A method for constructing and dynamically updating a digital twin of reservoir group flood control scheduling

Abstract

The application discloses a reservoir group flood control scheduling digital twin construction and dynamic updating method, relates to the technical field of water conservancy flood control scheduling digital twin, and comprises the following steps: obtaining upstream gate scheduling data, downstream channel monitoring data and channel station network reference data of a target object, performing event form boundary extraction, and generating cross-section reference data; extracting mutation time of water level and flow velocity from the cross-section reference data, calculating bivariate phase difference, and generating bireaction phase data containing the response phase interval of each cross-section. According to the scheme, the reachable window start and stop points of each event at each cross-section are calculated by comparing the adjacent cross-section response phase interval sequence, the candidate response is first constrained by the river channel propagation sequence and the event arrival range, the cross-event and cross-section misattribution is reduced, the bireaction candidate in the window is extracted, the phase connection interval is constructed, and the water level-flow velocity candidate continues to be constrained by the bivariate phase continuity of adjacent cross-sections.

Description

Technical Field

[0001] This invention relates to the field of digital twin technology for flood control scheduling in water conservancy, specifically a method for constructing and dynamically updating a digital twin of flood control scheduling for a reservoir group. Background Technology

[0002] With the development of basin flood control scheduling, joint operation of reservoir groups and digital twin water conservancy projects, the construction and dynamic updating of digital twins for flood control processes of reservoir groups usually requires the unified expression of upstream reservoir gate scheduling processes, downstream river or waterway monitoring processes and cross-sectional station network spatial benchmarks, and the formation of input data that can be called by hydrodynamic models, digital twin simulation modules or boundary condition interfaces before simulation starts or simulation is rolled out.

[0003] In existing technologies, methods for generating boundary conditions for digital twins or hydrodynamic simulations of flood control scheduling in reservoir groups typically begin by obtaining the outflow process line, gate opening process, or scheduling command time series from the upstream reservoir's scheduling system. Then, they obtain water level, velocity, or flow monitoring sequences from downstream hydrological stations. Subsequently, based on river distance, empirical propagation time, or pre-set boundaries using a hydrodynamic model, the upstream outflow process line is shifted to the downstream cross-sectional time axis. Alternatively, downstream single-station water level monitoring data is used to correct the boundary conditions. The advantages of these methods are their relatively direct implementation, their ability to utilize existing hydrological monitoring network and scheduling system data, and their suitability for scenarios where upstream outflow changes are relatively simple, downstream river propagation relationships are relatively stable, and monitoring cross-sectional responses are relatively synchronous. However, in the scenario of constructing and dynamically updating digital twins for flood control scheduling in reservoir groups, these methods suffer from propagation delays in the short-term continuous adjustment of multiple upstream gates, and the varying responses at different cross-sections of downstream curved waterways, as well as variations in water level and velocity. In cases of asynchronous operation, it is difficult to accurately determine which upstream flood discharge action and its corresponding propagation segment a response segment on a target boundary section belongs to. This is especially true when the response from the previous gate opening has not yet fully decayed and the subsequent gate adjustment has already generated new downstream disturbances. The above methods usually match responses based on fixed propagation time, empirical wave velocity, or the abrupt change of a single monitoring variable. This can easily lead to mistaking local responses that fall within the same temporal range but have discontinuous physical propagation relationships as valid boundary responses. For example, in a section of a curved channel, the water level rises first while the flow velocity changes with a lag. Adjacent downstream sections show different response sequences due to the river's curvature. If only a single time difference or fixed time delay is used to extract the boundary segment, responses generated under different gate adjustments may be mixed into the same boundary input. This can cause inconsistencies in the boundary conditions before the digital twin simulation starts regarding event attribution, section propagation sequence, and water level-flow velocity coordination. Summary of the Invention

[0004] The purpose of this invention is to provide a method for constructing and dynamically updating a digital twin of a reservoir group for flood control scheduling, so as to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the technical solution of the present invention is as follows:

[0006] In a first aspect, this invention discloses a method for constructing and dynamically updating a digital twin of flood control scheduling for a reservoir group, applied to the scenario of generating water level and flow velocity boundaries in the digital twin of flood control scheduling for a reservoir group, comprising the following steps:

[0007] Acquire upstream gate scheduling data, downstream waterway monitoring data, and waterway station network baseline data of the target object, extract event morphology boundaries, and generate cross-sectional baseline data;

[0008] Extract the abrupt change times of water level and flow velocity from the cross-sectional reference data and calculate the bivariate phase difference to generate bivariate phase data containing the response phase intervals of each cross-section.

[0009] Compare the sequential relationship of the response phase intervals of adjacent sections, and calculate the start and end points of the reachability window for each event in each section for the response phase intervals that meet the sequential relationship conditions, and generate event reachability window data.

[0010] Extract the portion of the dual-response phase data that falls into the event reachable window data as in-window dual-response candidates, and construct a phase acceptance interval based on the bivariate phase difference of adjacent sections, and then determine whether the bivariate phase difference of the in-window dual-response candidates falls into the phase acceptance interval.

[0011] Otherwise, calculate the phase residual between the bivariate phase difference and the phase transition interval, and remove in-window bi-response candidates whose phase residual exceeds a preset residual threshold, and then assemble the retained in-window bi-response candidates into piecewise variable time delay data.

[0012] The cross-sectional reference data is extracted and verified based on the segmented variable time delay data to generate water level and flow velocity boundary condition data.

[0013] Secondly, this invention discloses a system for constructing and dynamically updating a digital twin of a reservoir group for flood control scheduling, comprising:

[0014] The data acquisition module is used to acquire upstream gate scheduling data, downstream waterway monitoring data and waterway station network benchmark data of the target object, and extract event morphology boundaries to generate cross-sectional benchmark data.

[0015] The bivariate phase calculation module is used to extract the abrupt change time of water level and flow velocity from the cross-sectional reference data and calculate the bivariate phase difference to generate bivariate phase data containing the response phase interval of each cross-section.

[0016] The cross-section reachability window calculation module is used to compare the order of response phase intervals of adjacent cross-sections, and calculate the start and end points of the reachability window for each event in each cross-section for response phase intervals that meet the order of response conditions, and generate event reachability window data.

[0017] The candidate elimination module is used to extract the portion of the dual-response phase data that falls into the event reachable window data as dual-response candidates within the window, and construct a phase acceptance interval based on the bivariate phase difference of adjacent sections, and then determine whether the bivariate phase difference of the dual-response candidates within the window falls into the phase acceptance interval.

[0018] Otherwise, calculate the phase residual between the bivariate phase difference and the phase transition interval, and remove in-window bi-response candidates whose phase residual exceeds a preset residual threshold, and then assemble the retained in-window bi-response candidates into piecewise variable time delay data.

[0019] The water level and flow velocity boundary condition output module is used to extract and verify the cross-sectional reference data based on the segmented variable time delay data, and generate water level and flow velocity boundary condition data.

[0020] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0021] 1. This scheme compares the sequential relationship of response phase intervals between adjacent cross sections and calculates the start and end points of the reachable window for each event at each cross section. This ensures that candidate responses are first constrained by the river propagation sequence and the reach range of events, reducing misattribution across events and cross sections. It extracts dual-response candidates within the window and constructs phase continuity intervals, ensuring that water level-velocity candidates continue to be constrained by the dual-variable phase continuity of adjacent cross sections, eliminating pseudo-candidates that only fall within the time window but have discontinuous phases. It calculates phase residuals and removes candidates that exceed the threshold, making the retained candidates form traceable segmented variable time-delay data, improving the consistency and feasibility of subsequent boundary segment writing.

[0022] 2. This scheme extracts the river mileage of adjacent sections from the waterway station network benchmark data and proportionally converts the residual threshold, so that the threshold varies with the spatial spacing of the sections, avoiding misjudgment of response differences between long and short river sections using a fixed threshold; it judges whether the phase residual exceeds the threshold and removes the corresponding candidate, so that candidates with abnormal phase continuity within the window do not enter the time delay calculation, reducing cross-section pseudo-matches; it subtracts the event morphology boundary start point from the water level and flow velocity change time of the retained candidates to obtain bivariate time delays, so that the time delays are attributed to the same flood discharge event; it structurally associates the water level time delay and flow velocity time delay to form segmented variable time delay data that can be called for subsequent boundary writing. Attached Figure Description

[0023] The disclosure of this invention is illustrated with reference to the accompanying drawings. It should be understood that the drawings are for illustrative purposes only and are not intended to limit the scope of protection of this invention. In the drawings, the same reference numerals are used to refer to the same parts. Wherein:

[0024] Figure 1 A flowchart illustrating the steps of a method for constructing and dynamically updating a digital twin for flood control scheduling of a reservoir group, provided by this invention.

[0025] Figure 2 This is a schematic diagram of the process for generating dual-response phase data provided by the present invention;

[0026] Figure 3 This is a flowchart illustrating the process of generating event reachability window data provided by the present invention.

[0027] Figure 4 This is a schematic diagram of the process for generating piecewise variable time-delay data provided by the present invention;

[0028] Figure 5 This invention provides a schematic diagram of the module functions of a digital twin construction and dynamic update system for flood control scheduling of a reservoir group. Detailed Implementation

[0029] It is readily understood that, based on the technical solution of this invention, those skilled in the art can propose various interchangeable structural methods and implementations without altering the essential spirit of the invention. Therefore, the following detailed embodiments and accompanying drawings are merely illustrative examples of the technical solution of this invention and should not be considered as the entirety of the invention or as limitations or restrictions on the technical solution of this invention.

[0030] Existing methods for generating boundary conditions for digital twins or hydrodynamic simulations of reservoir group flood control scheduling typically rely on upstream outflow process lines, gate opening processes, or scheduling command time series, combined with downstream water level, velocity, or flow monitoring sequences. They match downstream cross-sectional boundary segments based on river distance, empirical propagation time, fixed time delays, or the moment of abrupt change in a single monitoring variable. This processing chain can generate boundary inputs when upstream outflow changes are relatively simple, downstream propagation relationships are relatively stable, and monitoring cross-sectional responses are relatively synchronous. However, when multiple upstream gates are continuously adjusted in a short period, and when there are propagation delays at different cross-sections of the downstream curved channel, and water level and velocity changes are not synchronized, existing methods struggle to determine which upstream flood discharge event and its corresponding propagation segment the response segment on the target boundary cross-section belongs to. This can easily lead to the misclassification of local responses that fall within the same temporal range but have inconsistent event attribution, cross-sectional propagation order, or water level-velocity coordination as valid boundary responses.

[0031] Before the digital twin simulation of flood control scheduling of the reservoir group was launched, the downstream disturbance caused by the previous gate opening of a certain upstream cascade reservoir had not yet completely decayed, and the subsequent gate opening adjustment had already generated new downstream disturbances. At this time, a target boundary section of a curved channel first showed a rise in water level and a lagging change in flow velocity, and adjacent downstream sections showed different response sequences due to the river's curvature. If the existing processing only extracts boundary segments according to fixed propagation time, empirical wave velocity, or the abrupt change of a single monitoring variable, it may mix the local responses generated by the two gate adjustments into the same boundary input, causing inconsistencies in event attribution, cross-section propagation sequence, and water level-flow velocity correspondence in this boundary segment.

[0032] If the above issues are not addressed, the water level and velocity boundary condition data generated by the reservoir group flood control scheduling digital twin before simulation initiation or rolling updates will propagate to subsequent hydrodynamic models or digital twin simulation modules along a path of unclear upstream event attribution—downstream section response matching deviation—mismatch in the coordination relationship between water level and velocity segments. This will cause response segments that do not belong to the same flood discharge action or do not conform to the section propagation sequence to be mixed into the target boundary section input data. As a result, the boundary conditions will shift in terms of temporal attribution, spatial propagation, and bivariate response relationship, and the input data called by subsequent simulations will not stably correspond to the water level and velocity response chain in the actual flood control scheduling process.

[0033] After introducing the basic concept of the present invention, the embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0034] Example 1:

[0035] This embodiment provides a method for constructing and dynamically updating a digital twin of flood control scheduling for a reservoir group. The method can be executed by a data processing system for the digital twin of flood control scheduling for a reservoir group. Physically, the data processing system includes a scheduling data access unit communicating with the upstream reservoir gate monitoring system, a monitoring data access unit communicating with the downstream waterway hydrological monitoring station network, a spatial reference library storing reference data for the waterway station network, and a computing server carrying the digital twin simulation module. The method is applied to the water level and velocity boundary generation scenario of the digital twin of flood control scheduling for a reservoir group. Specifically, before the digital twin simulation starts or before the simulation is rolled over, water level and velocity boundary condition data that can be called by the hydrodynamic model, the digital twin simulation module, or the boundary condition interface is generated for the target boundary section. This water level and velocity boundary condition data is then used as the dynamic boundary input for the target boundary section for the construction and dynamic updating of the digital twin of flood control scheduling for the reservoir group.

[0036] For ease of understanding, the following description uses a cascade reservoir group located upstream of a curved waterway, with its outflow controlled by multiple floodgates, as an example of the operational scenario. However, this does not constitute a limitation on the application scenario. The method is also applicable to watershed flood control scheduling scenarios with multiple gates operating jointly, multiple monitoring sections downstream, and propagation delays between sections due to the river's curvature, reservoir group joint flood control scheduling scenarios, and waterway navigation level and velocity assurance scenarios. The target object refers to a section of upstream and downstream river system included in this boundary generation process, with the upstream side being the reservoir outflow controlled by gate operation, and the downstream side being a curved waterway section with monitoring sections for water level and velocity.

[0037] In this embodiment, the upstream gate scheduling data refers to the time series data collected by the upstream reservoir gate monitoring system that reflects the gate opening and closing and reservoir discharge process, including at least the gate number, gate opening record and its collection time, and reservoir outflow record and its collection time.

[0038] Downstream channel monitoring data refers to time-series data collected by the downstream channel hydrological monitoring station network that reflects the water flow response process at each monitoring section, including at least the station number, water level record and its collection time, and flow velocity record and its collection time.

[0039] The reference data for the waterway station network refers to the reference data describing the spatial layout of the monitoring station network. It includes at least the station number, the river mileage, and the cross-section number. The river mileage refers to the distance of each monitoring cross-section along the middle channel line from the unified mileage reference point, and the cross-section number refers to the unique spatial identifier of each monitoring cross-section.

[0040] The overall processing flow of this embodiment is constructed by progressively converging a causal chain from upstream flood discharge to downstream section response. Its logical framework includes six interconnected processing steps, such as... Figure 1 As shown.

[0041] It should be noted that the overall processing flow is as follows:

[0042] First, upstream gate scheduling data, downstream channel monitoring data, and channel station network baseline data of the target object are acquired, and event morphology boundaries are extracted. The multi-source heterogeneous data is normalized to a unified event processing timeline and organized by cross-section. Simultaneously, the morphological range of each flood discharge effect on the timeline is defined, generating cross-sectional baseline data. Second, the abrupt change times of water level and flow velocity at each cross-section are extracted from the cross-sectional baseline data, and the bivariate phase difference between the two is calculated, generating bivariate phase data containing the response phase intervals of each cross-section. This upgrades the downstream response from isolated univariate abrupt change moments to response judgment units co-located with water level and flow velocity. Third, the sequential relationship of response phase intervals of adjacent cross-sections is compared along the river mileage. Only responses conforming to the upstream-downstream propagation sequence are recognized, and the reachability of each event at each cross-section is calculated based on the event morphology boundaries. The process involves several steps: First, the start and end points of the window are used to generate event reachable window data, forming absolute time convergence for candidate responses. Second, the portion of the dual-response phase data that falls within the event reachable window data is extracted as in-window dual-response candidates, and a phase continuity interval is constructed using the bivariate phase difference between adjacent sections. The relative phase continuity of the in-window dual-response candidates is then determined. Third, the phase residuals of the in-window dual-response candidates that do not fall within the phase continuity interval are calculated. Pseudo-delay candidates with phase residuals exceeding the residual threshold are removed, and the remaining in-window dual-response candidates are assembled into piecewise variable time-delay data. Finally, the cross-section reference data is time-delayed shifted and truncated based on the piecewise variable time-delay data to generate water level and flow velocity boundary condition data. This water level and flow velocity boundary condition data is then used as the dynamic boundary input for the target boundary section for the construction and dynamic updating of the digital twin.

[0043] Through the above technical solution, this embodiment uses the event pattern boundary derived from the upstream opening and outflow in the same direction as the internal anchor point, applies dual gating of absolute time and relative phase to the candidate response along the river mileage, and sends back the extracted and verified boundary fragments to drive the simulation module. Unlike the conventional processing method of shifting and extracting according to fixed propagation time or single variable mutation time, this embodiment constrains the event attribution of the response and the propagation order of adjacent sections, thereby producing water level and flow velocity boundary condition data that are consistent in event attribution, section order and water level-flow velocity coordination relationship. This data is used as the dynamic boundary input of the target boundary section and is directly called in the subsequent construction and rolling update of the digital twin simulation module.

[0044] This step follows the first step of the overall process, integrating upstream gate scheduling data, downstream channel monitoring data, and channel station network benchmark data in a unified time and spatial dimension. During the integration process, the form and range of each flood discharge action are defined, and cross-sectional benchmark data are output for all subsequent steps to use.

[0045] The upstream gate scheduling data and downstream waterway monitoring data are arranged in chronological order to generate an event processing timeline; the downstream waterway monitoring data are matched to the corresponding cross-sections based on the waterway station network benchmark data.

[0046] It should be noted that the event processing timeline refers to a unified time scale formed by merging and deduplicating the collection times of upstream gate scheduling data and downstream channel monitoring data in chronological order. The times recorded on this timeline are denoted as follows: The unit is seconds. The process of generating the event processing timeline is as follows: extract the gate opening record collection time and outflow record collection time from the upstream gate scheduling data, and the water level record collection time and flow velocity record collection time from the downstream channel monitoring data; arrange the above collection times in chronological order so that the time points covering the gate action change times and corresponding to the downstream monitoring sampling times together form the event processing timeline; for cases where only some data sources have samples at a certain time, the values ​​of the remaining data sources at that time are kept according to the principle of time proximity, based on the most recent valid sample, thereby ensuring that each moment on the event processing timeline has comparable upstream and downstream values.

[0047] To further explain, the process of matching downstream channel monitoring data to the corresponding cross-sections is as follows: read the station number, river mileage, and cross-section number from the channel station network baseline data, and construct a mapping relationship from station number to cross-section number; according to this mapping relationship, assign the water level records and flow velocity records identified by the station number in the downstream channel monitoring data to the corresponding cross-section number, and mark the upstream and downstream order of each cross-section in space with the river mileage, so that the downstream channel monitoring data is transformed from being organized by station to being organized by cross-section.

[0048] It should be noted that the river mileage in this step is used both to mark the spatial order of the cross-sections and as a spatial metric for comparing the phase intervals of responses of adjacent cross-sections along the river mileage, as well as for deriving the residual threshold. The downstream channel monitoring data is organized by cross-section number rather than by station number, so that the water level record and flow velocity record on the same cross-section can be paired. This provides a cross-section-based data organization basis for extracting the water level change time and flow velocity change time of the same cross-section and calculating the bivariate phase difference.

[0049] Calculate the changes in gate opening and outflow in the upstream gate scheduling data on the event processing time axis, and extract the start and end points of the continuous time intervals in which the direction of change in gate opening is consistent with the direction of change in outflow as the event form boundary.

[0050] It should be noted that the event form boundary refers to the time interval that defines the scope of a single flood discharge action on the event handling timeline. It is composed of the start point and the end point of the event form boundary, denoted as [missing information]. and All units are seconds; the process of extracting the event pattern boundary involves multiple steps with independent inputs and outputs, so it is explained in terms of sub-steps.

[0051] It should be noted that the process of extracting the event pattern boundary is as follows:

[0052] Sub-step 2.1: Under the same gate object, read the gate opening records at adjacent event processing timelines, calculate the gate opening difference between the next time and the previous time, and obtain the gate opening change. The input is the gate opening records arranged according to the event processing time axis under the same gate number. The output is the gate opening change and its direction at each adjacent time. The direction of change is determined by the sign of the gate opening change. A positive sign indicates that the opening increases, a negative sign indicates that the opening decreases, and zero indicates that the opening remains unchanged.

[0053] Sub-step 2.2: Under the same reservoir object, read the outflow records at adjacent event processing timelines, calculate the difference between the outflow at the next time point and the previous time point, and obtain the change in outflow. The unit is cubic meters per second; the input is the outflow records of the same reservoir outlet arranged according to the event processing time axis, and the output is the change in outflow at each adjacent time point and the direction of change.

[0054] Sub-step 2.3: Based on whether the direction of change of gate opening output from sub-step 2.1 is consistent with the direction of change of outflow output from sub-step 2.2 at the same time, merge time intervals with consistent directions and continuous time into one continuous time interval; the input is the direction of change of gate opening and the direction of change of outflow at each time, and the output is the continuous time interval with consistent directions. Consistent directions mean that at that time, the change of gate opening and the change of outflow are both positive or both negative, i.e., when the gate is open wider, the outflow increases; when the gate is closed smaller, the outflow decreases, and both point to the same flood discharge action.

[0055] Sub-step 2.4: Take the starting time of the continuous time interval obtained in sub-step 2.3 as the starting point of the event pattern boundary. The end time of this continuous time interval is taken as the endpoint of the event pattern boundary. The event pattern boundary is assigned an event number; the input is a continuous time interval with the same direction, and the output is the event pattern boundary identified by the event number. When there are multiple flood discharges, multiple non-overlapping event pattern boundaries are obtained in the above manner, each corresponding to a different flood discharge event.

[0056] For example, the calculation expressions for the change in gate opening and the change in outflow are as follows:

[0057] ;

[0058] in, Indicates the first The change in gate opening at each event processing timeline. Indicates the first Changes in outbound flow rate at each event processing timeline (unit: cubic meters per second). Indicates the first Gate opening records at each event processing timeline. Indicates the first The gate opening record at each event processing timeline. Indicates the first Outbound flow records at each event processing timeline (unit: cubic meters per second). Indicates the first Outbound flow records at each event processing timeline (unit: cubic meters per second).

[0059] Additionally, regarding the handling of situations where the direction of the gate opening change is inconsistent with the direction of the outflow change: when the gate opening change is not zero at a certain moment while the outflow change is zero or in the opposite direction, that moment is not included in any continuous time interval with the same direction, that is, it does not participate in the formation of the event pattern boundary; this is because the inconsistency between the opening change and the outflow change usually corresponds to measurement noise, gate fine-tuning not causing effective discharge, or transient states of asynchrony between upstream and downstream, excluding them can avoid misclassifying non-flood discharge effects as flood discharge events.

[0060] The downstream channel monitoring data, upstream gate scheduling data, and event pattern boundaries matched to the corresponding cross section are integrated to generate cross section benchmark data.

[0061] It should be noted that cross-sectional reference data refers to a unified data set indexed by the event processing timeline and the cross-section number, and marked with event morphology boundaries. The integration process is as follows: the water level and flow velocity records of the downstream channel monitoring data organized by cross-section number are written into the corresponding time of the event processing timeline; the gate opening records and outflow records of the upstream gate scheduling data are written into the corresponding time of the event processing timeline; and the event morphology boundaries identified by the event number are appended as time markers to the event processing timeline. This ensures that each cross-section in the cross-sectional reference data has water level and flow velocity records at every time, and that the flood discharge event to which it belongs and the corresponding start and end points of the event morphology boundaries can be found at each time.

[0062] Through the above technical solution, this embodiment normalizes multi-source heterogeneous upstream gate scheduling data and downstream channel monitoring data to the same event processing time axis and reorganizes them according to cross-section number. It defines the event form boundary by continuous intervals in the same direction as the gate opening change and the outflow change. Unlike the conventional method of processing data separately according to the timestamp of each data source and using downstream single variable mutation as the event starting point, this embodiment simultaneously completes spatiotemporal normalization and flood discharge action form reconstruction at both the data organization and event definition levels. This produces cross-section benchmark data marked with event form boundaries. This cross-section benchmark data is used as a unified data source for extracting downstream response mutation time and calculating the reachability window starting point in subsequent stages.

[0063] This step takes the cross-sectional baseline data and, for each flood discharge event, extracts the abrupt change times of water level and flow velocity at each downstream cross-section. It then calculates the signed time difference between the two on the event processing time axis and outputs dual-response phase data containing the response phase intervals of each cross-section. This elevates the downstream response from isolated, univariate abrupt change moments to a response determination unit co-located with water level and flow velocity. The processing procedure is as follows: Figure 2 As shown.

[0064] The search scope for downstream response is constructed by taking the starting point of the event pattern boundary as the search starting point and the end time of the change in outbound flow as the search ending point.

[0065] It should be noted that the downstream response search range refers to the time range within which the downstream response abrupt change time is defined on the event processing timeline for a specific flood discharge event. The search starting point is taken from the event pattern boundary of the flood discharge event. The search endpoint is taken from the end time of the change in outflow corresponding to the flood discharge event. The basis for constructing the downstream response search range is that the response of the downstream section cannot be earlier than the start of the upstream flood discharge in terms of causality, so the starting point of the event form boundary is taken as the search starting point; and the effective duration of the downstream response should not exceed the extension range of the change in outflow caused by this flood discharge, so the end time of the change in outflow is taken as the search endpoint.

[0066] To further clarify, the rule for determining the end time of the change in outbound flow is: when the end time of the change in outbound flow is later than the endpoint of the event pattern boundary. When the end time of the change in outbound flow is used as the search endpoint, the downstream response corresponding to the continued change in outbound flow in the same direction after the gate opening is completed is included in the retrieval; when the end time of the change in outbound flow is not later than the event mode boundary endpoint. At that time, the endpoint of the event form boundary This serves as the search endpoint. This ensures that the downstream response search scope neither misses ongoing responses nor oversteps its boundaries to include responses caused by the next flood discharge.

[0067] Within the downstream response search range, the water level recording time and flow velocity recording time that meet the hydrodynamic change conditions are extracted from the cross-sectional reference data according to the cross-section, and used as the water level change time and flow velocity change time.

[0068] It should be noted that the time of sudden water level change refers to the starting moment when a significant change in the water level record at a certain cross-section occurs in the same direction as the current flood discharge, denoted as . The velocity change time refers to the starting moment when a significant change in the velocity record on the same cross-section occurs in the same direction as the current flood discharge, denoted as . Both are in seconds. Hydrodynamic abrupt change conditions refer to the criteria used to identify the effective response front from noise fluctuations. These conditions include two requirements: consistent direction of change and cumulative change not falling below the abrupt change threshold. The abrupt change threshold is the lower limit of the cumulative change in water level or flow velocity used to distinguish the effective response front from the measured noise. The process of extracting the water level abrupt change time and the flow velocity abrupt change time by cross-section is as follows:

[0069] Sub-step 3.1: Extract the water level and flow velocity records of a certain section within the downstream response search range from the cross-sectional baseline data, calculate the water level difference and flow velocity difference at adjacent event processing timelines of the same section, and obtain the water level change and flow velocity change; the input is the water level and flow velocity records of this section within the downstream response search range, and the output is the water level change and flow velocity change at each adjacent time and their direction of change.

[0070] Sub-step 3.2: Compare the direction of water level change output from sub-step 3.1 with the direction of outflow change under the same flood discharge event, and determine whether the cumulative water level change within the continuous change interval with consistent direction is not lower than the abrupt change threshold. Extract the starting time of the first water level change interval that simultaneously satisfies both consistent direction and the threshold of cumulative change as the water level abrupt change time of that section. The inputs are the direction of water level change at each time point, the cumulative water level change, and the direction of outflow change during the event. The output is the time of water level change at the cross-section. "Same direction" means that the water level change and the outflow change are both positive or both negative, i.e., the water level rises when the gate is opened to increase discharge and falls when the gate is closed to reduce discharge.

[0071] Sub-step 3.3: Compare the direction of the velocity change output from sub-step 3.1 with the direction of the outflow change under the same flood discharge event, and determine whether the cumulative velocity change within the continuous interval with consistent direction is not lower than the abrupt change threshold. Extract the starting time of the first velocity change interval that simultaneously satisfies both consistent direction and the threshold of cumulative change as the velocity abrupt change time of that section. The inputs are the direction of the velocity change at each moment in the cross-section, the cumulative velocity change, and the direction of the outflow change during the event. The output is the velocity change time at the cross-section. The starting point of the first change interval with consistent direction and cumulative compliance is chosen instead of the moment of maximum change amplitude. This is to capture the arrival time of the response front, rather than the moment the response peak occurs, so that subsequent phase relationships reflect the order of propagation rather than the strength of the response.

[0072] It should be further explained that, due to measurement noise and minor fluctuations in water level and flow velocity records during monitoring, abrupt change thresholds are set to filter out noise. Because water level and flow velocity responses have different dimensions, abrupt change thresholds are set separately for water level and flow velocity. The water level threshold ranges from 0.02 meters to 0.10 meters, with a typical value of 0.05 meters; the flow velocity threshold ranges from 0.05 meters per second to 0.20 meters per second, with a typical value of 0.10 meters per second. Specific values ​​are determined based on the measurement accuracy of the monitoring section and the flow characteristics of the river section. Changes with cumulative changes below the corresponding abrupt change threshold are not identified as water level or flow velocity abrupt changes. The statistical interval for cumulative changes is defined by a continuous change interval in the same direction, measured from its starting point to the point where the direction changes.

[0073] The difference between the water level change time and the flow velocity change time belonging to the same cross section is calculated on the event processing time axis to obtain the bivariate phase difference; the phase direction is determined according to the order of occurrence of the water level change time and the flow velocity change time, and the response phase interval is formed by combining the bivariate phase difference, thereby generating dual response phase data.

[0074] It should be noted that the bivariate phase difference refers to the signed time difference between the time of abrupt change in water level and the time of abrupt change in flow velocity on the same cross-section along the event processing time axis, denoted as . The unit is seconds, used to characterize the relative order of arrival of the water level response front and the velocity response front on the cross section; the phase direction refers to the symbol determined by the order of the water level change time and the velocity change time, used to distinguish whether water level precedes or velocity precedes; the response phase interval refers to the time interval defined by the water level change time and the velocity change time on the cross section, with the starting point being the earlier of the two change times and the ending point being the later of the two change times, used to characterize the time span occupied by the response front of the cross section on the time axis.

[0075] For example, the expression for calculating the bivariate phase difference is:

[0076] ;

[0077] in, This represents the phase difference between two variables (unit: seconds). This indicates the time of sudden change in water level at this cross-section (unit: seconds). This indicates the time of abrupt change in flow velocity at this cross-section (unit: seconds).

[0078] when When the value is >0, the phase direction is marked as water level first, indicating that the water level front arrives at the cross section before the velocity front.

[0079] when When the phase is less than 0, the phase direction is indicated as velocity-first;

[0080] when When the value is 0, it indicates that the water level and flow velocity are synchronized. The output of this bivariate phase difference serves two purposes: firstly, as a phase marker for the cross-section when generating the response phase interval, and secondly, as a value for comparing the phase continuity of adjacent cross-sections when subsequently constructing the phase continuity interval.

[0081] To further explain, the process of generating dual-response phase data is as follows: the water level abrupt change times belonging to the same cross-section number under the same flood discharge event are... With the time of sudden change in flow rate Pairing; calculating the bivariate phase difference from the pairing results. The phase direction is determined; the earlier of the two is used as the starting point and the later one as the ending point to form the response phase interval of the cross section; the event number, cross section number, water level change time, flow velocity change time, bivariate phase difference and phase direction are combined to form a record of the cross section, and the records of each cross section are collected to obtain the bivariate phase data.

[0082] It should be noted that the water level and flow velocity records required for the water level change time and flow velocity change time are all taken from the cross-sectional reference data marked by the event shape boundary and organized by cross-section number. Therefore, the water level and flow velocity records of the same cross-section are naturally aligned with the same event processing time axis, and no secondary time alignment is required. The starting point of the event shape boundary serves as the search starting point of the downstream response search range in this stage, and will also serve as the deduction reference for the water level time delay and flow velocity time delay in the future. This reference reuse ensures that the phase difference, reachability window and time delay are established on the same time origin, avoiding phase misalignment caused by reference drift.

[0083] It should be noted that the applicable premises for bivariate phase difference calculation are as follows: the physical meaning of bivariate phase difference is based on the premise that the water level front and velocity front of the same cross section under the same flood discharge event originate from the same upstream flood discharge action; when a cross section only detects the water level change time but not the velocity change time or only the velocity change time but not the water level change time within the downstream response search range, the cross section does not generate a complete response phase interval under this event. Only the single change time that has been detected is retained as part of the phase information of the cross section. When the reachable window is calculated later, it will be geometrically derived and completed according to the response phase interval of adjacent cross sections, instead of forcibly using the missing change time in the bivariate phase difference calculation.

[0084] Through the above technical solution, this embodiment defines the downstream response search range by using the event form boundary start time and the end time of the outflow change, extracts the water level change time and flow velocity change time that are in the same direction as the flood discharge and whose cumulative change reaches the standard according to the cross section, and calculates the signed bivariate phase difference between the two. Unlike the conventional method of characterizing the downstream response by the change time of a single monitoring variable, this embodiment transforms the judgment object from a single variable time to the phase relationship of water level and flow velocity coordination at the response characterization level, thereby producing dual response phase data carrying phase direction and response phase interval. This dual response phase data is subsequently used to construct both the event reachability window and the phase acceptance interval, and is called as the common phase source for the two gating.

[0085] This step receives dual-response phase data, compares the sequential relationship of response phase intervals between adjacent sections along the river channel, and only accepts responses that conform to the upstream-downstream propagation sequence. It also defines the absolute time range of the response reachable at each section for each flood discharge event based on event morphology boundaries, outputting event reachability window data. This constitutes the first layer of gating for candidate responses. The processing procedure is as follows: Figure 3 As shown.

[0086] Following the order from upstream to downstream, read the starting point of the response phase interval of the adjacent upstream section and the adjacent downstream section respectively; determine whether the starting point of the response phase interval of the adjacent downstream section is earlier than the starting point of the response phase interval of the adjacent upstream section; otherwise, determine that the response phase interval of the corresponding adjacent downstream section satisfies the sequential relationship condition.

[0087] It should be noted that event reachability window data refers to the data set that defines the absolute time range of response reachability at each corresponding cross-section for each flood discharge event. It consists of the start point and end point of the reachability window, denoted as [missing information - likely a typo]. and All units are seconds; the sequence condition refers to the legal propagation timing constraint that the response of an adjacent downstream section must arrive before the response of an adjacent upstream section. The process of comparing the sequence of response phase intervals of adjacent sections is as follows:

[0088] Sub-step 5.1: Following the order of river mileage from upstream to downstream, read the starting points of the response phase intervals of adjacent upstream sections and adjacent downstream sections from the dual response phase data; the input is the response phase intervals of each section sorted by river mileage, and the output is the starting points of the two response phase intervals for adjacent sections. Adjacent upstream sections refer to sections with smaller river mileage located on the upstream side, and adjacent downstream sections refer to sections with larger river mileage located on the downstream side.

[0089] Sub-step 5.2: Determine whether the starting point of the response phase interval of the adjacent downstream section is earlier than the starting point of the response phase interval of the adjacent upstream section; the input is the two starting points of the response phase intervals output from sub-step 5.1, and the output is the determination result of whether the sequential relationship condition is met. When the starting point of the response phase interval of the adjacent downstream section is not earlier than the starting point of the response phase interval of the adjacent upstream section, it is determined that the response phase intervals of the adjacent downstream section meet the sequential relationship condition, and a phase sequence succession identifier is generated; when the starting point of the response phase interval of the adjacent downstream section is earlier than the starting point of the response phase interval of the adjacent upstream section, it is determined that the sequential relationship condition is not met, and the response phase interval of the adjacent downstream section does not participate in the formation of the reachability window of this event.

[0090] Additionally, regarding the handling of situations where the sequential relationship condition is not met: when the starting point of the response phase interval of an adjacent downstream section is earlier than the starting point of the response phase interval of an adjacent upstream section, it means that the response detected by the downstream section appeared earlier than the response of the upstream section in time. This violates the physical order of the flood discharge effect from upstream to downstream. Usually, what is detected by the downstream section is the residual response or other disturbances of the previous flood discharge effect that has not yet decayed. Therefore, it is judged as not meeting the sequential relationship condition and excluded from the formation of the reachability window of this event, thereby blocking the reverse-order response from entering the subsequent processing at the reachability window level.

[0091] The later start point in time between the start point of the response phase interval that satisfies the sequential relationship condition and the start point of the event form boundary is taken as the start point of the reachable window, and the end point of the response phase interval of the adjacent downstream section that satisfies the sequential relationship condition is determined as the end point of the reachable window; when the end point of the reachable window is not earlier than the start point of the reachable window, the start point and end point of the reachable window of the corresponding section are combined to generate event reachable window data.

[0092] It should be noted that the process of calculating the start and end points of the reachability window is as follows: For a target section with a phase sequence identifier, the start point of its response phase interval is compared with the start point of the event pattern boundary of the flood discharge event. In comparison, a later time on the event processing timeline is selected as the starting point of the reachability window. The endpoint of the phase interval of the target section response is taken as the endpoint of the reachability window. When the endpoint of the reachable window is not earlier than the start point of the reachable window, the start point and the endpoint of the reachable window, together with the event number and the section number, are recorded in the event reachable window data. When the endpoint of the reachable window is earlier than the start point of the reachable window, it is determined that there is no valid reachable window for the section under this event and it is not recorded.

[0093] For example, the expressions for calculating the start and end points of the reachable window are:

[0094] ;

[0095] in, Indicates the starting point of the reachable window (unit: seconds). Indicates the reachability window endpoint (unit: seconds). Indicates the starting point (in seconds) of the phase interval of the target section response that satisfies the sequential relationship condition. Indicates the end point of the phase interval of the response (unit: seconds). Indicates the starting point of the event mode boundary of the flood discharge event (unit: seconds). This indicates taking the larger value.

[0096] Choosing the latter of the response phase interval start point and the event pattern boundary start point as the reachability window start point ensures that the reachability window start point is neither earlier than the start of upstream flood discharge nor earlier than the emergence of the response front at this section, thus being simultaneously subject to upstream causal constraints and local section response constraints. The outputs of this reachability window start point and reachability window end point are subsequently used as absolute time boundaries for screening dual response candidates within the window.

[0097] It should also be noted that, for cases where the target section does not have its own complete response phase interval under this flood discharge event, the start and end points of the reachable window are determined by geometric derivation from adjacent sections. The process is as follows: when the target section does not have its own corresponding response phase interval, but its adjacent upstream and downstream sections both have response phase intervals and there is a phase sequence succession marker between them, the end point of the response phase interval of the adjacent upstream section is compared with the start point of the event form boundary of the flood discharge event, and the later one is selected as the start point of the reachable window of the target section; the start point of the response phase interval of the adjacent downstream section is selected as the end point of the reachable window of the target section; when the end point of the reachable window is not earlier than the start point of the reachable window, it is recorded in the event reachable window data.

[0098] Furthermore, the basis for this geometric derivation is that, under a valid propagation timeline, the response of the target section must occur after the response of its adjacent upstream section and before the response of its adjacent downstream section. Therefore, taking the end point of the phase interval of the response of the adjacent upstream section as the earliest time when the response of the target section is achievable and the beginning point of the phase interval of the response of the adjacent downstream section as the latest time when the response of the target section is achievable, even when the abrupt change time of the target section itself is missing, the response phase intervals of its upstream and downstream adjacent sections can still spatially squeeze out the achievable range of the section's response, thereby avoiding the interruption of the boundary generation of the entire propagation chain at that section due to the lack of monitoring of individual sections.

[0099] It should be noted that regarding the configuration boundary for calculating the event reachability window: the above comparison of river mileage and geometric derivation are based on the premise that each cross-section is monotonically arranged along the middle channel of the river and that there are no branching or backflow between adjacent cross-sections; when there are branching, confluence or significant backflow in the river section, making it impossible for the river mileage to uniquely represent the upstream and downstream propagation order, the upstream and downstream adjacency relationship between cross-sections is re-determined according to the mainstream direction of the water flow before comparison, or the measured mainstream propagation path is used instead of the river mileage as the sorting basis, so as to ensure that the determination of the sequential relationship conditions and the squeeze direction of the geometric derivation are consistent with the actual propagation direction.

[0100] Through the above technical solution, this embodiment applies a sequential constraint of downstream not earlier than upstream to the response phase interval of adjacent cross sections along the river mileage, and uses the latter of the response phase interval start point and event form boundary start point as the reachable window start point. Unlike the conventional processing method of uniformly shifting all downstream cross sections with a fixed time delay, this embodiment applies absolute time gating to candidate responses at the cross section propagation order level, thereby producing event reachable window data for each cross section and each event. This event reachable window data is subsequently used as the first absolute time screening boundary for dual response candidates within the window, and as the basis for verifying the legality of the leading-edge time during the final interception and verification.

[0101] This step takes over the event reachable window data and the dual-response phase data. First, it extracts the dual-response candidates within the window using absolute time gating. Then, it constructs the phase takeover interval based on the continuity of the phase difference between the two variables in adjacent sections. It applies relative phase gating to the dual-response candidates within the window and outputs the dual-response candidates within the window that have been retained by double gating.

[0102] Extract the portion of the two-response phase data that falls within the event reachable window as in-window two-response candidates.

[0103] It should be noted that in-window dual-response candidates refer to dual-response phase data records whose water level change time and flow velocity change time both fall between the start and end points of the corresponding reachable window. These are candidate responses retained after the first layer of absolute time gating. The process of extracting in-window dual-response candidates is as follows: compare the dual-response phase data and event reachable window data under the same flood discharge event and the same cross-section; when the water level change time of that cross-section... With the time of sudden change in flow rate None earlier than the start of the reachable window And all of them are no later than the reachable window endpoint. When the water level change time and flow velocity change time of a record are combined, they are extracted as candidates for in-window dual responses. If either the water level change time or the flow velocity change time falls outside the range between the start and end points of the reachable window, the record is discarded. This process achieves basic convergence of the time delay range in absolute time.

[0104] A phase transition interval is constructed based on the bivariate phase difference between adjacent sections, and then it is determined whether the bivariate phase difference of the candidate bivariate response within the window falls into the phase transition interval.

[0105] It should be noted that the phase transition interval refers to the range of values ​​defined by the bivariate phase difference between the adjacent upstream section and the adjacent downstream section, with the lower limit denoted as . The upper limit endpoint is denoted as The units are all seconds, used to constrain the phase continuity relationship that the current cross-sectional bivariate phase difference should be between the phase differences of adjacent upstream and downstream cross-sections. The process of constructing the phase continuity interval is as follows:

[0106] Sub-step 6.1: For the current cross section, extract the bivariate phase difference corresponding to the adjacent upstream cross section and the bivariate phase difference corresponding to the adjacent downstream cross section from the bivariate response phase data; the input is the bivariate phase difference of each cross section sorted by river mileage, and the output is the bivariate phase difference of the two adjacent upstream and downstream cross sections of the current cross section.

[0107] Sub-step 6.2: Compare the numerical values ​​of the bivariate phase difference between adjacent upstream sections and the bivariate phase difference between adjacent downstream sections, and determine the smaller value as the lower limit endpoint. The larger value is determined as the upper limit endpoint. The input is the two bivariate phase differences output from substep 6.1, and the output is the lower and upper limits of the phase transition interval. The lower limit is set by the smaller value and the upper limit by the larger value, so that the phase transition interval is not dependent on the prior assumption of which phase difference is larger between the upstream and downstream sections, but is determined solely by the envelope of the actual phase difference values.

[0108] Sub-step 6.3: Using the lower and upper limit endpoints as boundaries, construct the phase transition interval of the current section. Determine the bivariate phase difference corresponding to the two response candidates within the current cross-sectional window. Whether it falls within the phase acceptance interval; the input is the bivariate phase difference between the phase acceptance interval and the dual-response candidate within the current cross-sectional window, and the output is the determination result of whether it falls within the interval. If it falls within the interval, the phase acceptance is determined to be satisfied, and the dual-response candidate within the window is directly retained; if it does not fall within the interval, the calculation of the phase residual is performed.

[0109] It should be noted that the two endpoints of the phase transition interval are taken from the bivariate phase difference between adjacent upstream and downstream sections under the same flood discharge event. This bivariate phase difference is derived from the water level change time and flow velocity change time extracted based on the boundary of the same event pattern. Since the relative order of the water level front and flow velocity front of adjacent sections should change continuously with spatial position under legitimate propagation, the bivariate phase difference of the current section falls into the transition interval formed by the bivariate phase differences of adjacent upstream and downstream sections. This means that the response of this section is continuously connected with the response of the adjacent section in phase, thus forming a relative phase gating that is different from absolute time gating.

[0110] Otherwise, calculate the phase difference between the two variables and the phase residual of the phase transition interval, and remove in-window dual response candidates whose phase residual exceeds the preset residual threshold. Then, assemble the remaining in-window dual response candidates into piecewise variable time delay data.

[0111] It should be noted that the phase residual refers to the minimum time distance between the bivariate phase difference and the nearest endpoint of the phase continuation interval when the bivariate phase difference of the candidate bivariate response within the window does not fall into the phase continuation interval, denoted as . The unit is seconds, used to measure the degree to which the candidate response deviates from the phase continuity relationship between adjacent sections; the residual threshold refers to the upper limit of the allowable phase residual, denoted as . The unit is seconds, used to determine whether the deviation corresponding to the phase residual is within the acceptable tolerance range. The process of calculating the phase residual and assembling the piecewise variable time-delay data involves multiple steps with independent inputs and outputs, and is therefore described as sub-steps. The processing procedure is as follows: Figure 4 As shown.

[0112] It should be noted that the process of calculating the phase residual and assembling the piecewise variable time-delay data is as follows:

[0113] Sub-step 7.1: When the bivariate phase difference of the candidate two responses within the window Not falling into the phase transition zone At that time, the smaller of the bivariate phase difference and the distance between the two endpoints of the phase transition interval is calculated as the phase residual. The input is the candidate bivariate phase difference and the lower and upper limits of the phase transition interval, and the output is the phase residual.

[0114] For example, the expression for calculating the phase residual is:

[0115] ;

[0116] in, Represents the phase residual (unit: seconds). This represents the bivariate phase difference (in seconds) of the two-response candidates within the current cross-sectional window. Indicates the lower limit endpoint of the phase transition interval (unit: seconds). Indicates the upper limit of the phase transition interval (unit: seconds). This indicates taking the smaller value.

[0117] Sub-step 7.2: Extract the river mileage distance between the adjacent upstream and downstream sections of the current section from the waterway station network benchmark data. The residual threshold is generated by proportionally converting the river channel distance. The phase residual obtained in sub-step 7.1 With residual threshold The comparison process involves taking the phase residual and the residual threshold as inputs and outputting the decision on whether to remove a candidate. If the phase residual is greater than the residual threshold, the candidate with the dual response within the window is determined to not satisfy the phase continuity relationship between adjacent sections and is removed as a pseudo-delay candidate. If the phase residual is not greater than the residual threshold, the candidate with the dual response within the window is retained.

[0118] For example, the expression for calculating the residual threshold is:

[0119] ;

[0120] in, Represents the residual threshold (unit: seconds). This indicates the distance (in meters) between the current cross-section and the adjacent upstream and downstream cross-sections. This indicates the proportional conversion factor (unit: seconds per meter).

[0121] It should be noted that, since the larger the distance between adjacent cross sections, the greater the allowable fluctuation of the response phase difference with spatial variation, the residual threshold is set to increase proportionally with the river mileage distance between adjacent cross sections, with a proportional conversion factor. The value ranges from 0.005 seconds per meter to 0.05 seconds per meter, with a typical value of 0.02 seconds per meter. The specific value is determined based on the average propagation wave velocity of the river section and the density of the monitoring section; the higher the average wave velocity, the better. The smaller the value.

[0122] It should be noted that the residual threshold is derived from the distance between adjacent cross-sections, and the distance between the cross-sections is taken from the initial integrated waterway station network reference data. Since the residual threshold is dynamically generated with the cross-section spacing rather than being a uniform constant, a larger phase acceptance tolerance is allowed in river sections with sparse cross-sections, and a stricter phase acceptance tolerance is applied in river sections with dense cross-sections. This avoids under-removal or over-removal under different cross-section spacings with a single fixed tolerance. This linkage generates a phase continuity discrimination scale that matches the spatial structure of the river section.

[0123] Sub-step 7.3: For the in-window dual-response candidates that are directly retained in sub-step 6.3 or retained after comparison in sub-step 7.2, extract their corresponding water level change times. With the time of sudden change in flow rate Subtract the starting point of the event form boundary of the corresponding flood discharge event respectively. The water level time lag was obtained. With flow velocity time delay The input consists of the water level change time, flow velocity change time and corresponding event mode boundary start point of the reserved in-window dual response candidates. The output consists of the water level time delay and flow velocity time delay of the cross section under the event.

[0124] For example, the calculation expressions for water level time delay and flow velocity time delay are as follows:

[0125] ;

[0126] in, Indicates water level time delay (unit: seconds). Indicates flow velocity delay (unit: seconds). Indicates the water level change time (in seconds) for the retained dual-response candidates within the window. Indicates the time of abrupt change in flow velocity (unit: seconds). Indicates the starting point of the event pattern boundary of the corresponding flood discharge event (unit: seconds).

[0127] Water level time delay and flow velocity time delay respectively characterize the time taken for the flood discharge to propagate from its morphological origin to the cross section and cause water level response and flow velocity response. The outputs of the two constitute the values ​​of piecewise variable time delay data.

[0128] Sub-step 7.4: Delay water level With flow velocity time delay The data is structured and associated based on event number, river segment number, and cross-section number, and assembled into segmented variable time-delay data. The input consists of the water level and velocity time delays for each event at each cross-section, and the output is segmented variable time-delay data indexed by event number, river segment number, and cross-section number. The river segment number is a unique identifier for the river segment between two adjacent cross-sections, enabling the segmented variable time-delay data to represent the propagation time delays of different events in different river segments, rather than using a single time delay for the entire river channel.

[0129] Additionally, it is necessary to add the following applicable boundary conditions for phase residual calculation: the phase residual is based on the premise that the current section and its adjacent upstream and downstream sections have comparable bivariate phase differences under the same flood discharge event; when one of the adjacent sections lacks a bivariate phase difference under this event, making it impossible for the phase connection interval to be formed by two endpoints, then phase connection discrimination and phase residual calculation are not performed on the dual response candidates within the current section window, only the filtering results of the first absolute time gating of the event reachable window are retained, and the section is marked as having only passed absolute time gating in the segmented variable time delay data for differentiated treatment during subsequent interception and verification.

[0130] Through the above technical solution, this embodiment constructs a phase transition interval with the phase difference between adjacent cross sections as the endpoint, and removes pseudo-time delay candidates whose phase residual exceeds the limit by using a residual threshold converted according to the river mileage ratio for candidates that do not fall into the transition interval. Unlike the conventional processing method that only uses absolute time or fixed time delay to determine the validity of the response, this embodiment applies a second gating to candidates that have passed the absolute time gating at the level of relative phase continuity, thereby producing segmented variable time delay data after removing pseudo-overlapping across events. This segmented variable time delay data is subsequently used as the time delay source for time delay translation and truncation verification of cross section reference data.

[0131] This step receives segmented variable time-delay data, performs time-delay translation and truncation on the time series of the target boundary section in the cross-sectional reference data, and performs truncation verification on the writing side using event reachable window data and phase acceptance interval, outputting the final water level and flow velocity boundary condition data.

[0132] Read the water level time delay and flow velocity time delay of the corresponding section in the segmented variable time delay data; add the corresponding water level time delay and flow velocity time delay to the start and end points of the event pattern boundary of the corresponding event in the section reference data, respectively, to obtain the target response interception time range; within the target response interception time range, extract water level segments and flow velocity segments from the section reference data.

[0133] It should be noted that the target boundary section refers to the downstream monitoring section corresponding to the boundary condition data required by the digital twin simulation module; the water level segment refers to the sequence of water level records extracted from the cross-section reference data within the target response intercept time range; and the flow velocity segment refers to the sequence of flow velocity records extracted within the same range. The process of obtaining the target response intercept time range and extracting segments is as follows: read the water level time delay and flow velocity time delay of the target boundary section under the corresponding flood discharge event from the segmented variable time delay data; add the water level time delay to the start point and end point of the event form boundary of the flood discharge event respectively to obtain the intercept time range of the water level segment; add the flow velocity time delay to obtain the intercept time range of the flow velocity segment; the two together constitute the target response intercept time range; extract the water level record sequence and flow velocity record sequence of the target boundary section from the cross-section reference data within the target response intercept time range, and use them as the water level segment and flow velocity segment respectively.

[0134] By superimposing water level and flow velocity time delays at the start and end points of the event pattern boundary, respectively, the water level and flow velocity segments are located to the response time period of the target boundary section according to their actual propagation time delays, rather than simultaneously translating water level and flow velocity with the same time delay, thus preserving the phase difference between the water level and flow velocity responses. The output of this target response intercept time range is verified together with the event reachability window data and phase transition interval in subsequent intercept verification.

[0135] The water level and flow velocity segments are subjected to truncation verification, and the water level and flow velocity segments that pass the truncation verification are combined to generate water level and flow velocity boundary condition data.

[0136] It should be noted that the water level and velocity boundary condition data refers to the set of boundary condition data that is used by the digital twin simulation module, composed of time-series aligned and combined water level and velocity segments that have passed truncation verification. Truncation verification refers to the end-point verification of the legality of the leading edge time and the compliance of the phase before the segments are written into the boundary condition data. The process of performing truncation verification and generating water level and velocity boundary condition data is as follows:

[0137] Sub-step 8.1: Determine whether the water level change time corresponding to the water level segment and the flow velocity change time corresponding to the flow velocity segment both fall between the start and end points of the reachability window data of the target boundary section under the corresponding flood discharge event; the input is the change time corresponding to the water level segment and the flow velocity segment and the start and end points of the reachability window of the target boundary section, and the output is the determination result of whether the leading edge time falls within the reachability window.

[0138] Sub-step 8.2: When sub-step 8.1 determines that the leading edge time falls within the reachable window, further determine whether the bivariate phase difference corresponding to the water level segment and the flow velocity segment falls within the phase acceptance interval of the target boundary section, or whether the phase residual is not greater than the residual threshold; the input is the bivariate phase difference and phase acceptance interval corresponding to the segment, and the output is the determination result of whether the phase is qualified.

[0139] Sub-step 8.3: When the current edge time falls within the reachable window and the phase is qualified, and the segment has a corresponding water level time delay and flow velocity time delay source in the segmented variable time delay data, the interception verification is determined to be successful, and the water level segment and flow velocity segment are assigned an interception verification success flag; when any of the following three conditions are met: the current edge time does not fall within the reachable window, the phase is unqualified, or the time delay source is missing, the interception verification is determined to be unsuccessful, and the writing of the segment is blocked; the input is the aforementioned two judgment results and the existence of the time delay source, and the output is the judgment result of whether the interception verification is successful.

[0140] Sub-step 8.4: Extract the water level and flow velocity segments that have passed the interception verification. According to the time step of the target boundary section, pair the water level and flow velocity values ​​that exist in pairs at the same writing time according to the corresponding elevation benchmark requirements, and write them into the boundary water level sequence and boundary flow velocity sequence. Combine them to generate water level and flow velocity boundary condition data. The input is the water level and flow velocity segments that have passed the interception verification, and the output is water level and flow velocity boundary condition data containing the boundary water level sequence and boundary flow velocity sequence.

[0141] Additionally, regarding the handling and output field composition for cases where the interception verification fails: when a segment fails the interception verification, it is not written into the boundary water level sequence or boundary velocity sequence. Instead, an interception record is generated. The interception record includes at least the event number, section number, type of intercepted segment, reason for interception (one of the following three: leading edge window crossing, phase non-compliance, or missing time delay source), and the corresponding time range. For segments that pass the interception verification, each time point of the water level and velocity boundary condition data includes at least the target boundary section number, writing time, boundary water level value and its elevation datum, boundary velocity value and its direction, the event number to which it belongs, and the segment writing permission identifier. For example, a boundary condition data record that has been verified by interception can be represented as follows: the section number is B, the writing time is the 205th second, the boundary water level is the corresponding water level under a certain elevation datum, the boundary flow velocity value and its direction along the main flow direction, the event number is this flood discharge event, and the fragment writing permission identifier is passed; an interception record can be represented as follows: the section number is B, the type of the intercepted fragment is water level fragment, the reason for interception is front-end window crossing, and the corresponding time range is a certain residual response period.

[0142] It should also be noted that the intercept verification calls the event reachable window data and phase connection interval again on the write side, forming a double gating reuse that echoes the candidate screening process. The reason for performing intercept verification on the write side after candidate screening is that time delay translation may cause boundary shifts in the leading time or phase relationship of individual segments after the superposition of time delays. The end verification on the write side can intercept such offset residues and prevent erroneous segments from contaminating the water level and flow velocity boundary condition data and thus the digital twin.

[0143] Through the above technical solution, this embodiment determines the target response interception time range by superimposing water level time delay and flow velocity time delay at the start and end points of the event form boundary, and performs end interception verification on the writing side using reachable windows and phase acceptance intervals. Unlike the conventional processing method of directly writing to the boundary after a single fixed time delay translation, this embodiment applies a double verification of frontier legality and phase qualification after time delay translation and before fragment writing, thereby producing water level and flow velocity boundary condition data that has been verified multiple times in terms of event attribution and water level-flow velocity coordination relationship. This water level and flow velocity boundary condition data serves as the dynamic boundary input of the target boundary section for the construction and rolling update of the digital twin simulation module.

[0144] This step takes the water level and flow velocity boundary condition data and uses it as the dynamic boundary input of the target boundary section for the construction and dynamic updating of the digital twin of the reservoir group flood control scheduling. This ensures that the high-confidence boundary data produced by the aforementioned steps are ultimately applied to the simulation drive of the digital twin.

[0145] It should be noted that the process of using water level and velocity boundary condition data for the construction and dynamic updating of the digital twin is as follows: In the construction phase of the digital twin simulation module, the boundary water level sequence and boundary velocity sequence in the water level and velocity boundary condition data are read and set as the downstream boundary conditions at the target boundary section. Together with the upstream reservoir outflow process line, river topography data, and river roughness parameters, they constitute a complete hydrodynamic calculation boundary, thereby completing the initial construction of the digital twin simulation module in the target river section. In the dynamic updating phase of the digital twin simulation module, whenever new sampling occurs in the upstream gate scheduling data and downstream navigation channel monitoring data, the water level and velocity boundary condition data for the new time period are regenerated according to the aforementioned steps of generating cross-section reference data, generating dual response phase data, generating event reachability window data, and performing dual gating screening and interception verification. The newly generated boundary water level sequence and boundary velocity sequence are then used to replace or add to the boundary conditions of the target boundary section, thereby driving the digital twin simulation module to perform rolling hydrodynamic calculations, so that the simulation state of the digital twin is updated synchronously with the actual flood discharge and propagation process.

[0146] It should also be noted that, since the water level and velocity boundary condition data have been verified by dual gating and write-side interception in terms of event attribution, cross-section propagation order, and water level-velocity coordination relationship, when driving the digital twin, the boundary conditions received by the target boundary section can correctly correspond to the flood discharge event and propagation river section to which it belongs. This avoids the inconsistency in event attribution, cross-section propagation order, and water level-velocity coordination relationship of the boundary conditions before the simulation starts due to mixing the responses caused by different flood discharge actions into the same boundary input. This is beneficial for the digital twin to maintain the correspondence between the simulation state and the real flow state in the scenario of continuous scheduling of multiple gates.

[0147] It should be noted that regarding dynamically updated data boundaries: when the water level and velocity boundary condition data of the target boundary section are insufficient to form a complete boundary water level sequence and boundary velocity sequence due to missing monitoring data or failure to pass the interception verification within a certain update period, the digital twin simulation module continues to perform rolling calculations using the boundary conditions of the previous valid update period and marks that period as boundary conditions to be supplemented. The boundary conditions will be replaced after subsequent monitoring data is recovered and passes the interception verification, thereby avoiding abrupt changes in the simulation state caused by driving the simulation with incomplete or unverified boundary data.

[0148] Through the above technical solution, this embodiment uses the water level and flow velocity boundary condition data verified by dual gating and interception as the downstream boundary condition of the target boundary section and sends it back to the digital twin simulation module for rolling recalculation with new sampling. Unlike the conventional driving method that writes the boundary obtained by fixed time delay translation at once without event-level verification, this embodiment constrains the construction and updating of the digital twin at the event ownership consistency level of the simulation boundary. This allows the simulation state of the digital twin in the scenario of continuous scheduling of multiple gates to be continuously driven by high-confidence boundary data. This driving process constitutes the final implementation of this method in the research on the construction and dynamic updating of digital twins for flood control scheduling of reservoir groups.

[0149] Example 2:

[0150] Please see Figure 5 A system for constructing and dynamically updating a digital twin for flood control scheduling of a reservoir group, comprising:

[0151] The data acquisition module is used to acquire upstream gate scheduling data, downstream waterway monitoring data and waterway station network benchmark data of the target object, and extract event morphology boundaries to generate cross-sectional benchmark data.

[0152] The bivariate phase calculation module is used to extract the abrupt change time of water level and flow velocity from the cross-sectional reference data and calculate the bivariate phase difference to generate bivariate phase data containing the response phase interval of each cross-section.

[0153] The cross-section reachability window calculation module is used to compare the order of response phase intervals of adjacent cross-sections, and calculate the start and end points of the reachability window for each event in each cross-section for response phase intervals that meet the order of response conditions, and generate event reachability window data.

[0154] The candidate elimination module is used to extract the portion of the dual response phase data that falls into the event reachable window data as dual response candidates within the window, and construct a phase acceptance interval based on the bivariate phase difference of adjacent sections, and then determine whether the bivariate phase difference of the dual response candidates within the window falls into the phase acceptance interval.

[0155] Otherwise, calculate the phase residual between the bivariate phase difference and the phase transition interval, and remove in-window bi-response candidates whose phase residual exceeds the preset residual threshold. Then, assemble the remaining in-window bi-response candidates into piecewise variable time delay data.

[0156] The water level and flow velocity boundary condition output module is used to extract and verify the cross-sectional reference data based on the segmented variable time delay data, and generate water level and flow velocity boundary condition data.

[0157] This embodiment has the same technical effects as Embodiment 1.

[0158] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. The data mentioned in this application have undergone normalization and other preprocessing to unify dimensions during formula calculations.

[0159] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for constructing and dynamically updating a digital twin of flood control scheduling for a reservoir group, applied to the scenario of generating water level and flow velocity boundaries in the digital twin of flood control scheduling for a reservoir group, characterized in that... Includes the following steps: Acquire upstream gate scheduling data, downstream waterway monitoring data, and waterway station network baseline data of the target object, extract event morphology boundaries, and generate cross-sectional baseline data; Extract the abrupt change times of water level and flow velocity from the cross-sectional reference data and calculate the bivariate phase difference to generate bivariate phase data containing the response phase intervals of each cross-section. Compare the sequential relationship of the response phase intervals of adjacent sections, and calculate the start and end points of the reachability window for each event in each section for the response phase intervals that meet the sequential relationship conditions, and generate event reachability window data. Extract the portion of the dual-response phase data that falls into the event reachable window data as in-window dual-response candidates, and construct a phase acceptance interval based on the bivariate phase difference of adjacent sections, and then determine whether the bivariate phase difference of the in-window dual-response candidates falls into the phase acceptance interval. Otherwise, calculate the phase residual between the bivariate phase difference and the phase transition interval, and remove in-window bi-response candidates whose phase residual exceeds a preset residual threshold, and then assemble the retained in-window bi-response candidates into piecewise variable time delay data. The cross-sectional reference data is extracted and verified based on the segmented variable time delay data to generate water level and flow velocity boundary condition data.

2. The method for constructing and dynamically updating a digital twin for flood control scheduling of a reservoir group according to claim 1, characterized in that: The process of acquiring upstream gate scheduling data, downstream waterway monitoring data, and waterway station network baseline data of the target object, extracting event morphology boundaries, and generating cross-sectional baseline data specifically includes: The upstream gate scheduling data and the downstream waterway monitoring data are arranged in chronological order to generate an event processing timeline; The downstream waterway monitoring data is matched to the corresponding cross section based on the waterway station network benchmark data. Calculate the changes in gate opening and outflow on the event processing time axis of the upstream gate scheduling data, and extract the start and end points of the continuous time interval in which the direction of change of gate opening is consistent with the direction of change of outflow as the event form boundary. The downstream channel monitoring data matched to the corresponding cross section, the upstream gate scheduling data, and the event pattern boundary are integrated to generate cross section reference data.

3. The method for constructing and dynamically updating a digital twin for flood control scheduling of a reservoir group according to claim 2, characterized in that: Extracting the abrupt change times of water level and flow velocity from the cross-sectional reference data and calculating the bivariate phase difference to generate bivariate phase data containing the response phase intervals of each cross-section specifically includes: The downstream response search range is constructed by taking the starting point of the event pattern boundary as the search starting point and the end time of the change in outbound flow as the search ending point. Within the downstream response search range, the water level recording time and flow velocity recording time that meet the hydrodynamic change conditions are extracted from the cross-sectional reference data according to the cross-section, and are used as the water level change time and flow velocity change time. The difference between the water level change time and the flow velocity change time belonging to the same cross section is calculated on the event processing time axis to obtain the bivariate phase difference; The phase direction is determined by the order of occurrence of the water level change time and the flow velocity change time, and the response phase interval is formed by combining the bivariate phase difference, thereby generating dual response phase data.

4. The method for constructing and dynamically updating a digital twin for flood control scheduling of a reservoir group according to claim 3, characterized in that: The water level recording time and flow velocity recording time that satisfy the hydrodynamic abrupt change conditions are extracted from the cross-sectional reference data according to the cross-section, and are specifically included as the water level abrupt change time and flow velocity abrupt change time: Extract water level and flow velocity records from the cross-sectional reference data within the downstream response search range, and calculate the water level change and flow velocity change at adjacent times under the same cross-section based on the water level and flow velocity records. The consistency of the direction of the water level change is compared with the direction of the outflow change under the same event, and the starting point of the first water level change interval with the same direction is extracted as the water level change time. The direction of the change in flow velocity is compared with the direction of the change in outflow rate, and the starting point of the first flow velocity change interval with the same direction is extracted as the flow velocity mutation time.

5. The method for constructing and dynamically updating a digital twin for flood control scheduling of a reservoir group according to claim 3, characterized in that: The sequence of response phase intervals between adjacent sections is compared, and for response phase intervals that satisfy the sequence condition, the start and end points of the reachability window for each event at each section are calculated. The event reachability window data specifically includes: Following the order from upstream to downstream, the starting points of the response phase intervals of adjacent upstream sections and adjacent downstream sections are read respectively; Determine whether the starting point of the response phase interval of the adjacent downstream section is earlier than the starting point of the response phase interval of the adjacent upstream section; otherwise, determine that the response phase interval of the corresponding adjacent downstream section satisfies the sequential relationship condition. The later starting point of the response phase interval that satisfies the sequential relationship condition and the starting point of the event form boundary is taken as the starting point of the reachable window, and the ending point of the response phase interval of the adjacent downstream section that satisfies the sequential relationship condition is determined as the ending point of the reachable window. When the endpoint of the reachable window is not earlier than the start point of the reachable window, the start point and the endpoint of the reachable window of the corresponding section are combined to generate event reachable window data.

6. The method for constructing and dynamically updating a digital twin for flood control scheduling of a reservoir group according to claim 1, characterized in that: The phase transition interval is constructed based on the bivariate phase difference between adjacent sections, specifically including: For the current cross section, extract the bivariate phase difference corresponding to its adjacent upstream cross section in the bivariate phase data, and extract the bivariate phase difference corresponding to its adjacent downstream cross section in the bivariate phase data. Compare the numerical values ​​of the bivariate phase difference between the adjacent upstream cross-section and the adjacent downstream cross-section: The smaller bivariate phase difference is determined as the lower limit endpoint, and the larger bivariate phase difference is determined as the upper limit endpoint. The phase transition interval of the current section is constructed using the lower limit endpoint and the upper limit endpoint as boundaries.

7. The method for constructing and dynamically updating a digital twin for flood control scheduling of a reservoir group according to claim 5, characterized in that: Eliminating in-window double-response candidates whose phase residuals exceed a preset residual threshold, and then assembling the remaining in-window double-response candidates into piecewise variable time-delay data, specifically includes: The river mileage distance between the adjacent upstream section and the adjacent downstream section of the current section is extracted from the waterway station network reference data, and the river mileage distance is proportionally converted to generate a preset residual threshold. Determine whether the phase residual is greater than a preset residual threshold; if so, remove the corresponding in-window dual response candidate. Extract the water level change time and flow velocity change time corresponding to the retained dual response candidates within the window, and subtract the start time in the event pattern boundary of the corresponding event from the water level change time and the flow velocity change time to obtain the water level time delay and flow velocity time delay respectively; The water level time delay and the flow velocity time delay are structurally correlated to generate segmented variable time delay data.

8. The method for constructing and dynamically updating a digital twin for flood control scheduling of a reservoir group according to claim 7, characterized in that: Based on the segmented variable time-delay data, the cross-sectional reference data is truncated and verified to generate water level and flow velocity boundary condition data, specifically including: Read the water level time delay and the flow velocity time delay of the corresponding section from the segmented variable time delay data; The target response intercept time range is obtained by adding the corresponding water level time delay and flow velocity time delay to the start and end points of the event shape boundary of the corresponding event in the cross-sectional reference data, respectively. Within the target response interception time range, water level segments and flow velocity segments are extracted from the cross-sectional reference data; The water level segment and the flow velocity segment are subjected to interception verification, and the water level segment and the flow velocity segment that pass the interception verification are combined to generate water level and flow velocity boundary condition data.

9. The method for constructing and dynamically updating a digital twin for flood control scheduling of a reservoir group according to claim 8, characterized in that: The process of performing truncation verification on the water level segment and the flow velocity segment, and combining the verified water level segment and flow velocity segment to generate water level and flow velocity boundary condition data specifically includes: Determine whether the water level change time corresponding to the water level segment and the flow velocity change time corresponding to the flow velocity segment both fall between the start and end points of the reachable window data of the event reachable window of the corresponding cross section; If yes, determine whether the phase difference between the two variables corresponding to the water level segment and the flow velocity segment falls within the phase transition interval. If yes, determine that the interception verification is successful, and align the water level segment and the flow velocity segment that have passed the interception verification according to the time sequence to generate water level and flow velocity boundary condition data.

10. A system for constructing and dynamically updating a digital twin for flood control scheduling of a reservoir group, characterized in that, include: The data acquisition module is used to acquire upstream gate scheduling data, downstream waterway monitoring data and waterway station network benchmark data of the target object, and extract event morphology boundaries to generate cross-sectional benchmark data. The bivariate phase calculation module is used to extract the abrupt change time of water level and flow velocity from the cross-sectional reference data and calculate the bivariate phase difference to generate bivariate phase data containing the response phase interval of each cross-section. The cross-section reachability window calculation module is used to compare the order of response phase intervals of adjacent cross-sections, and calculate the start and end points of the reachability window for each event in each cross-section for response phase intervals that meet the order of response conditions, and generate event reachability window data. The candidate elimination module is used to extract the portion of the dual-response phase data that falls into the event reachable window data as dual-response candidates within the window, and construct a phase acceptance interval based on the bivariate phase difference of adjacent sections, and then determine whether the bivariate phase difference of the dual-response candidates within the window falls into the phase acceptance interval. Otherwise, calculate the phase residual between the bivariate phase difference and the phase transition interval, and remove in-window bi-response candidates whose phase residual exceeds a preset residual threshold, and then assemble the retained in-window bi-response candidates into piecewise variable time delay data. The water level and flow velocity boundary condition output module is used to extract and verify the cross-sectional reference data based on the segmented variable time delay data, and generate water level and flow velocity boundary condition data.

Images