Large river system flood intelligent conduction prediction method and system

Abstract

The application discloses a large river system flood intelligent conduction prediction method and system, relates to the technical field of flood intelligent prediction, and comprises the following steps: carrying out measured flow history backtracking on a real-time flood simulation sequence, generating shadow correction content and benchmark prediction content, carrying out same-path double-track simulation playback and cross-section alignment on the benchmark prediction content and the shadow correction content, locking a deviation convergence boundary, and generating a minimum re-calculation range; based on the minimum re-calculation range, re-calculation is carried out on the shadow correction content cross river sections, a unique conduction mark of this round is generated for each river section, each river section is limited to only accepting the re-calculation content generated by the previous river section in this round, a cross-river-section re-calculation sequence is formed, the cross-river-section re-calculation sequence is subjected to deviation convergence boundary truncation, and the benchmark prediction content reserved outside the boundary is connected, and a local re-calculation conduction sequence is formed. The application realizes self-adaptive convergence of the influence range of the measured flow correction and keeps the continuity and consistency of the multi-path conduction process.

Description

Technical Field

[0001] This invention relates to the field of intelligent flood forecasting technology, and in particular to an intelligent transmission forecasting method and system for large-scale river system floods. Background Technology

[0002] With the development of hydrological monitoring, numerical rainfall forecasting, and watershed computing technologies, flood forecasting of large river systems is gradually shifting from single-station empirical extrapolation to spatiotemporal continuous calculation based on river network topology. Existing technologies typically aggregate the data from rain gauges, meteorological forecasts, and cross-sectional flow monitoring into river network computing nodes. Based on the upstream and downstream connections, they complete rainfall-runoff conversion, river propagation time difference correction, and confluence regulation calculations. After the arrival of new measured flow, they restart the subsequent flood evolution to obtain the flow process and arrival time of each cross-section.

[0003] When large river systems contain multiple tributaries, parallel propagation paths, and multiple path confluence locations, existing technologies for correcting new measured flow rates typically proceed downstream along a predetermined propagation chain. This makes it difficult to define the endpoint of the impact in a timely manner at the point where the baseline forecast and the corrected forecast re-align. Consequently, calculations may continue to be repeated in downstream river segments where the deviation has already converged. Furthermore, the lack of a unified current-cycle propagation constraint during asynchronous updates in different river segments can easily lead to overlap between previous and current-cycle calculations at tributary access and confluence locations, affecting the clarity of local correction boundaries and the consistency of continuous propagation. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides an intelligent transmission forecasting method for large-scale river system floods to solve the problems of difficulty in accurately defining the recalculation boundary after measured flow reversal and easy overlap within the transmission process in large-scale river system flood forecasting.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides an intelligent flood propagation forecasting method for large-scale river systems, comprising: acquiring the upstream and downstream connections of each section of a large-scale river system; tracking the flood propagation sequence of the upstream and downstream connections to generate the cross-sectional propagation path; and performing backtracking correction and propagation sequence rearrangement on the cross-sectional propagation path to form a large-scale river system propagation topology; based on the large-scale river system propagation topology, performing time-series fusion and inflow conversion of measured and predicted rainfall to generate inflow sequences between adjacent sections; performing propagation evolution calculation and inter-section confluence regulation superposition on the inflow sequences between adjacent sections to generate a real-time flood simulation sequence; performing historical backtracking of measured flow on the real-time flood simulation sequence to generate shadow correction content and baseline forecast content; and performing backtracking of the baseline forecast content and shadow correction content... The system performs dual-track simulation playback along the same path and cross-section alignment to lock the deviation convergence boundary and generate a minimum recalculation range. Based on the minimum recalculation range, the shadow correction content is recalculated segment by segment, generating a unique transmission marker for each segment in this round. Each segment is limited to receiving only the recalculated content generated in this round from the previous segment, forming a segment-by-segment recalculation sequence. The segment-by-segment recalculation sequence is truncated by the deviation convergence boundary and connected with the baseline forecast content retained outside the boundary to form a local recalculation transmission sequence. Based on the local recalculation transmission sequence, the downstream flood transmission and confluence are conducted along the propagation topology of the large river system, and the flood forecast content of each cross-section is updated according to the subsequent measured flow, outputting the intelligent transmission forecast result of the large river system flood.

[0008] As a preferred embodiment of the intelligent transmission forecasting method for large-scale river system floods described in this invention, the generation of the cross-sectional transmission path is specifically as follows:

[0009] The upstream and downstream connection relationships of each section of a large river system are obtained and converted into node connection entries. Each section is converted into a section calculation node. The section calculation node identifier is unified according to the section location. Abnormal node connection entries are processed separately and a section connection index table is generated.

[0010] Based on the cross-section connection index table, starting from the cross-section calculation node without an upstream predecessor, the subsequent cross-section calculation nodes are visited level by level according to the connection direction from upstream to downstream. The order of connection between cross-section calculation nodes is recorded, and the corresponding propagation branch relationship is preserved at the tributary access position and the multi-path convergence position to generate the cross-section propagation path.

[0011] As a preferred embodiment of the intelligent transmission forecasting method for large-scale river system floods described in this invention, the formation of the large-scale river system propagation topology is specifically as follows:

[0012] Perform backtracking correction on the cross-section propagation path according to the direction from the end cross-section to the starting cross-section, verify the connection status of adjacent cross-section calculation nodes, merge the path positions that repeatedly pass through the same cross-section calculation node, correct the propagation direction of the path positions that are accessed in reverse, and add the corresponding adjacent cross-section calculation nodes to the connection interruption positions according to the cross-section connection index table to generate the backtracking correction path.

[0013] The propagation order of the retrospective correction path is rearranged according to the propagation hierarchy of the river system flood. Each cross-section calculation node is written into a unique topological position, and the preceding cross-section calculation nodes and the subsequent cross-section calculation nodes are encapsulated into the corresponding topological positions to form a large-scale river system propagation topology.

[0014] As a preferred embodiment of the intelligent transmission forecasting method for large-scale river system floods described in this invention, the generation of the real-time flood simulation sequence is specifically as follows:

[0015] Based on the propagation topology of large river systems, measured rainfall and estimated rainfall are aggregated to the corresponding cross-section calculation nodes and arranged in order. Measured rainfall and estimated rainfall are distinguished at time-overlapping locations, and the corresponding time order is retained at time-missing locations to generate a rainfall time-series aligned sequence.

[0016] Using the measured rainfall coverage period as the first segment and the estimated rainfall carryover period as the second segment, the rainfall time series alignment sequence is fused, and the fused continuous rainfall time series is solidified to the corresponding cross-section calculation node to generate a rainfall time series fusion sequence.

[0017] Based on the rainfall time series fusion sequence, the rainfall content of each cross-section calculation node is converted into runoff and inflow. The inflow content generated by the preceding cross-section calculation node and the corresponding generation time are linked to the propagation path of the corresponding subsequent cross-section calculation node. The inflow content of each tributary is retained at the tributary access point, and the inflow content corresponding to different propagation paths is retained at the multi-path confluence point, generating the inflow sequence between adjacent cross-sections.

[0018] Based on the inflow sequence between adjacent cross sections, the propagation evolution calculation and inter-section confluence regulation and storage are superimposed on the inflow of water between adjacent cross sections. According to the connection relationship between the calculation nodes of the preceding cross section and the calculation nodes of the subsequent cross section in the propagation topology of the large river system, the inflow content that can be received is written into the calculation node of the subsequent cross section, and the remaining inflow content is retained to the next time position to generate a real-time flood simulation sequence.

[0019] As a preferred embodiment of the intelligent transmission forecasting method for large-scale river system floods described in this invention, the generation of shadow correction content and baseline forecast content is specifically as follows:

[0020] Based on the real-time flood simulation sequence, the cross-sectional calculation node and historical time position corresponding to the measured flow are located, the measured flow of the real-time flood simulation sequence is historically reversed, and the historical time position is locked as the common evolution starting point of the baseline forecast content and the shadow correction content, generating the historical reversed position of the measured flow.

[0021] Using the historical return position of the measured flow as the boundary, the continuous simulation content after the boundary position in the real-time flood simulation sequence is retained as a baseline, and the shadow replay is performed starting from the measured flow written at the boundary position, generating baseline forecast content and shadow correction content respectively.

[0022] As a preferred embodiment of the intelligent transmission forecasting method for large-scale river system floods described in this invention, the generation of the minimum recalculation range is specifically as follows:

[0023] Based on the baseline forecast content and the shadow correction content, a dual-track simulation playback is performed along the same propagation path in the propagation topology of a large river system. The dual-track simulation playback refers to placing the baseline forecast content and the shadow correction content in independent bearing positions, and synchronously pushing them downstream along the same propagation path and the same time sequence, respectively retaining the evolution status of the baseline forecast content and the shadow correction content at each cross-section calculation node, and generating a dual-track playback sequence.

[0024] The two playback contents in the dual-track playback sequence are compared section by section. The sections with deviations are continuously connected, and the first section where the two playback contents begin to continuously fit together is locked as the corresponding deviation convergence boundary to generate a deviation alignment sequence.

[0025] The start and end positions of the recalculation of the deviation alignment sequence are defined. The historical backtracking position of the measured flow is taken as the starting point, and the deviation convergence boundary corresponding to the transmission path of each section is taken as the ending point. The non-converged positions of the deviation between the starting point and each ending point are continuously retained, and the section calculation nodes outside the recalculation termination position are removed to generate the minimum recalculation range.

[0026] As a preferred embodiment of the intelligent transmission forecasting method for large-scale river system floods described in this invention, the step of forming a segment-by-segment recalculation sequence is specifically as follows:

[0027] Extract the cross-section content covered by the minimum recalculation range from the shadow correction content. Based on the connection relationship between the calculation nodes of the preceding cross-section and the calculation nodes of the subsequent cross-section in the large river system propagation topology, divide the cross-section content to the recalculation position of the river segment corresponding to the adjacent cross-section, and arrange the recalculation positions of each river segment in the order from upstream to downstream to generate a list of recalculated river segments.

[0028] The recalculated river segment list is recalculated and the recalculation round is assigned to generate a recalculation round identifier for this round. This recalculation round identifier is then written into the recalculation position of each river segment in the recalculated river segment list. The recalculation round identifier, the historical backtracking position of the measured flow, and the recalculation position of the starting river segment are merged and coded to obtain a unique transmission marker for this round corresponding to the starting river segment. The recalculation round identifier, the unique transmission marker, the recalculation position of the starting river segment, and the corresponding shadow correction content are written in the same position to generate a unique transmission marker table for this round.

[0029] Based on the unique transmission marker table for this round, the shadow correction content is recalculated segment by segment. Only the recalculated content of the previous segment carrying the corresponding unique transmission marker for this round is accepted. The current unique transmission marker for this round and the recalculated position of the next segment are recursively encoded to form a segment-by-segment recalculation sequence.

[0030] As a preferred embodiment of the intelligent transmission forecasting method for large-scale river system floods described in this invention, the step of forming a locally recalculated transmission sequence is as follows:

[0031] The deviation convergence boundary is truncated for the recalculated sequence of each river segment. After the river segment where the deviation convergence boundary is located is recalculated for the propagation path of each section, the generation of the unique propagation mark for the next river segment in this round is stopped, and the recalculated content of each river segment within the deviation convergence boundary is retained to generate the boundary truncated segment.

[0032] Based on the connection between the calculation nodes of the preceding section and the calculation nodes of the subsequent section, the boundary truncated segments are recalculated and replaced within the boundary and the benchmark reused outside the boundary. The recalculated content of the river segment within each deviation convergence boundary replaces the corresponding benchmark forecast content, while the benchmark forecast content outside each deviation convergence boundary is retained, forming a local recalculation transmission sequence.

[0033] As a preferred embodiment of the intelligent transmission forecasting method for large-scale river system floods described in this invention, the output of the intelligent transmission forecasting results for large-scale river system floods is specifically as follows:

[0034] Based on the connection relationship between the calculation nodes of the preceding section and the calculation nodes of the succeeding section in the propagation topology of large river systems, the local recalculated propagation sequence is recursively pushed down the river segment by segment and truncated and connected at the deviation convergence boundary to generate the downstream propagation propagation segment.

[0035] The tributary access points and multi-path confluence points in the downstream propagation segment are merged and merged. The flood content transmitted by different preceding section calculation nodes is connected to the same subsequent section calculation node according to the arrival time. The flood content after merging is continuously spliced ​​in time to generate the confluence splice segment.

[0036] The flow segment of the confluence connection segment is updated in segments. The subsequent measured flow is written into the corresponding cross-section calculation node and the corresponding time position, replacing the flood forecast content at the same position. The updated position is locked by actual measurement, and the flood forecast content at the position where the subsequent measured flow has not yet been obtained is retained, generating a cross-section forecast update list.

[0037] The forecast content of the cross-section forecast update list is rolled and fixed. The latest measured flow locked by the calculation node of each cross section is used as the starting point for subsequent propagation. The flood forecast content that has not been replaced by the subsequent measured flow is arranged in sequence according to the propagation topology of the large river system and the continuous time position, and the intelligent propagation forecast result of the large river system flood is output.

[0038] Secondly, the present invention provides a large-scale intelligent river system flood transmission forecasting system, comprising:

[0039] The topology construction module is used to obtain the upstream and downstream connection relationships of each section of a large river system, track the flood propagation order of the upstream and downstream connection relationships, generate the cross-sectional conduction path, and perform backtracking correction and propagation order rearrangement of the cross-sectional conduction path to form the propagation topology of the large river system.

[0040] The inflow simulation module is used to perform time-series fusion and inflow conversion of measured and predicted rainfall based on the propagation topology of large river systems, generate inflow sequences between adjacent sections, perform propagation evolution calculation and inter-section confluence regulation superposition of inflow sequences between adjacent sections, and generate real-time flood simulation sequences.

[0041] The deviation positioning module is used to perform measured flow history reversal on the real-time flood simulation sequence, generate shadow correction content and baseline forecast content, perform same-path dual-track simulation playback and cross-section alignment on the baseline forecast content and shadow correction content, lock the deviation convergence boundary, and generate the minimum recalculation range.

[0042] The local recalculation module is used to recalculate the shadow correction content segment by segment based on the minimum recalculation range, generate a unique transmission marker for each segment in this round, and limit each segment to only receive the recalculation content generated in this round from the previous segment, forming a segment-by-segment recalculation sequence. The segment-by-segment recalculation sequence is truncated by the deviation convergence boundary and connected with the baseline forecast content retained outside the boundary to form a local recalculation transmission sequence.

[0043] The transmission forecast module is used to perform downstream flood transmission and confluence continuation along the propagation topology of large river systems based on the local recalculation of the transmission sequence, and update the flood forecast content of each section according to the subsequent measured flow, and output the intelligent transmission forecast results of large river system floods.

[0044] The beneficial effects of this invention are as follows: by performing dual-track simulation playback and cross-section alignment on the baseline forecast content and the shadow correction content, the flow deviation and arrival time deviation of each cross-section are quantified, the deviation convergence boundary at which the two contents begin to continuously overlap is locked, and the river segments with unconverged deviations and the overlapping river segments outside the boundary are retained accordingly. This achieves adaptive convergence of the influence range of measured flow correction, provides accurate start and end positions for subsequent recalculation of each river segment and boundary truncation, reduces invalid repeated calculations, and maintains the continuity and consistency of the multi-path transmission process in large river systems. Attached Figure Description

[0045] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0046] Figure 1 This is a flowchart of a method for intelligent transmission forecasting of floods in large river systems.

[0047] Figure 2 This is a schematic diagram of a large-scale intelligent flood transmission and forecasting system for river systems.

[0048] Figure 3 This is a flowchart generated for the historical callback and minimum recalculation range of the measured traffic.

[0049] Figure 4 The flowchart shows the generation of the propagation sequence for recalculation of each river segment and local recalculation. Detailed Implementation

[0050] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0051] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0052] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0053] Reference Figures 1-4As one embodiment of the present invention, this embodiment provides a method for intelligent transmission forecasting of floods in large river systems, comprising the following steps:

[0054] S1: Obtain the upstream and downstream connections of each section of a large river system, track the flood propagation order of the upstream and downstream connections, generate the cross-sectional conduction path, and perform backtracking correction and propagation order rearrangement on the cross-sectional conduction path to form the propagation topology of the large river system.

[0055] S1.1: Obtain the upstream and downstream connection relationships of each section of a large river system and convert them into node connection entries. Convert each section into a section calculation node, unify the section calculation node identifier according to the section location, and separate abnormal node connection entries to generate a section connection index table.

[0056] Upstream and downstream connection refers to the connection between adjacent sections in a large river system according to the direction of water flow propagation. It is used to indicate the upstream preceding section and the downstream succeeding section corresponding to each section, and to retain the connection between the tributary access point and the confluence of multiple paths.

[0057] A cross-section calculation node refers to an independent calculation location formed by the transformation of each actual cross-section. It is used to carry the cross-section calculation node identifier, the predecessor cross-section calculation node, the successor cross-section calculation node, as well as the subsequently written rainfall content, water inflow content, and flood forecast content. It also serves as the basic location for performing cross-section-by-cross-section propagation calculations in the propagation topology of large river systems.

[0058] The cross-sectional location refers to the river affiliation, river segment affiliation, and upstream / downstream order of the cross-section in a large river system. It is used to distinguish cross-sections in different main streams, different tributaries, and different propagation levels, and serves as a unified cross-sectional calculation node identifier and a basis for verifying upstream and downstream connections.

[0059] Furthermore, the upstream and downstream connections of each cross-section in a large river system are obtained. Following the direction of water flow from upstream to downstream, the connection status of adjacent cross-sections is checked item by item. The positions of the preceding and succeeding cross-sections for each group of adjacent cross-sections are recorded and converted into node connection entries. Each cross-section is located based on these node connection entries, and each cross-section is converted into an independent cross-section calculation node. The positions of the preceding and succeeding cross-sections in the node connection entries are mapped to the corresponding cross-section calculation nodes. Cross-section calculations are then uniformly arranged according to the river, river segment, and upstream / downstream order of the cross-section. Calculate node identifiers to ensure that only one node identifier corresponds to the same cross section; verify the cross section node identifiers and upstream / downstream directions in the node connection entries, and remove node connection entries with missing cross section node identifiers, duplicate connections of the same cross section node, conflicting upstream / downstream directions, or interrupted connections from the normal connection range, and write them into the abnormal connection positions in the cross section connection index table as the basis for subsequent connection review and backtracking correction; arrange the remaining node connection entries according to the cross section node identifiers and upstream / downstream order to generate the cross section connection index table.

[0060] S1.2: Based on the cross-section connection index table, starting from the cross-section calculation node without an upstream predecessor, the subsequent cross-section calculation nodes are visited level by level according to the connection direction from upstream to downstream. The connection order between cross-section calculation nodes is recorded, and the corresponding propagation branch relationship is preserved at the tributary access position and the multi-path convergence position to generate the cross-section propagation path.

[0061] Based on the cross-section connection index table, cross-section calculation nodes without upstream predecessors are located one by one according to the cross-section calculation node identifier, and the cross-section calculation nodes without upstream predecessors are written to the beginning of the path; according to the correspondence between the predecessor cross-section calculation node and the successor cross-section calculation node in the node connection entry, the successor cross-section calculation node is visited level by level along the upstream to downstream direction, and the predecessor cross-section calculation node, successor cross-section calculation node and the order of connection formed by each visit are written into the continuous path position in sequence.

[0062] At the tributary access location, the preceding section calculation nodes corresponding to each tributary are connected to the corresponding successor section calculation nodes, and are placed into different propagation paths according to the tributary access order. At the multi-path confluence location, the preceding section calculation nodes at the ends of different propagation paths are connected to the same successor section calculation node, and the confluence access order of each propagation path is retained. The section propagation path is generated by arranging the path first position, continuous path position, propagation branch relationship and confluence location in sequence.

[0063] S1.3: Perform backtracking correction on the cross-section propagation path according to the direction from the end cross-section to the starting cross-section, verify the connection status of adjacent cross-section calculation nodes, merge the path positions that repeatedly pass through the same cross-section calculation node, correct the propagation direction of the path positions that are accessed in reverse, and add the corresponding adjacent cross-section calculation nodes to the connection interruption positions according to the cross-section connection index table to generate the backtracking correction path.

[0064] Following the direction from the end section to the starting section, the preceding section calculation nodes are traced back level by level from the end section calculation node in the section propagation path. The identification of adjacent section calculation nodes and their upstream and downstream connection relationships are checked according to the section connection index table. When the section calculation node corresponding to the previous tracing position is connected to the current section calculation node in the section connection index table, the preceding and following positions of the adjacent section calculation nodes are retained and arranged continuously according to the tracing sequence.

[0065] For each consecutive path location, the cross-section calculation node identifier is checked. Duplicate path locations with the same cross-section calculation node identifier and the same preceding and subsequent cross-section calculation nodes are merged into the same location. The order of the subsequent cross-section calculation nodes arranged before the preceding cross-section calculation nodes is swapped. For connection interruption locations that lack preceding or subsequent cross-section calculation nodes, the corresponding adjacent cross-section calculation nodes are located from the cross-section connection index table and the interruption locations are filled in. Then, the paths are arranged in the order of upstream to downstream to generate a backtracking correction path.

[0066] It should be noted that the upstream and downstream connection status refers to the state in which the upstream and downstream correspondence, connection order and flood propagation direction of the calculation nodes of adjacent cross sections in the cross section connection index table remain continuous and consistent.

[0067] S1.4: The propagation order of the retrospective correction path is rearranged according to the propagation hierarchy of the river system flood. Each cross-section calculation node is written into a unique topological position, and the preceding cross-section calculation nodes and the subsequent cross-section calculation nodes are encapsulated into the corresponding topological positions to form a large-scale river system propagation topology.

[0068] The preceding cross-section calculation node refers to the adjacent cross-section calculation node located upstream of the current cross-section calculation node and transmitting flood information to the current cross-section calculation node.

[0069] The successor cross-section calculation node refers to the adjacent cross-section calculation node located downstream of the current cross-section calculation node and inheriting the flood content of the current cross-section calculation node.

[0070] Following the order of flood propagation hierarchy in the river system, starting from the initial cross-section calculation node in the retrospective correction path, the downstream connected cross-section calculation nodes are located level by level. Cross-section calculation nodes without upstream precursors are arranged in the first layer, and the subsequent cross-section calculation nodes that directly inherit from the first layer are arranged in the next layer. The remaining cross-section calculation nodes are arranged in the same way, and the propagation hierarchy corresponding to each tributary is retained at the tributary access location. At the multi-path confluence location, the confluence cross-section calculation nodes are arranged after all the precursor cross-section calculation nodes.

[0071] Each cross-sectional computing node after the propagation hierarchy is assigned a unique topological position. The cross-sectional computing node identifier is written into the corresponding unique topological position. The predecessor cross-sectional computing node and the successor cross-sectional computing node that directly connect to the current cross-sectional computing node are extracted from the backtracking correction path. The predecessor cross-sectional computing node identifier is written into the upstream connection position of the current unique topological position, and the successor cross-sectional computing node identifier is written into the downstream connection position of the current unique topological position. All unique topological positions are connected in the order of propagation hierarchy to form a large river system propagation topology.

[0072] Large-scale river system propagation topology refers to the river system conduction structure formed by writing each cross-section calculation node into a unique topological position according to the propagation level of flood from upstream to downstream, and recording the predecessor cross-section calculation nodes, successor cross-section calculation nodes, tributary access relationships and multi-path convergence relationships corresponding to each cross-section calculation node.

[0073] S2: Based on the propagation topology of large river systems, time-series fusion and inflow conversion of measured and predicted rainfall are performed to generate inflow sequences between adjacent sections. Propagation evolution calculation and inter-section confluence regulation are performed on the inflow sequences between adjacent sections to generate real-time flood simulation sequences.

[0074] S2.1: Based on the propagation topology of large river systems, measured rainfall and estimated rainfall are aggregated to the corresponding cross-section calculation nodes and arranged in order. Measured rainfall and estimated rainfall are distinguished at time-overlapping positions, and the corresponding time order is retained at time-missing positions to generate a rainfall time-series aligned sequence.

[0075] Furthermore, based on the propagation topology of large river systems, the cross-sectional calculation nodes corresponding to the measured and predicted rainfall are located one by one according to the topological location of the cross-sectional calculation nodes. The measured and predicted rainfall are assigned to the corresponding cross-sectional calculation nodes according to the cross-sectional location. The occurrence time of the measured and predicted rainfall is extracted, and the rainfall content in the same cross-sectional calculation node is arranged in chronological order, while retaining the source type of the measured and predicted rainfall.

[0076] Overlapping rainfall data with the same time sequence in each cross-section calculation node is distinguished, and measured rainfall and predicted rainfall are written to the corresponding positions under the same time sequence to avoid the two types of rainfall data overlapping each other; the positions where rainfall data is missing between adjacent rainfall times are located, and the corresponding time sequence is filled in according to the time interval before and after, while the vacancy state is retained, and rainfall data is not filled in for the missing time positions; the rainfall time sequence is generated by arranging the cross-section calculation nodes and continuous time sequence in the large river system propagation topology.

[0077] It should be noted that the measured rainfall comes from rainfall records from rain gauges, meteorological stations, and weather radars within the coverage area of ​​large river systems, and the measured rainfall reflects the rainfall occurrence time and corresponding cross-sectional location.

[0078] The estimated rainfall comes from the rainfall forecasts for the future period issued by meteorological forecasting agencies. The estimated rainfall includes the expected rainfall time and the corresponding cross-sectional location, and is used to receive the rainfall process after the actual rainfall coverage period.

[0079] S2.2: Using the measured rainfall coverage period as the front segment and the estimated rainfall receiving period as the back segment, the rainfall time series alignment sequence is fused, and the fused continuous rainfall time series is solidified to the corresponding cross-section calculation node to generate the rainfall time series fusion sequence.

[0080] Furthermore, based on the rainfall time-series alignment sequence, the start and end times of the measured rainfall coverage are located item by item according to the cross-sectional calculation nodes. The measured rainfall within the measured rainfall coverage period is retained in the earlier time position. The end position of the measured rainfall coverage period is used as the connection position between the preceding and following periods. The estimated rainfall after the end position is extracted and arranged in the later time position according to the expected rainfall time. The estimated rainfall that overlaps with the measured rainfall coverage period is not written again.

[0081] The preceding and following time positions are continuously connected. The measured and predicted rainfall corresponding to the same cross-section calculation node are written into the continuous rainfall time position in a unified time order. For the time missing positions retained in the rainfall time series alignment sequence, the corresponding time order is retained, and the measured or predicted rainfall is not added. According to the cross-section calculation node order in the large river system propagation topology, the continuous rainfall time position is fixed to the corresponding cross-section calculation node to generate a rainfall time series fusion sequence.

[0082] S2.3: Based on the rainfall time series fusion sequence, the rainfall content of each cross-section calculation node is converted into runoff and inflow. The inflow content generated by the preceding cross-section calculation node and the corresponding generation time are linked to the propagation path of the corresponding subsequent cross-section calculation node. The inflow content corresponding to each tributary is retained at the tributary access position, and the inflow content corresponding to different propagation paths is retained at the multi-path confluence position, generating the inflow sequence between adjacent cross-sections.

[0083] Furthermore, based on the rainfall time-series fusion sequence, according to the order of the cross-sectional calculation nodes in the large river system propagation topology, the rainfall content corresponding to the continuous time position of each cross-sectional calculation node is extracted item by item; the continuous rainfall positions are connected according to the occurrence time, the original time order is retained for the time missing positions, and the existing runoff generation and runoff calculation relationship is used to convert the rainfall content of each time position into the corresponding inflow content, and then the inflow content is written into the same time position of the corresponding cross-sectional calculation node.

[0084] Based on the connection relationship between the preceding and succeeding cross-section calculation nodes, the inflow content and corresponding generation time generated by the preceding cross-section calculation node are linked to the propagation path of the corresponding succeeding cross-section calculation node. At the tributary access location, the inflow content and corresponding generation time of each tributary are retained, and the inflow content of different tributaries is not merged in advance. At the multi-path confluence location, the inflow content, generation time and preceding cross-section calculation node position corresponding to different propagation paths are retained, and the inflow content of different propagation paths is not merged in advance. The inflow sequence between adjacent cross-sections is generated by arranging them according to the order of cross-section calculation nodes and generation time.

[0085] It should be noted that the tributary access location refers to the location where the preceding section calculation node of the tributary is connected to the succeeding section calculation node of the main stream, which is used to retain the inflow content of each tributary and the corresponding generation time.

[0086] The multi-path convergence point refers to the location where the preceding section calculation nodes corresponding to two or more propagation paths are connected to the same subsequent section calculation node. It is used to retain the incoming water content, generation time and preceding section calculation node location corresponding to different propagation paths, and to undertake subsequent propagation evolution calculations.

[0087] S2.4: Based on the inflow sequence between adjacent cross sections, the propagation evolution calculation and inter-section confluence regulation and storage superposition of the inflow of water between adjacent cross sections are performed. According to the connection relationship between the calculation nodes of the preceding cross section and the calculation nodes of the subsequent cross section in the propagation topology of the large river system, the inflow content that can be received is written into the calculation node of the subsequent cross section, and the remaining inflow content is retained to the next time position to generate a real-time flood simulation sequence.

[0088] Furthermore, based on the inflow sequence between adjacent cross sections, the computational nodes of adjacent precursor and successor cross sections are located item by item according to the propagation topology of the large river system. The generation time of each inflow content in the precursor cross section computational node is extracted, and the arrival time of the inflow content to the successor cross section computational node is calculated according to the propagation duration corresponding to the adjacent cross sections. The expression is as follows:

[0089] ;

[0090] In the formula, The arrival time of the node is calculated after the incoming water reaches the subsequent cross section; The generation time of the incoming water content in the calculation node of the precursor section; The propagation time between the preceding section calculation node and the succeeding section calculation node is defined; the continuous time position in the succeeding section calculation node is located according to the arrival time, and the incoming water content with the same arrival time is written to the same continuous time position, while the incoming water content with different arrival times is written to the corresponding continuous time positions respectively.

[0091] The inflow data after propagation evolution calculation is extracted based on continuous time positions. The inflow data arriving at different precursor section calculation nodes within the same continuous time position is superimposed with the inflow data that has not yet been received downstream from the previous continuous time position for interval confluence regulation and storage. The expression is as follows:

[0092]

[0093] In the formula, This represents the total water inflow within the interval corresponding to the current continuous time position. For the first The amount of water arriving at each preceding cross-section node at the current continuous time position; The number of predecessor cross-section calculation nodes that connect to the same successor cross-section calculation node; This represents the inflow volume that has not yet been received downstream in the previous continuous time position. Based on the correspondence between upstream inflow and downstream outflow during historical floods at adjacent cross sections, the inflow content that can be received downstream at the current continuous time position is extracted from the total inflow volume of the interval. The remaining inflow content is retained for the next continuous time position. Then, the inflow content after interval confluence regulation and superposition is arranged step by step according to the receiving relationship between the calculation nodes of the preceding cross section and the calculation nodes of the subsequent cross section to generate a real-time flood simulation sequence.

[0094] S3: Perform historical backtracking of measured flow on the real-time flood simulation sequence to generate shadow correction content and baseline forecast content. Perform dual-track simulation playback and cross-section alignment on the baseline forecast content and shadow correction content along the same path to lock the deviation convergence boundary and generate the minimum recalculation range.

[0095] S3.1: Based on the real-time flood simulation sequence, locate the cross-sectional calculation node and historical time position corresponding to the measured flow, perform historical backtracking of the measured flow in the real-time flood simulation sequence, and lock the historical time position as the common evolution starting point of the baseline forecast content and the shadow correction content to generate the historical backtracking position of the measured flow.

[0096] Furthermore, based on the real-time flood simulation sequence, the cross-sectional calculation nodes corresponding to the measured flow are located one by one according to the cross-sectional calculation node identifiers. The measurement time carried by the measured flow is extracted, and the historical time position consistent with the measurement time is found in the continuous time position of the corresponding cross-sectional calculation node. In the case where the measurement time is missing a corresponding position, the corresponding time sequence is filled in according to the adjacent time positions before and after, and then the measured flow is written into the historical time position to replace the original simulation flow content of the historical time position. At the same time, the real-time flood simulation sequence content and the original time sequence before the historical time position are retained.

[0097] Using the historical time position of the actual flow as the backtracking boundary, the real-time flood simulation sequence content that still needs to be propagated after the historical time position is located along the topology of the large river system is locked as the common evolution starting point of the baseline forecast content and the shadow correction content to be pushed downstream; the cross-section calculation node identifier, historical time position and the actual flow written are aligned, and the connection relationship between the common evolution starting point and the subsequent continuous time positions is preserved to generate the historical backtracking position of the actual flow.

[0098] The baseline forecast content is the original flow process and arrival time formed by the real-time flood simulation sequence extending downstream from the historical backtracking position of the measured flow before the measured flow is written into the historical time position.

[0099] The shadow correction involves writing the measured flow rate into the historical rollback position of the measured flow rate and replacing the original simulated flow rate. Then, using the same river segment propagation parameters and river flood calculation relationships, the flow process and arrival time are re-estimated downstream.

[0100] S3.2: Using the historical return position of the measured flow as the boundary, the continuous simulation content after the boundary position in the real-time flood simulation sequence is retained as the baseline, and the shadow replay is performed starting from the measured flow written at the boundary position, generating the baseline forecast content and the shadow correction content respectively.

[0101] Furthermore, using the historical return position of measured flow as the boundary, and according to the cross-sectional calculation node identifier and continuous time position, the continuous simulation content after the boundary position in the real-time flood simulation sequence is extracted. The original cross-sectional calculation node sequence, time sequence, and the succession relationship between the preceding cross-sectional calculation node and the subsequent cross-sectional calculation node are retained in the continuous simulation content. The continuous simulation content is then arranged sequentially from the historical return position of measured flow to generate the baseline forecast content.

[0102] The measured flow rate written to the historical callback position of the measured flow rate is used as the starting point of the shadow replay. The river segment propagation parameters, river flood calculation relationships and continuous time positions used in the real-time flood simulation sequence are retained. The measured flow rate replaces the original simulated flow rate carried by the common evolution starting point. Along the propagation topology of the large river system, the outflow content of the replaced predecessor section calculation node, the inflow content of the corresponding river segment interval and the water storage content of the river segment at the previous time position are sequentially used to perform river flood calculation to obtain the shadow outflow content of the successor section calculation node and the water storage content of the river segment at the next time position. At the tributary access position and the multi-path confluence position, the shadow outflow content of different propagation paths is continued according to the corresponding arrival time to generate shadow correction content.

[0103] S3.3: Based on the baseline forecast content and the shadow correction content, perform dual-track simulation playback along the same propagation path in the propagation topology of the large river system. Dual-track simulation playback means placing the baseline forecast content and the shadow correction content in independent bearing positions, and synchronously pushing them downstream along the same propagation path and the same time sequence. The evolution status of the baseline forecast content and the shadow correction content at each cross-section calculation node is retained to generate a dual-track playback sequence.

[0104] Furthermore, based on the baseline forecast content and the shadow correction content, and taking the historical backtracking position of the measured flow as a common starting point, the cross-sectional calculation nodes in the same propagation path are located step by step along the propagation topology of the large river system. Within each cross-sectional calculation node, the baseline forecast content carrying position and the shadow correction content carrying position are set respectively. The baseline forecast content and the shadow correction content are written synchronously in the same time order and transmitted downstream step by step according to the same preceding cross-sectional calculation node and subsequent cross-sectional calculation node connection relationship.

[0105] At the tributary access point, the baseline forecast content and the shadow correction content are transmitted along the corresponding propagation branches respectively; at the multi-path convergence point, they are written into the same subsequent section calculation node according to the same arrival time and the same convergence order; the baseline forecast content and the shadow correction content in each section calculation node retain the flood content, arrival time and the preceding and following connection positions respectively, and the two contents are not allowed to overlap each other, and are arranged one-to-one according to the section calculation node order and continuous time position to generate a dual-track playback sequence.

[0106] The evolution state refers to the flood content, arrival time, and preceding and following locations when the baseline forecast content and the shadow correction content arrive at each cross-section calculation node along the same propagation path.

[0107] S3.4: Perform segment-by-segment comparison on the two playback contents in the dual-track playback sequence, continuously connect the segment positions with deviations, and lock the first segment position where the two playback contents begin to continuously fit together as the corresponding deviation convergence boundary to generate a deviation alignment sequence.

[0108] Furthermore, according to the sequence of cross-sectional calculation nodes and continuous time positions in the dual-track playback sequence, the baseline forecast content and the shadow correction content are aligned item by item at the same cross-sectional calculation nodes and the same time positions. The baseline forecast flow and baseline arrival time carried by the baseline forecast content, as well as the shadow correction flow and shadow arrival time carried by the shadow correction content, are extracted. The minimum flow resolution value calibrated by the cross-sectional flow monitoring equipment is retrieved, and the time interval between adjacent continuous time positions in the dual-track playback sequence is extracted. The minimum flow resolution value is used as the minimum distinguishable flow difference, and the time interval between adjacent continuous time positions is used as the minimum distinguishable time difference.

[0109] The expression for calculating the double-track deviation is as follows:

[0110] ;

[0111] In the formula, This represents the double-track deviation corresponding to the current cross-section calculation node. This represents the propagation order of the current cross-section calculation node in the corresponding cross-section propagation path; The baseline forecast flow is the baseline forecast content carried at the corresponding time location of the calculation node at the current section; The shadow correction traffic carried by the shadow correction content at the corresponding time position of the current section calculation node; The reference arrival time is the time when the baseline forecast content reaches the current section calculation node; The arrival time of the shadow correction content to the current section calculation node; The minimum interval between adjacent distinguishable flow readings extracted from the calibration records of cross-sectional flow monitoring equipment is used to characterize the minimum distinguishable flow difference; The smallest time interval greater than zero extracted from adjacent consecutive time positions in a dual-track playback sequence is used to characterize the minimum distinguishable time difference.

[0112] When the deviation of the two tracks is not greater than 1, the corresponding cross-section calculation node is retained as the fitting cross-section position; when the deviation of the two tracks is greater than 1, the corresponding cross-section calculation node is marked as the deviation cross-section position. The fitting cross-section position and the deviation cross-section position are arranged from upstream to downstream along the transmission path of each cross-section, and the deviation cross-section positions that are adjacent to each other and whose deviation of the two tracks is continuously greater than 1 are continuously connected.

[0113] When a fitting section with a double-track deviation of no more than 1 appears after the deviation section position, the calculation nodes of subsequent sections are checked along the corresponding section propagation path. When the double-track deviation in the subsequent section calculation nodes is consistently no more than 1, the first section calculation node corresponding to the continuous fitting start position is locked as the deviation convergence boundary. When the double-track deviation in the subsequent section calculation nodes is greater than 1 again, the boundary locking of the previous fitting section position is canceled, and the check continues downstream from the next deviation section position. The deviation section positions, fitting section positions, and deviation convergence boundaries corresponding to each section propagation path are arranged in sequence to generate a deviation alignment sequence.

[0114] It should be noted that the expression converts the difference between the baseline forecast flow and the shadow correction flow into a dimensionless flow deviation through the minimum distinguishable flow difference, and converts the difference between the baseline arrival time and the shadow arrival time into a dimensionless time deviation through the minimum distinguishable time difference. The product of the two types of deviations is used to reflect the interaction effect when the flow deviation and the arrival time deviation occur simultaneously. Then, a unified dual-track deviation quantity is formed by taking the square root of the sum of the squares, which is used to objectively distinguish the deviation section position and the fitting section position, and to lock the deviation convergence boundary along the section transmission path.

[0115] S3.5: Recalculate the start and end positions of the deviation alignment sequence, taking the historical backtracking position of the measured flow as the starting point and the deviation convergence boundary corresponding to the transmission path of each section as the ending point. The non-converged positions of the deviation between the starting point and each ending point are continuously retained, and the section calculation nodes outside the recalculation termination positions are removed to generate the minimum recalculation range.

[0116] Furthermore, based on the deviation alignment sequence, the historical backtracking position of the measured flow and the deviation convergence boundary are located according to the transmission path of each section. The historical backtracking position of the measured flow is written into the recalculation start position of the corresponding section transmission path. For the section transmission path that can locate the deviation convergence boundary, the deviation convergence boundary is written into the recalculation termination position. For the section transmission path that cannot locate the deviation convergence boundary, the end section calculation node of the corresponding section transmission path is located along the propagation topology of the large river system, and the end section calculation node is written into the recalculation termination position.

[0117] Along the propagation path of each section, the positions of the deviation sections are extracted step by step from the starting position of the recalculation downstream. The calculation nodes of the sections where the baseline forecast content and the shadow correction content are not yet continuously aligned, as well as the preceding and following positions, are retained. The starting position of the recalculation, the position where the deviation has not converged, and the ending position of the recalculation are arranged continuously from upstream to downstream. The calculation nodes of the sections other than the ending position of the recalculation are removed. When the tributary propagation path corresponds to different ending positions of the recalculation, the continuous positions in the propagation path of each section are retained to generate the minimum recalculation range.

[0118] The minimum recalculation range refers to the continuous river segment in the transmission path of each section, from the historical return position of the measured flow to the corresponding deviation convergence boundary, that needs to be recalculated and does not include the content already attached outside the boundary.

[0119] S4: Based on the minimum recalculation range, the shadow correction content is recalculated segment by segment to generate a unique transmission marker for each segment in this round. Each segment is limited to receiving only the recalculated content generated in this round from the previous segment, forming a segment-by-segment recalculation sequence. The segment-by-segment recalculation sequence is truncated by the deviation convergence boundary and connected with the baseline forecast content retained outside the boundary to form a local recalculation transmission sequence.

[0120] S4.1: Extract the cross-section content covered by the minimum recalculation range from the shadow correction content. Based on the connection relationship between the calculation nodes of the preceding cross-section and the calculation nodes of the subsequent cross-section in the large river system propagation topology, divide the cross-section content to the recalculation position of the river segment corresponding to the adjacent cross-section, and arrange the recalculation positions of each river segment in the order from upstream to downstream to generate a list of recalculated river segments.

[0121] Furthermore, the cross-sectional content covered by the minimum recalculation range is extracted from the shadow correction content according to the cross-sectional calculation node identifier and continuous time position. The shadow correction content, arrival time, and cross-sectional calculation node order corresponding to the cross-sectional content are retained. The cross-sectional calculation nodes between the starting point of the minimum recalculation range and the deviation convergence boundary are located along the propagation path of each cross-section. Based on the propagation topology of the large river system, the predecessor cross-sectional calculation node identifier and the successor cross-sectional calculation node identifier corresponding to the adjacent cross-sectional calculation node are compared item by item. When the successor cross-sectional calculation node identifier recorded by the current cross-sectional calculation node is consistent with the adjacent downstream cross-sectional calculation node identifier, and the predecessor cross-sectional calculation node identifier recorded by the adjacent downstream cross-sectional calculation node is consistent with the current cross-sectional calculation node identifier, the corresponding succession relationship is retained.

[0122] The cross-sectional content carried between adjacent preceding cross-section calculation nodes and successor cross-section calculation nodes is divided into the same river segment recalculation position, so that each river segment recalculation position corresponds to the upstream cross-section calculation node, the downstream cross-section calculation node, and the shadow correction content, respectively. At the tributary access position, the river segment recalculation position corresponding to each tributary is retained. At the multi-path confluence position, the river segment recalculation positions corresponding to different propagation paths are arranged before the common successor cross-section calculation node. Then, all river segment recalculation positions are arranged item by item according to the propagation order from upstream to downstream to generate a list of recalculated river segments.

[0123] S4.2: Recalculate the recalculation rounds of the recalculated river segment list, generate the recalculation round identifier for this round, and write the recalculation round identifier for this round into the recalculation position of each river segment in the recalculated river segment list. Combine and encode the recalculation round identifier for this round, the historical backtracking position of the measured flow, and the recalculation position of the starting river segment to obtain the unique transmission mark for this round corresponding to the starting river segment. Write the recalculation round identifier for this round, the unique transmission mark for this round, the recalculation position of the starting river segment, and the corresponding shadow correction content in the same position to generate the unique transmission mark table for this round.

[0124] Furthermore, based on the recalculated river segment list, the recalculation position of the first river segment is located in the order from upstream to downstream. This first recalculation position is locked as the starting recalculation position, and the preceding cross-section calculation node identifier, the subsequent cross-section calculation node identifier, and the historical backtracking position of the measured flow corresponding to the starting recalculation position are extracted. The recalculation round identifiers are arranged according to the order in which the historical backtracking of the measured flow triggers the recalculation, so that each local recalculation corresponds to an independent recalculation round identifier. The recalculation rounds are then arranged according to their order of occurrence. The fixed sequence of the sub-identifier, the historical return position of the measured flow, the calculation node identifier of the preceding section, the calculation node identifier of the subsequent section, and the recalculated position of the starting river segment is used to generate a unique transmission marker for the starting river segment in this round. The recalculated position of the starting river segment is used as the writing position. The recalculated round identifier, the unique transmission marker, the recalculated position of the starting river segment, and the corresponding shadow correction content are written into the same bearing position and fixed according to the order of the recalculated position of the starting river segment to generate a unique transmission marker table for this round.

[0125] It should be noted that the recalculation round identifier is used to distinguish the local recalculation process triggered by the historical rollback of different measured flow rates, and the unique transmission marker in this round is used to indicate the transmission location of the recalculation content of each river segment within the corresponding recalculation round.

[0126] S4.3: Based on the unique transmission marker table of this round, the shadow correction content is recalculated segment by segment. Only the recalculated content of the previous segment carrying the corresponding unique transmission marker of this round is accepted. The current unique transmission marker of this round and the recalculated position of the next segment are recursively encoded to form a segment-by-segment recalculation sequence.

[0127] Furthermore, based on the unique transmission marker table for this round, the recalculation position of the starting river segment is located according to the upstream to downstream arrangement in the recalculated river segment list. The shadow correction content corresponding to the starting river segment is used as the upstream input. Combining the inflow content of the river segment interval corresponding to the starting river segment, the water storage content of the river segment at the previous time position, and the river segment propagation parameters, the outflow content corresponding to the downstream section calculation node and the water storage content of the river segment at the next time position are calculated using the river flood calculation relationship. The calculated arrival time is written into the corresponding continuous time position to form the recalculation content of the starting river segment. The recalculation round identifier and the unique transmission marker of the starting river segment for this round are carried in the recalculation content of the starting river segment.

[0128] The recalculation round identifier carried by the recalculation content of the preceding river segment is extracted for the subsequent river segment and checked against the recalculation round identifier carried by the recalculation position of the subsequent river segment. If the recalculation round identifier matches, the recalculation content of the preceding river segment is accepted; otherwise, it is not accepted. At the confluence of multiple paths, the recalculation content of each preceding river segment with the same recalculation round identifier is collected according to arrival time. Then, combined with the inflow content of the corresponding river segment interval, the water storage content of the river segment at the previous time position, and the river segment propagation parameters, the river flood calculation is performed to form the recalculation content of the subsequent river segment. The unique transmission marker of each preceding river segment in this round is arranged according to the propagation order and jointly encoded with the recalculation round identifier and the recalculation position of the subsequent river segment to generate the unique transmission marker of the subsequent river segment in this round. The same acceptance and recalculation relationship is used to iterate segment by segment to form a segment-by-segment recalculation sequence.

[0129] It should be noted that the unique transmission marker for this round is updated segment by segment as the river segment is recalculated. This is used to distinguish the recalculated content generated by different river segments within the same recalculation round. The same unique transmission marker for this round is not used as a condition for succession for river segments before and after.

[0130] S4.4: The deviation convergence boundary is truncated for the recalculated sequence of each river segment. After the river segment where the deviation convergence boundary is located is recalculated for the transmission path of each section, the generation of the unique transmission mark for the next river segment in this round is stopped, and the recalculated content of each river segment within the deviation convergence boundary is retained to generate the boundary truncated segment.

[0131] Based on the recalculation sequence for each river segment, the recalculation position and corresponding deviation convergence boundary of each river segment are extracted according to the transmission path of each cross section. The calculation node identifier of the downstream cross section at the recalculation position of the river segment is compared with the calculation node identifier of the cross section corresponding to the deviation convergence boundary in sequence from upstream to downstream. When the calculation node identifiers of the two cross sections do not match, the unique transmission mark of the current river segment in this round is retained, and the recalculation content of the current river segment continues to be carried over to the recalculation position of the next river segment. When the calculation node identifiers of the two cross sections match, the current river segment is locked as the river segment where the deviation convergence boundary is located.

[0132] When the shadow correction content of the river segment where the deviation convergence boundary is located is recalculated, the recalculated content of the river segment formed by the river segment where the deviation convergence boundary is located is retained. The current unique transmission marker of this round is no longer connected with the recalculated position of the next river segment, nor is the corresponding unique transmission marker of the next river segment generated. The recalculated content of the river segment between the recalculated position of the starting river segment and the river segment where the deviation convergence boundary is located is retained according to the transmission path of each section. The river segment position after the deviation convergence boundary is moved out and arranged continuously from upstream to downstream according to the original upstream to downstream order to generate the boundary truncation segment.

[0133] The content of river segment recalculation refers to the flood content generated after recalculating the previous river segment recalculation content or the shadow correction content of the starting river segment at the corresponding river segment recalculation location, rechecking the arrival time, and writing it into the downstream end section calculation node.

[0134] S4.5: According to the connection relationship between the calculation nodes of the preceding section and the calculation nodes of the subsequent section, the boundary truncated segment is recalculated and replaced within the boundary and the benchmark reused outside the boundary. The recalculated content of the river segment within each deviation convergence boundary replaces the corresponding benchmark forecast content, and the benchmark forecast content outside each deviation convergence boundary is retained, forming a local recalculation transmission sequence.

[0135] Furthermore, based on the boundary truncation segment, the recalculated content of the river segment within the deviation convergence boundary is extracted according to the transmission path of each section. Based on the calculation node identifier of the preceding section, the calculation node identifier of the subsequent section, and the continuous time position, the corresponding position is located in the baseline forecast content. The recalculated content of the river segment is written into the corresponding position of the baseline forecast content, replacing the original flood content carried by the same section calculation node and the same time position, while retaining the original propagation sequence and the successive relationship of the recalculated content of the river segment.

[0136] The baseline forecast content outside the boundary is extracted from the corresponding position of each deviation convergence boundary downstream and is not written into the recalculation content of the river segment. The baseline forecast content is retained according to the original sequence of cross-section calculation nodes and continuous time position. The subsequent cross-section calculation nodes of the river segment where the deviation convergence boundary is located are aligned with the preceding cross-section calculation nodes corresponding to the first baseline forecast content outside the boundary. The recalculation content of the river segment inside the boundary is continuously arranged with the baseline forecast content outside the boundary to form a local recalculation transmission sequence.

[0137] S5: Based on the local recalculation of the transmission sequence, the downstream flood transmission and confluence are carried out along the propagation topology of the large river system, and the flood forecast content of each section is updated according to the subsequent measured flow, and the intelligent transmission forecast results of the large river system flood are output.

[0138] S5.1: Based on the connection relationship between the calculation nodes of the preceding section and the calculation nodes of the succeeding section in the propagation topology of a large river system, the local recalculated propagation sequence is recursively pushed down the river segment by segment and truncated and connected at the deviation convergence boundary to generate the downstream propagation propagation segment.

[0139] Furthermore, based on the local recalculation of the transmission sequence, the preceding and subsequent cross-sectional calculation nodes in the transmission path of each cross-section are located according to the propagation topology of the large river system. The cross-sectional calculation node identifier of the downstream end of the current river segment is compared with the cross-sectional calculation node identifier of the upstream end of the next river segment. When the cross-sectional calculation node identifiers are consistent, the flood content carried by the current river segment is written into the corresponding continuous time position of the next river segment, and the flood content formed in the next river segment is continued to be transmitted to the next subsequent cross-sectional calculation node, and the process is repeated segment by segment from upstream to downstream.

[0140] When the river segment reaches the deviation convergence boundary, the recalculation of the river segment within the deviation convergence boundary is stopped and extended downstream. The first baseline forecast content retained outside the deviation convergence boundary is located. The subsequent cross-section calculation nodes of the river segment where the deviation convergence boundary is located are aligned with the preceding cross-section calculation nodes corresponding to the first baseline forecast content. The recalculated river segment within the boundary and the baseline forecast content outside the boundary are taken over according to the continuous time position. Then, they are arranged sequentially along the transmission path of each cross-section to generate the downstream transmission sequence.

[0141] It should be noted that the deviation convergence boundary refers to the first cross-section position where the baseline forecast content and the shadow correction content begin to continuously overlap along the same cross-section transmission path, which is used to define the downstream termination position of the recalculated content of the river segment.

[0142] S5.2: Merge and combine the tributary access points and multi-path confluence points in the downstream propagation segment, connect the flood content transmitted by different preceding section calculation nodes to the same subsequent section calculation node according to the arrival time, and perform continuous time continuation on the flood content after merging to generate a confluence continuation segment.

[0143] The tributary access point refers to the location where the calculation node of the preceding section of the tributary connects with the calculation node of the succeeding section of the main stream, and is used to receive the flood content transmitted by the tributary.

[0144] The multi-path convergence point refers to the location where the preceding cross-section calculation nodes corresponding to two or more cross-section transmission paths are connected to the same subsequent cross-section calculation node, which is used to aggregate the flood content transmitted by different cross-section transmission paths.

[0145] Furthermore, based on the downstream propagation segment, the tributary access locations are located item by item along the propagation topology of the large river system. The flood content transmitted by the calculation nodes of the corresponding precursor sections of the tributaries and the flood content transmitted by the calculation nodes of the corresponding precursor sections of the main stream are retained separately. The arrival times corresponding to different flood contents are extracted and arranged to the same successor section calculation node in order of arrival time. Flood contents with the same arrival time are written to the same continuous time position, and flood contents with different arrival times are written to their respective continuous time positions.

[0146] The confluence points of multiple paths are located along the transmission paths of each section. The flood content transmitted by the calculation nodes of different preceding sections is matched item by item according to the calculation node identifier and arrival time, and connected to the common successor section calculation node. The flood content that has not yet been received downstream from the previous continuous time position is retained to the next continuous time position, and the flood content after the influx is continuously arranged according to the time sequence of the successor section calculation node to generate the confluence connection segment.

[0147] S5.3: Perform flow segment updates on the confluence connection segment, write the subsequent measured flow into the corresponding cross-section calculation node and the corresponding time position, replace the flood forecast content at the same position, lock the updated position with actual measurement, and retain the flood forecast content for the position where the subsequent measured flow has not yet been obtained, and generate a cross-section forecast update list.

[0148] Furthermore, based on the confluence segment, the cross-sectional calculation nodes corresponding to the subsequent measured flow are located one by one according to the cross-sectional calculation node identifier. The measurement time carried by the subsequent measured flow is extracted, and the same time position is found in the continuous time position of the corresponding cross-sectional calculation node. If the same time position is missing, the corresponding time position is filled in according to the order of adjacent time. The subsequent measured flow is written into the corresponding cross-sectional calculation node and the corresponding time position, and the original flood forecast content at the same position is replaced.

[0149] The status of subsequent measured flow writing is checked item by item along the time sequence of each cross-section calculation node. The positions where subsequent measured flow is continuously written are locked as measured positions. The correspondence between subsequent measured flow, measurement time and cross-section calculation node is preserved, and subsequent flood forecast content is prohibited from being overwritten again. For positions where subsequent measured flow has not yet been written, the flood forecast content and original time sequence in the confluence connection segment are preserved. Then, the cross-section forecast update list is generated according to the cross-section calculation node sequence and continuous time position.

[0150] S5.4: The forecast content of the cross-section forecast update list is rolled and fixed. The latest measured flow locked by each cross-section calculation node is used as the starting point for subsequent propagation. The flood forecast content that has not been replaced by the subsequent measured flow is arranged in sequence according to the propagation topology of the large river system and the continuous time position, and the intelligent propagation forecast result of the large river system flood is output.

[0151] Furthermore, based on the cross-sectional forecast update list, the subsequent measured flows that have been measured and locked are collected according to the cross-sectional calculation node identifier, and the measured locked positions are arranged from first to last according to the measurement time; the measured locked position with the latest measurement time is extracted, and the latest locked subsequent measured flows are written into the propagation starting point of the corresponding cross-sectional calculation node. The correspondence between subsequent measured flows, measurement time and cross-sectional calculation nodes is retained, and the flood content that has been measured and locked before the propagation starting point is removed from the subsequent rolling forecast range.

[0152] The flood forecast content that has not been replaced by subsequent measured flow is extracted item by item from the propagation starting point corresponding to each cross-section calculation node and arranged in order of continuous time position. Based on the connection relationship between the preceding cross-section calculation node and the succeeding cross-section calculation node in the propagation topology of the large river system, the flood forecast content of the preceding cross-section calculation node is connected to the corresponding succeeding cross-section calculation node, and the corresponding propagation order is retained at the tributary access position and the multi-path confluence position. The latest locked subsequent measured flow and subsequent flood forecast content are fixed and arranged according to the cross-section calculation node order and continuous time position, and the intelligent propagation forecast result of the large river system flood is output.

[0153] It should be noted that the intelligent transmission forecast results of floods in large river systems refer to the forecast content formed by continuously arranging the subsequent measured flow of each cross-section according to the propagation topology of the large river system, starting from the latest measured flow locked by the calculation nodes of each cross-section. This forecast is used to understand the flow changes, arrival time and confluence process of floods propagating downstream along the main stream and tributaries, and to provide a basis for flood warning, cross-section monitoring and flood control scheduling.

[0154] This embodiment also provides a large-scale intelligent river system flood transmission forecasting system, including:

[0155] The topology construction module is used to obtain the upstream and downstream connection relationships of each section of a large river system, track the flood propagation order of the upstream and downstream connection relationships, generate the cross-sectional conduction path, and perform backtracking correction and propagation order rearrangement of the cross-sectional conduction path to form the propagation topology of the large river system.

[0156] The inflow simulation module is used to perform time-series fusion and inflow conversion of measured and predicted rainfall based on the propagation topology of large river systems, generate inflow sequences between adjacent sections, perform propagation evolution calculation and inter-section confluence regulation superposition of inflow sequences between adjacent sections, and generate real-time flood simulation sequences.

[0157] The deviation positioning module is used to perform measured flow history reversal on the real-time flood simulation sequence, generate shadow correction content and baseline forecast content, perform same-path dual-track simulation playback and cross-section alignment on the baseline forecast content and shadow correction content, lock the deviation convergence boundary, and generate the minimum recalculation range.

[0158] The local recalculation module is used to recalculate the shadow correction content segment by segment based on the minimum recalculation range, generate a unique transmission marker for each segment in this round, and limit each segment to only receive the recalculation content generated in this round from the previous segment, forming a segment-by-segment recalculation sequence. The segment-by-segment recalculation sequence is truncated by the deviation convergence boundary and connected with the baseline forecast content retained outside the boundary to form a local recalculation transmission sequence.

[0159] The transmission forecast module is used to perform downstream flood transmission and confluence continuation along the propagation topology of large river systems based on the local recalculation of the transmission sequence, and update the flood forecast content of each section according to the subsequent measured flow, and output the intelligent transmission forecast results of large river system floods.

[0160] In summary, this invention achieves adaptive convergence of the influence range of measured flow correction by performing simultaneous dual-track simulation playback and cross-section alignment on the baseline forecast content and the shadow correction content, quantifying the flow deviation and arrival time deviation of each cross-section, locking the deviation convergence boundary where the two contents begin to continuously converge, and retaining the river segments where the deviation has not converged and the river segments that have converged outside the boundary. This provides accurate start and end positions for subsequent recalculation of each river segment and boundary truncation, reduces invalid repeated calculations, and maintains the continuity and consistency of the multi-path transmission process in large river systems.

[0161] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for intelligent transmission forecasting of floods in large river systems, characterized in that, include: The upstream and downstream connections of each section of a large river system are obtained, the flood propagation order of the upstream and downstream connections is tracked, the cross-sectional conduction paths are generated, and the cross-sectional conduction paths are backtracked, corrected, and rearranged to form the propagation topology of the large river system. Based on the propagation topology of large river systems, time-series fusion and inflow conversion of measured and predicted rainfall are performed to generate inflow sequences between adjacent sections. Propagation evolution calculation and inter-section confluence regulation are performed on the inflow sequences between adjacent sections to generate real-time flood simulation sequences. The measured flow history of the real-time flood simulation sequence is backtracked to generate shadow correction content and baseline forecast content. The baseline forecast content and shadow correction content are simulated and replayed along the same path and aligned section by section to lock the deviation convergence boundary and generate the minimum recalculation range. Based on the minimum recalculation range, the shadow correction content is recalculated segment by segment, generating a unique transmission marker for each segment in this round, and limiting each segment to only receive the recalculated content generated in this round from the previous segment, forming a segment-by-segment recalculation sequence. The segment-by-segment recalculation sequence is truncated by the deviation convergence boundary and connected with the baseline forecast content retained outside the boundary to form a local recalculation transmission sequence. Based on the local recalculation of the transmission sequence, the downstream flood transmission and confluence are carried out along the propagation topology of the large river system, and the flood forecast content of each section is updated according to the subsequent measured flow, outputting the intelligent transmission forecast results of the large river system flood.

2. The intelligent transmission forecasting method for large-scale river system floods as described in claim 1, characterized in that, The generated cross-sectional conduction path is as follows: The upstream and downstream connection relationships of each section of a large river system are obtained and converted into node connection entries. Each section is converted into a section calculation node. The section calculation node identifier is unified according to the section location. Abnormal node connection entries are processed separately and a section connection index table is generated. Based on the cross-section connection index table, starting from the cross-section calculation node without an upstream predecessor, the subsequent cross-section calculation nodes are visited level by level according to the connection direction from upstream to downstream. The order of connection between cross-section calculation nodes is recorded, and the corresponding propagation branch relationship is preserved at the tributary access position and the multi-path convergence position to generate the cross-section propagation path.

3. The intelligent transmission forecasting method for large-scale river system floods as described in claim 2, characterized in that, The formation of a large river system propagation topology is as follows: Perform backtracking correction on the cross-section propagation path according to the direction from the end cross-section to the starting cross-section, verify the connection status of adjacent cross-section calculation nodes, merge the path positions that repeatedly pass through the same cross-section calculation node, correct the propagation direction of the path positions that are accessed in reverse, and add the corresponding adjacent cross-section calculation nodes to the connection interruption positions according to the cross-section connection index table to generate the backtracking correction path. The propagation order of the retrospective correction path is rearranged according to the propagation hierarchy of the river system flood. Each cross-section calculation node is written into a unique topological position, and the preceding cross-section calculation nodes and the subsequent cross-section calculation nodes are encapsulated into the corresponding topological positions to form a large-scale river system propagation topology.

4. The intelligent transmission forecasting method for large-scale river system floods as described in claim 3, characterized in that, The generation of the real-time flood simulation sequence is as follows: Based on the propagation topology of large river systems, measured rainfall and estimated rainfall are aggregated to the corresponding cross-section calculation nodes and arranged in order. Measured rainfall and estimated rainfall are distinguished at time-overlapping locations, and the corresponding time order is retained at time-missing locations to generate a rainfall time-series aligned sequence. Using the measured rainfall coverage period as the first segment and the estimated rainfall carryover period as the second segment, the rainfall time series alignment sequence is fused, and the fused continuous rainfall time series is solidified to the corresponding cross-section calculation node to generate a rainfall time series fusion sequence. Based on the rainfall time series fusion sequence, the rainfall content of each cross-section calculation node is converted into runoff and inflow. The inflow content generated by the preceding cross-section calculation node and the corresponding generation time are linked to the propagation path of the corresponding subsequent cross-section calculation node. The inflow content of each tributary is retained at the tributary access point, and the inflow content corresponding to different propagation paths is retained at the multi-path confluence point, generating the inflow sequence between adjacent cross-sections. Based on the inflow sequence between adjacent cross sections, the propagation evolution calculation and inter-section confluence regulation and storage are superimposed on the inflow of water between adjacent cross sections. According to the connection relationship between the calculation nodes of the preceding cross section and the calculation nodes of the subsequent cross section in the propagation topology of the large river system, the inflow content that can be received is written into the calculation node of the subsequent cross section, and the remaining inflow content is retained to the next time position to generate a real-time flood simulation sequence.

5. The intelligent transmission forecasting method for large-scale river system floods as described in claim 4, characterized in that, The generation of shadow correction content and baseline forecast content are as follows: Based on the real-time flood simulation sequence, the cross-sectional calculation node and historical time position corresponding to the measured flow are located, the measured flow of the real-time flood simulation sequence is historically reversed, and the historical time position is locked as the common evolution starting point of the baseline forecast content and the shadow correction content, generating the historical reversed position of the measured flow. Using the historical return position of the measured flow as the boundary, the continuous simulation content after the boundary position in the real-time flood simulation sequence is retained as a baseline, and the shadow replay is performed starting from the measured flow written at the boundary position, generating baseline forecast content and shadow correction content respectively.

6. The intelligent transmission forecasting method for large-scale river system floods as described in claim 5, characterized in that, The specific steps for generating the minimum recalculation range are as follows: Based on the baseline forecast content and the shadow correction content, a dual-track simulation playback is performed along the same propagation path in the propagation topology of a large river system. The dual-track simulation playback refers to placing the baseline forecast content and the shadow correction content in independent bearing positions, and synchronously pushing them downstream along the same propagation path and the same time sequence, respectively retaining the evolution status of the baseline forecast content and the shadow correction content at each cross-section calculation node, and generating a dual-track playback sequence. The two playback contents in the dual-track playback sequence are compared section by section. The sections with deviations are continuously connected, and the first section where the two playback contents begin to continuously fit together is locked as the corresponding deviation convergence boundary to generate a deviation alignment sequence. The start and end positions of the recalculation of the deviation alignment sequence are defined. The historical backtracking position of the measured flow is taken as the starting point, and the deviation convergence boundary corresponding to the transmission path of each section is taken as the ending point. The non-converged positions of the deviation between the starting point and each ending point are continuously retained, and the section calculation nodes outside the recalculation termination position are removed to generate the minimum recalculation range.

7. The intelligent transmission forecasting method for large-scale river system floods as described in claim 6, characterized in that, The formation of the segment-by-segment recalculation sequence is as follows: Extract the cross-section content covered by the minimum recalculation range from the shadow correction content. Based on the connection relationship between the calculation nodes of the preceding cross-section and the calculation nodes of the subsequent cross-section in the large river system propagation topology, divide the cross-section content to the recalculation position of the river segment corresponding to the adjacent cross-section, and arrange the recalculation positions of each river segment in the order from upstream to downstream to generate a list of recalculated river segments. The recalculated river segment list is recalculated and the recalculation round is assigned to generate a recalculation round identifier for this round. This recalculation round identifier is then written into the recalculation position of each river segment in the recalculated river segment list. The recalculation round identifier, the historical backtracking position of the measured flow, and the recalculation position of the starting river segment are merged and coded to obtain a unique transmission marker for this round corresponding to the starting river segment. The recalculation round identifier, the unique transmission marker, the recalculation position of the starting river segment, and the corresponding shadow correction content are written in the same position to generate a unique transmission marker table for this round. Based on the unique transmission marker table for this round, the shadow correction content is recalculated segment by segment. Only the recalculated content of the previous segment carrying the corresponding unique transmission marker for this round is accepted. The current unique transmission marker for this round and the recalculated position of the next segment are recursively encoded to form a segment-by-segment recalculation sequence.

8. The intelligent transmission forecasting method for large-scale river system floods as described in claim 7, characterized in that, The formation of the locally recalculated propagation sequence is as follows: The deviation convergence boundary is truncated for the recalculated sequence of each river segment. After the river segment where the deviation convergence boundary is located is recalculated for the propagation path of each section, the generation of the unique propagation mark for the next river segment in this round is stopped, and the recalculated content of each river segment within the deviation convergence boundary is retained to generate the boundary truncated segment. Based on the connection between the calculation nodes of the preceding section and the calculation nodes of the subsequent section, the boundary truncated segments are recalculated and replaced within the boundary and the benchmark reused outside the boundary. The recalculated content of the river segment within each deviation convergence boundary replaces the corresponding benchmark forecast content, while the benchmark forecast content outside each deviation convergence boundary is retained, forming a local recalculation transmission sequence.

9. The intelligent transmission forecasting method for large-scale river system floods as described in claim 8, characterized in that, The output of intelligent transmission forecast results for large-scale river system floods is as follows: Based on the connection relationship between the calculation nodes of the preceding section and the calculation nodes of the succeeding section in the propagation topology of large river systems, the local recalculated propagation sequence is recursively pushed down the river segment by segment and truncated and connected at the deviation convergence boundary to generate the downstream propagation propagation segment. The tributary access points and multi-path confluence points in the downstream propagation segment are merged and merged. The flood content transmitted by different preceding section calculation nodes is connected to the same subsequent section calculation node according to the arrival time. The flood content after merging is continuously spliced ​​in time to generate the confluence splice segment. The flow segment of the confluence connection segment is updated in segments. The subsequent measured flow is written into the corresponding cross-section calculation node and the corresponding time position, replacing the flood forecast content at the same position. The updated position is locked by actual measurement, and the flood forecast content at the position where the subsequent measured flow has not yet been obtained is retained, generating a cross-section forecast update list. The forecast content of the cross-section forecast update list is rolled and fixed. The latest measured flow locked by the calculation node of each cross section is used as the starting point for subsequent propagation. The flood forecast content that has not been replaced by the subsequent measured flow is arranged in sequence according to the propagation topology of the large river system and the continuous time position, and the intelligent propagation forecast result of the large river system flood is output.

10. A large-scale river system flood intelligent transmission forecasting system, based on the large-scale river system flood intelligent transmission forecasting method according to any one of claims 1 to 9, characterized in that, include: The topology construction module is used to obtain the upstream and downstream connection relationships of each section of a large river system, track the flood propagation order of the upstream and downstream connection relationships, generate the cross-sectional conduction path, and perform backtracking correction and propagation order rearrangement of the cross-sectional conduction path to form the propagation topology of the large river system. The inflow simulation module is used to perform time-series fusion and inflow conversion of measured and predicted rainfall based on the propagation topology of large river systems, generate inflow sequences between adjacent sections, perform propagation evolution calculation and inter-section confluence regulation superposition of inflow sequences between adjacent sections, and generate real-time flood simulation sequences. The deviation positioning module is used to perform measured flow history reversal on the real-time flood simulation sequence, generate shadow correction content and baseline forecast content, perform same-path dual-track simulation playback and cross-section alignment on the baseline forecast content and shadow correction content, lock the deviation convergence boundary, and generate the minimum recalculation range. The local recalculation module is used to recalculate the shadow correction content segment by segment based on the minimum recalculation range, generate a unique transmission marker for each segment in this round, and limit each segment to only receive the recalculation content generated in this round from the previous segment, forming a segment-by-segment recalculation sequence. The segment-by-segment recalculation sequence is truncated by the deviation convergence boundary and connected with the baseline forecast content retained outside the boundary to form a local recalculation transmission sequence. The transmission forecast module is used to perform downstream flood transmission and confluence continuation along the propagation topology of large river systems based on the local recalculation of the transmission sequence, and update the flood forecast content of each section according to the subsequent measured flow, and output the intelligent transmission forecast results of large river system floods.

Images