to as a parallel analysis. In the simulation, it acts as a spatial sensor to quantify the disturbances in various localities within the watershed in real time. In the process of error source tracing, it can directly map and locate the macroscopic water imbalance to one or more specific geographic spatial regions and the suspected events occurring in those regions, thereby improving the accuracy of source tracing.
[0036] This application also proposes a hydrological extrapolation and error tracing system under the influence of multiple factors, including a processor and a memory. The memory stores a computer program, and when the computer program is executed by the processor, it implements the steps of the hydrological extrapolation and error tracing method under the influence of multiple factors as described above.
[0037] Furthermore, although exemplary embodiments have been described herein, their scope includes any and all embodiments based on this disclosure that have equivalent elements, modifications, omissions, combinations (e.g., schemes involving intersections of various embodiments), adaptations, or changes. They are not limited to the examples described in this specification or during the implementation of this application, and such examples are to be construed as non-exclusive.
[0038] The above description is intended to be illustrative and not restrictive. For example, the above examples (or one or more of them) can be used in combination with each other. Other embodiments can be used by those skilled in the art when reading the above description.
[0039] The above embodiments are merely exemplary embodiments of this disclosure. Those skilled in the art can make various modifications or equivalent substitutions to this invention within the scope of the disclosure, and such modifications or equivalent substitutions should also be considered to fall within the protection scope of this invention.
Claims
1. A method for hydrological extrapolation and error tracing under the influence of multiple factors, characterized in that, include: Using the inlet section of the target reservoir as the outlet, a digital elevation model (DEM) is created for the catchment area above it, and a computational grid is defined. A static attribute vector is established for each computing grid, and the static attribute vector includes parameters describing the association between the corresponding computing grid and hydrological scheduling; Acquire multi-source data streams and map the acquired multi-source data streams to the partitioned computational grid; In each hydrological extrapolation calculation step, the overall water balance and dynamic influence field of the reservoir are calculated. The dynamic influence field is calculated by calculating the dynamic influence weight of each calculation grid in the calculation step, and the calculation grids with dynamic influence weights exceeding a preset threshold are aggregated into an influence field for spatial correction based on the influence field. In any hydrological extrapolation calculation step, spatial source tracing is performed if the error obtained from the extrapolation continues to exceed the threshold, so as to determine the abnormal grid based on water balance and dynamic influence field.
2. The hydrological extrapolation and error tracing method under the influence of multiple factors as described in claim 1, characterized in that, Using the inlet section of the target reservoir as the outlet, a digital elevation model (DEM) is created for the catchment area upstream of it, and a computational grid is defined, including: Based on the DEM, the topographic index of each raster cell within the catchment area is calculated. The topographic index includes normalized runoff accumulation, slope, and humidity index, where runoff accumulation represents the upstream catchment area, and the topographic humidity index (TWI) is... ,in For unit width bus area, The slope angle; All pixels within the catchment area are clustered unsupervised based on their attributes in the multidimensional feature space to create a computational grid.
3. The hydrological extrapolation and error tracing method under the influence of multiple factors as described in claim 2, characterized in that, Mapping the acquired multi-source data streams to the partitioned computational grid includes: Based on the humidity index of the terrain index, the humidity index is classified to determine the water retention probability according to the humidity classification, and the water retention probability is mapped to the corresponding computing grid. The acquired rainfall data is converted into rainfall intensity sequences for each computational grid using spatial interpolation. Based on the monitoring logs, determine the scheduling events and their scope of impact, and convert the monitoring logs into an event sequence for the computing grid within the scope of impact.
4. The hydrological extrapolation and error tracing method under the influence of multiple factors as described in claim 3, characterized in that, In each hydrological simulation step, the calculation of the overall water balance and dynamic influence field of the reservoir includes: The event source grid is determined based on rainfall data or monitoring logs. The event source grid is the central grid corresponding to the core impact area of the corresponding event. Based on the central grid and associated rainfall or scheduling data, the dynamic impact radius is determined; Based on the dynamic influence radius, a search is performed in the neighborhood of the event source grid to obtain the dynamic computation grid; The dynamic computational grid and the event source grid are updated and combined in each computation step to construct a dynamic influence field.
5. The hydrological extrapolation and error tracing method under the influence of multiple factors as described in claim 4, characterized in that, It also includes the following anomaly pattern detection process: If, within a set period, the dynamic influence weights of multiple computational grids with continuous relationships are consistently greater than a set threshold during periods without rainfall and monitoring events, or if the rainfall-runoff response relationship in the corresponding area consistently deviates from the original trend and its contribution to rainfall consistently exceeds the expected runoff volume, then it is determined that a new confluence path exists. Based on multiple computational grids with continuous relationships, and combined with existing runoff paths in the previously divided computational grids, the nearest or newly added runoff paths are determined, and existing runoff paths are updated.
6. The hydrological extrapolation and error tracing method under the influence of multiple factors as described in claim 4, characterized in that, The errors obtained from the simulation include cumulative reservoir capacity error, flow process error, and water balance closure error, among which, The cumulative reservoir capacity error is the difference between the simulated reservoir capacity calculated in the hydrological extrapolation calculation step and the actual reservoir capacity determined by the measured reservoir water level. The flow process error is the Nash efficiency coefficient between the calculated flow process curve at the inlet section and the inlet flow obtained through water balance inversion. The water balance closure error is the difference between the change in reservoir water storage and the total inflow, outflow and loss of water during the same period within the time period Δt.
7. The hydrological extrapolation and error tracing method under the influence of multiple factors as described in claim 6, characterized in that, In the event of triggered spatial source tracing, the anomaly grids identified based on water balance and dynamic influence fields include: Set the maximum backtracking time to determine the dynamic influence field that each confluence path passes through, as the expected influence field; For the expected impact field, calculate the expected contribution of each computing grid at each time period of the backtrack duration; and decompose the measured value of the error term in the associated confluence path based on the maximum backtrack duration. Determine the deviation between the expected contribution value and the measured value of the decomposition for each computational grid to identify anomalous grids.
8. The hydrological extrapolation and error tracing method under the influence of multiple factors as described in claim 7, characterized in that, Also includes: The deviations between the expected contribution value and the measured value of the decomposition for each computational grid are sorted. The first k sorted computational grids are used for spatial distance and terrain index similarity analysis. The computational grids that belong to the error sources are determined based on the results of similarity analysis.
9. A hydrological extrapolation and error tracing system under the influence of multiple factors, characterized in that, It includes a processor and a memory, wherein the memory stores a computer program, and when the computer program is executed by the processor, it implements the steps of the hydrological extrapolation and error tracing method under the influence of multiple factors as described in any one of claims 1 to 8.