A 3D Flood Simulation Method and System Based on Scene Generation and Parameter Control
By constructing a flood simulation scenario library and parameter control, generating candidate scenarios and performing incremental updates, the problem of low efficiency in multi-scenario simulation under complex working conditions in existing technologies has been solved, and more efficient three-dimensional flood simulation has been achieved.
Patent Information
- Application Number
- CN202611597362.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-10-15
- Publication Date
- 2026-11-10
AI Technical Summary
Existing flood simulation technologies rely on pre-set simulation conditions operated manually, which makes it difficult to adapt to multi-scenario simulations under complex conditions. Furthermore, parameters need to be repeatedly configured when conditions change, resulting in low efficiency and insufficient resource utilization.
By constructing a scenario library, generating multiple candidate scenarios, dividing them into fixed and adjustable parameters, determining the target scenario using scenario similarity, and incrementally updating and correcting the adjustable parameters, three-dimensional flood simulation is achieved.
It improves the scene adaptability and processing efficiency of 3D flood simulation, reduces redundant processing and parameter configuration, and enhances scene coverage under complex working conditions.
Smart Images

Figure CN123069936A_ABST
Abstract
Description
Technical Field
[0001] This disclosure relates to the field of data processing technology, and more specifically, to a method and system for three-dimensional flood simulation based on scene generation and parameter control. Background Technology
[0002] With the increasing frequency and complexity of extreme rainfall events, urban flooding disasters are becoming more frequent and complex. To assist in flood control and drainage analysis, existing technologies typically use flood models to simulate the flood evolution process under different rainfall conditions to obtain corresponding projection results.
[0003] However, existing flood simulation methods typically rely on pre-set simulation scenarios. Scenario construction and parameter adjustment are heavily dependent on manual operation, limiting the range of simulation scenarios they can cover and making it difficult to adapt to the multi-scenario simulation needs under complex conditions. Furthermore, when simulation conditions change, relevant parameters usually need to be reconfigured and recalculated, easily leading to repetitive processing and low efficiency in continuous parameter tuning and multi-scenario trial calculations. In addition, different simulation scenarios are relatively independent, making it difficult to fully utilize existing simulation resources, further increasing the processing workload under complex flood scenarios. Therefore, existing technologies still need improvement in terms of scenario adaptability and processing efficiency in flood simulation. Summary of the Invention
[0004] The purpose of this disclosure is to provide a flood three-dimensional simulation method, a flood three-dimensional simulation system, an electronic device, and a computer-readable storage medium based on scene generation and parameter control. This method can improve the scene adaptability and processing efficiency of flood three-dimensional simulation by constructing a simulation scene library, incrementally updating the adjustable parameters that need to be adjusted, determining the target simulation scene based on scene similarity, and correcting the adjustable parameters in the target simulation scene.
[0005] Other features and advantages of this disclosure will become apparent from the following detailed description, or may be learned in part by practice of this disclosure.
[0006] According to a first aspect of the present disclosure, a method for three-dimensional flood simulation based on scene generation and parameter control is provided, comprising: acquiring flood simulation requirements; performing semantic parsing on the flood simulation requirements to generate structured simulation instructions; combining and filtering simulation parameters based on regional rainstorm frequency statistics, hydrophysical constraints, and parameter combination rules to generate multiple candidate simulation scenarios, and constructing a simulation scenario library based on the multiple candidate simulation scenarios; dividing the simulation parameters into fixed parameters and adjustable parameters; determining the adjustable parameters to be adjusted based on the structured simulation instructions while keeping the fixed parameters unchanged, and incrementally updating the adjustable parameters to be adjusted; extracting the current scene features corresponding to the flood simulation requirements, determining the scene similarity between the current scene features and each existing scene feature in the simulation scenario library, and determining the target simulation scenario based on the scene similarity; correcting the adjustable parameters in the target simulation scenario based on the differences between the incrementally updated adjustable parameters and the corresponding parameters in the target simulation scenario; and performing three-dimensional flood simulation based on the parameter-corrected target simulation scenario to obtain the simulation results.
[0007] In an example embodiment of this disclosure, based on the aforementioned scheme, the step of obtaining flood simulation requirements and performing semantic parsing on the flood simulation requirements to generate structured simulation instructions includes: obtaining the flood simulation requirements; performing semantic parsing on the flood simulation requirements to determine the corresponding rainfall parameters, river hydrological parameters, engineering scheduling parameters, pipeline hydraulic parameters, and waterlogging risk parameters; and encapsulating the determined parameters in a structured manner according to a preset instruction format to obtain the structured simulation instructions.
[0008] In an example embodiment of this disclosure, based on the aforementioned scheme, the semantic parsing of the flood simulation requirements includes: acquiring a flood-labeled dataset, which includes the correspondence between natural language questions and standard simulation parameters; adapting the semantic parsing model to the domain based on the flood-labeled dataset; retrieving flood cases, engineering scheduling procedures, hydrological parameter thresholds, and simulation condition constraints that match the flood simulation requirements from a flood domain knowledge base; inputting the retrieved content and the flood simulation requirements into the domain-adapted semantic parsing model, and determining the corresponding standard simulation parameters according to the output format constraints.
[0009] In an example embodiment of this disclosure, based on the aforementioned scheme, the step of combining and filtering the projection parameters based on regional rainstorm frequency statistics, hydrophysical constraints, and parameter combination rules to generate multiple candidate projection scenarios includes: determining the range of rainfall parameters corresponding to different rainstorm return periods based on the regional rainstorm frequency statistics; combining rainfall intensity, rainfall duration, rainstorm return period, spatial distribution of rainfall patterns, engineering scheduling status, spatial projection range, and simulation duration within the rainfall parameter range to obtain multiple candidate parameter combinations; verifying the multiple candidate parameter combinations based on the hydrophysical constraints; removing candidate parameter combinations that have any of the following conditions from the verified candidate parameter combinations according to the parameter combination rules; and generating multiple candidate projection scenarios based on the remaining candidate parameter combinations.
[0010] In an example embodiment of this disclosure, based on the aforementioned scheme, the step of generating multiple candidate simulation scenarios according to the remaining candidate parameter combinations includes: dividing the candidate simulation scenarios into conventional gradient rainstorm scenarios, short-term extreme heavy rainfall scenarios, multi-condition superimposed composite scenarios, and excessive extreme flood scenarios according to the rainfall parameters and engineering operation status corresponding to each candidate parameter combination. Constructing a scenario library based on multiple candidate scenarios includes: associating each candidate scenario with its corresponding standard parameter configuration, initial hydraulic condition, boundary constraints, and linkage calculation rules; classifying, encoding, and storing the associated candidate scenarios in a structured manner to construct the scenario library.
[0011] In an example embodiment of this disclosure, based on the foregoing scheme, determining the adjustable parameters to be adjusted according to the structured deduction instructions includes: obtaining the previous round of deduction state, the previous round of deduction state including the fixed parameters and adjustable parameters used in the previous round of deduction; determining the target state for the current round according to the structured deduction instructions; comparing the target state for the current round with the previous round of deduction state to determine the corresponding parameter increment; if the parameter increment is zero, using the previous round of deduction results; if the parameter increment is not zero, determining the adjustable parameters to be adjusted according to the parameter increment.
[0012] In an example embodiment of this disclosure, based on the aforementioned scheme, the incremental update of the adjustable parameters to be adjusted includes: determining the pipeline topology, underlying surface attributes, terrain data, and water system boundaries as the fixed parameters; determining the rainfall parameters, engineering scheduling parameters, risk assessment parameters, and simulation time parameters as the adjustable parameters; when the structured simulation instruction does not contain adjustment information for the fixed parameters, maintaining the corresponding values of the fixed parameters in the current simulation state in the previous simulation state; incrementally updating the adjustable parameters to be adjusted according to the parameters, and maintaining the corresponding values of the unchanged adjustable parameters in the previous simulation state.
[0013] In an example embodiment of this disclosure, based on the aforementioned scheme, after updating the adjustable parameters to be adjusted according to the parameter increment, the method further includes: performing parameter over-limit judgment on the updated adjustable parameters to determine whether the updated adjustable parameters are within the corresponding allowable range; performing scheduling logic verification based on the project operation status and project scheduling parameters to determine whether the project scheduling parameters match the project operation status; performing parameter coupling verification on the updated adjustable parameters and fixed parameters based on the hydrophysical constraints; and, if the parameter over-limit judgment, the scheduling logic verification, and the parameter coupling verification all pass, using the updated adjustable parameters for scenario matching and parameter correction.
[0014] In an example embodiment of this disclosure, based on the aforementioned scheme, determining the scene similarity between the current scene features and the existing scene features in the simulation scene library includes: extracting rainfall features, spatial simulation range, engineering scheduling conditions, simulation time scale, and risk judgment threshold from the flood simulation requirements to generate a current scene feature vector; obtaining the existing scene feature vectors corresponding to each of the existing scenes; calculating the weighted feature distance between the current scene feature vector and each of the existing scene feature vectors according to the feature weights corresponding to each feature dimension; and determining the corresponding scene similarity based on the weighted feature distance and the maximum feature distance.
[0015] In an example embodiment of this disclosure, based on the aforementioned scheme, determining the target derivation scene according to the scene similarity includes: sorting multiple existing scenes according to the scene similarity; determining the existing scene with the highest scene similarity as the target derivation scene; and correcting the adjustable parameters in the target derivation scene according to the difference between the incrementally updated adjustable parameters and the corresponding parameters in the target derivation scene, including: comparing the adjustable parameters in the target derivation scene with the incrementally updated adjustable parameters to determine the adjustable parameters with differences; keeping the derivation parameters in the target derivation scene without differences unchanged, and correcting the adjustable parameters with differences according to the incrementally updated adjustable parameters to obtain the target derivation scene with corrected parameters.
[0016] In an example embodiment of this disclosure, based on the aforementioned scheme, the step of performing a three-dimensional flood simulation based on the target simulation scenario after parameter correction to obtain the simulation result includes: generating a corresponding rainfall process according to the rainfall parameters in the target simulation scenario after parameter correction; applying corresponding engineering scheduling conditions according to the engineering scheduling parameters; performing pipeline hydraulic calculations based on the rainfall process and the engineering scheduling conditions to obtain the pipeline node flow rate and node water level; performing surface waterlogging evolution calculations based on the pipeline hydraulic calculation results to obtain the surface water depth and inundation range; and performing risk assessment based on the pipeline node flow rate, the node water level, the surface water depth, and the inundation range to obtain the simulation result.
[0017] In an example embodiment of this disclosure, based on the aforementioned scheme, the step of determining the risk based on the pipeline node flow rate, the node water level, the surface water depth, and the inundation range to obtain the projection result includes: associating the pipeline node flow rate, the node water level, the surface water depth, and the inundation range according to the same time node to obtain time-series flood status data corresponding to multiple time nodes; determining the risk of each time-series flood status data according to a preset risk determination threshold to obtain the flood risk status corresponding to each time node; determining the flood peak time based on the multiple time-series flood status data, and obtaining the node peak flow rate, node maximum water level, regional maximum water depth, and inundation range corresponding to the flood peak time; and obtaining the projection result based on the flood risk status corresponding to each time node and the node peak flow rate, node maximum water level, regional maximum water depth, and inundation range corresponding to the flood peak time.
[0018] In an example embodiment of this disclosure, based on the aforementioned scheme, the step of performing risk assessment on each of the time-series flood status data according to a preset risk assessment threshold to obtain the flood risk status corresponding to each of the time nodes includes: obtaining the flow threshold, water level threshold, and risk classification threshold corresponding to the water accumulation depth of the pipeline node; for each of the time nodes, comparing the flow of the pipeline node with the corresponding flow threshold, comparing the water level of the node with the corresponding water level threshold, and comparing the surface water accumulation depth with the corresponding risk classification threshold to obtain the flow exceeding limit status, water level exceeding limit status, and water accumulation risk level corresponding to each of the time nodes; and determining the flood risk status corresponding to each of the time nodes based on the flow exceeding limit status, the water level exceeding limit status, and the water accumulation risk level. The process of obtaining the projection results based on the flood risk status corresponding to each time node and the peak flow, highest water level, maximum water depth, and inundation range corresponding to the flood peak time includes: determining the duration of flow exceeding the limit based on the time nodes continuously in the flow exceeding the limit state, and determining the duration of water level exceeding the limit based on the time nodes continuously in the water level exceeding the limit state; determining the magnitude of flow exceeding the limit based on the peak flow and the corresponding flow threshold, and determining the magnitude of water level exceeding the limit based on the highest water level and the corresponding water level threshold; and generating the projection results based on the flood risk status corresponding to each time node, the duration of flow exceeding the limit, the duration of water level exceeding the limit, the magnitude of flow exceeding the limit, the magnitude of water level exceeding the limit, the maximum water depth, and the inundation range.
[0019] In an example embodiment of this disclosure, based on the aforementioned scheme, after obtaining the simulation results, the method further includes: associating the scene features, simulation parameters, and simulation results corresponding to this three-dimensional flood simulation to form a simulation scene record; storing the simulation scene record in the simulation scene library, and recording the corresponding simulation rounds and parameter differences; and when obtaining new flood simulation requirements in the future, using the simulation scene record as an existing scene to participate in scene similarity calculation.
[0020] In an example embodiment of this disclosure, based on the aforementioned scheme, after obtaining the deduction result, the method further includes: associating the deduction result with knowledge objects in a flood knowledge graph; performing semantic mapping on the deduction scenario, deduction parameters, and entities, relationships, and rules in the flood knowledge graph, respectively, and calibrating the corresponding parameter thresholds and resolving conflicting information; constraining the deduction result based on the disaster-causing mechanism, risk transmission path, engineering scheduling boundary, and parameter thresholds in the flood knowledge graph; and inputting the constrained deduction result into a semantic reasoning model to perform flood causal analysis, process correlation analysis, and risk transmission analysis to obtain flood risk analysis results.
[0021] According to a second aspect of the present disclosure, a three-dimensional flood simulation system based on scene generation and parameter control is provided. The system includes: a demand parsing module, used to acquire flood simulation demands, perform semantic parsing on the flood simulation demands, and generate structured simulation instructions; a scene generation module, used to combine and filter simulation parameters based on regional rainstorm frequency statistics, hydrophysical constraints, and parameter combination rules to generate multiple candidate simulation scenes, and construct a simulation scene library based on the multiple candidate simulation scenes; and a parameter update module, used to divide the simulation parameters into fixed parameters and adjustable parameters, and, while keeping the fixed parameters unchanged, adjust the simulation parameters according to the results... The structured simulation command determines the adjustable parameters that need to be adjusted and performs incremental updates on the adjustable parameters; the scene matching module is used to extract the current scene features corresponding to the flood simulation requirements, determine the scene similarity between the current scene features and each existing scene feature in the simulation scene library, and determine the target simulation scene based on the scene similarity; the parameter correction module is used to correct the adjustable parameters in the target simulation scene based on the differences between the incrementally updated adjustable parameters and the corresponding parameters in the target simulation scene; the three-dimensional simulation module is used to perform three-dimensional flood simulation based on the parameter-corrected target simulation scene to obtain the simulation results.
[0022] According to a third aspect of the present disclosure, an electronic device is provided, comprising: a processor; and a memory storing computer-readable instructions, which, when executed by the processor, implement the flood three-dimensional simulation method based on scene generation and parameter control as described in the first aspect.
[0023] According to a fourth aspect of the present disclosure, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the flood three-dimensional simulation method based on scene generation and parameter control as described in the first aspect.
[0024] The technical solutions provided in this disclosure can include the following beneficial effects: According to the flood 3D simulation method based on scene generation and parameter control in this example embodiment, on the one hand, the flood simulation requirements are semantically parsed to form structured simulation instructions. Combined with regional rainstorm frequency statistics, hydrophysical constraints, and parameter combination rules, the simulation parameters are combined and filtered to enable the generated candidate simulation scenarios to cover different flood conditions and form a callable simulation scenario library. This reduces the reliance on manually preset single simulation conditions and improves the scene coverage under complex flood conditions. On the other hand, by dividing the simulation parameters into fixed parameters and adjustable parameters, only the adjustable parameters that need to be adjusted are incrementally updated when the fixed parameters remain unchanged. This eliminates the need to repeatedly process the unchanged parameters when the simulation conditions change, thereby reducing the repetitive configuration and calculation in the continuous parameter tuning and multi-scenario trial calculation process. Furthermore, by extracting the current scene features and comparing the scene similarity with the existing scenes in the simulation scenario library, the target simulation scene is determined. Only the adjustable parameters that differ from the target simulation scene are corrected, so that existing similar simulation scenes can continue to be used, reducing the amount of processing for repeatedly constructing scenes between different simulation tasks. Therefore, it is possible to reduce redundant processing while adapting to different flood simulation needs, thereby improving the scenario adaptability and processing efficiency of three-dimensional flood simulation.
[0025] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description
[0026] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure. It is obvious that the drawings described below are merely some embodiments of this disclosure, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.
[0027] Figure 1 The illustration shows a flowchart of a three-dimensional flood simulation method based on scene generation and parameter control according to some embodiments of the present disclosure.
[0028] Figure 2 The illustration shows a schematic diagram of a three-dimensional flood simulation process according to some embodiments of the present disclosure.
[0029] Figure 3 The illustration shows a schematic diagram of the flood evolution process in a 5-year return period rainfall scenario according to some embodiments of the present disclosure.
[0030] Figure 4 The diagram illustrates the peak value projection results for a 110mm rainfall condition according to some embodiments of the present disclosure.
[0031] Figure 5 The illustration shows a schematic diagram of the water receding process under 130mm rainfall conditions according to some embodiments of the present disclosure.
[0032] Figure 6 The diagram illustrates a block diagram of a flood three-dimensional simulation system based on scene generation and parameter control according to some embodiments of the present disclosure.
[0033] Figure 7 The schematic diagram illustrates the structural schematic of a computer system of an electronic device according to some embodiments of the present disclosure.
[0034] In the accompanying drawings, the same or corresponding reference numerals indicate the same or corresponding parts. Detailed Implementation
[0035] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this specification. Rather, they are merely examples of systems and methods consistent with some aspects of this specification as detailed in the appended claims.
[0036] The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to be limiting of this specification. The singular forms “a,” “the,” and “the” as used in this specification and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.
[0037] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this disclosure will be more thorough and complete, and will fully convey the concept of the exemplary embodiments to those skilled in the art.
[0038] Furthermore, the described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. Numerous specific details are provided in the following description to give a thorough understanding of embodiments of this disclosure. However, those skilled in the art will recognize that the technical solutions of this disclosure can be practiced without one or more of the specific details, or other methods, components, systems, steps, etc., can be employed. In other instances, well-known methods, systems, implementations, or operations are not shown or described in detail to avoid obscuring various aspects of this disclosure.
[0039] Furthermore, the accompanying drawings are for illustrative purposes only and are not necessarily drawn to scale. The block diagrams shown in the drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor systems and / or microcontroller systems.
[0040] In this exemplary embodiment, a method for three-dimensional flood simulation based on scene generation and parameter control is first provided. This method can be implemented by a three-dimensional flood simulation system based on scene generation and parameter control. Figure 1 The illustration schematically shows a flowchart of a three-dimensional flood simulation method based on scene generation and parameter control according to some embodiments of the present disclosure. (Reference) Figure 1 As shown, the flood 3D simulation method based on scene generation and parameter control may include the following steps: Step S110: Obtain flood simulation requirements, perform semantic parsing on the flood simulation requirements, and generate structured simulation instructions; Step S120: Based on the regional rainstorm frequency statistics, hydrophysical constraints, and parameter combination rules, the simulation parameters are combined and filtered to generate multiple candidate simulation scenarios, and a simulation scenario library is constructed based on the multiple candidate simulation scenarios. Step S130: Divide the inference parameters into fixed parameters and adjustable parameters. With the fixed parameters remaining unchanged, determine the adjustable parameters that need to be adjusted according to the structured inference instructions, and perform incremental updates on the adjustable parameters that need to be adjusted. Step S140: Extract the current scene features corresponding to the flood simulation requirements, determine the scene similarity between the current scene features and the existing scene features in the simulation scene library, and determine the target simulation scene based on the scene similarity. Step S150: Based on the difference between the incrementally updated adjustable parameters and the corresponding parameters in the target simulation scenario, the adjustable parameters in the target simulation scenario are corrected. Step S160: Perform a three-dimensional flood simulation based on the target simulation scenario after parameter correction to obtain the simulation results.
[0041] According to the flood 3D simulation method based on scene generation and parameter control in this example embodiment, on the one hand, the flood simulation requirements are semantically parsed to form structured simulation instructions. Combined with regional rainstorm frequency statistics, hydrophysical constraints, and parameter combination rules, the simulation parameters are combined and filtered to enable the generated candidate simulation scenarios to cover different flood conditions and form a callable simulation scenario library. This reduces the reliance on manually preset single simulation conditions and improves the scene coverage under complex flood conditions. On the other hand, by dividing the simulation parameters into fixed parameters and adjustable parameters, only the adjustable parameters that need to be adjusted are incrementally updated when the fixed parameters remain unchanged. This eliminates the need to repeatedly process the unchanged parameters when the simulation conditions change, thereby reducing the repetitive configuration and calculation in the continuous parameter tuning and multi-scenario trial calculation process. Furthermore, by extracting the current scene features and comparing the scene similarity with the existing scenes in the simulation scenario library, the target simulation scene is determined. Only the adjustable parameters that differ from the target simulation scene are corrected, so that existing similar simulation scenes can continue to be used, reducing the amount of processing for repeatedly constructing scenes between different simulation tasks. Therefore, it is possible to reduce redundant processing while adapting to different flood simulation needs, thereby improving the scenario adaptability and processing efficiency of three-dimensional flood simulation.
[0042] The following will further explain the three-dimensional flood simulation method based on scene generation and parameter control in this example embodiment.
[0043] In step S110, the flood simulation requirements are obtained, the flood simulation requirements are semantically parsed, and structured simulation instructions are generated.
[0044] The flood simulation request can represent the simulation conditions or objectives proposed when conducting flood evolution analysis for a target area. It can be input by the user or generated by business requests related to flood simulation. For example, the flood simulation request can be input in natural language or other predefined parsable formats, and its content can include the simulation area, rainfall conditions, simulation period, engineering operation conditions, and the flood status to be obtained.
[0045] For example, flood simulation requirements could be: "to simulate the flood evolution process of the target area under a 5-year return period rainfall condition to determine the inundation situation at different time stages"; or "to conduct a flood simulation of the target area under a 110mm rainfall condition to determine the flow rate, water level, and surface water accumulation at the pipeline nodes during the peak flood stage"; or "to conduct a flood simulation of the target area with an average rainfall of 35mm in the next 3 hours to determine the changes in water accumulation during the corresponding period." In the case of continuous flood simulations, the flood simulation requirements could also be adjustments based on the conditions of the previous flood simulation. For example, if the previous flood simulation used a 110mm rainfall condition, the current flood simulation requirement could be: "to adjust the rainfall to 130mm and then conduct a new flood simulation"; or, under existing simulation conditions, to propose increasing rainfall intensity, extending the simulation duration, or changing the current engineering operating conditions. Structured simulation instructions can represent instruction data that is formed by organizing the simulation information in the flood simulation requirements according to a preset data structure, and can be used for subsequent scenario generation and parameter control.
[0046] Specifically, flood simulation requirements input in natural language or other parsable formats can be acquired. Information describing the simulation objects, conditions, and objectives within these requirements is identified, and this identified information is converted into simulation parameters corresponding to the 3D flood simulation. These parameters are then organized according to a unified data structure to generate structured simulation instructions. This allows flood simulation requirements in different expressions to be converted into a unified simulation data format, reducing parameter configuration discrepancies caused by differences in requirement expression and improving the consistency of subsequent simulation processing.
[0047] In step S120, based on the regional rainstorm frequency statistics, hydrophysical constraints, and parameter combination rules, the simulation parameters are combined and filtered to generate multiple candidate simulation scenarios, and a simulation scenario library is constructed based on the multiple candidate simulation scenarios.
[0048] Among them, the regional rainstorm frequency statistics can be represented as statistical results obtained from historical rainfall data or existing rainstorm statistics of the target area, used to characterize different rainfall occurrence characteristics.
[0049] Hydrophysical constraints can represent conditions that restrict the inference parameters or combinations thereof based on the physical relationships between various hydrological elements during the formation and evolution of floods. Parameter combination rules can represent rules used to limit the combinations of multiple inference parameters to form effective inference conditions. Candidate inference scenarios can represent flood conditions jointly characterized by a set of interrelated inference parameters and satisfying corresponding inference conditions. Inference scenario library can represent a dataset used to store multiple inference scenarios and their corresponding inference information.
[0050] Specifically, the applicable value range or variation range of the simulation parameters can be determined based on the statistical results of regional rainstorm frequency, and different simulation parameters can be combined within the corresponding range. For the obtained parameter combinations, they can be judged according to hydrophysical constraints and parameter combination rules, eliminating parameter combinations that do not meet the corresponding constraints or combination relationships, and forming candidate simulation scenarios for the remaining parameter combinations. Subsequently, multiple candidate simulation scenarios are uniformly organized and stored to form a simulation scenario library. In this way, while reducing the manual configuration of simulation conditions one by one, the formed simulation scenarios can balance the number of scenarios and the rationality of parameters, improving the scenario coverage under different flood conditions.
[0051] In step S130, the deduction parameters are divided into fixed parameters and adjustable parameters. While the fixed parameters remain unchanged, the adjustable parameters that need to be adjusted are determined according to the structured deduction instructions, and the adjustable parameters that need to be adjusted are incrementally updated.
[0052] Fixed parameters represent simulation parameters that do not need to change with the changing needs of flood simulations during the current continuous simulation process. Adjustable parameters represent simulation parameters that can be adjusted according to different flood simulation needs. Incremental updates represent a parameter update method that updates only the parameters whose values have changed relative to their existing values, while keeping the unchanged parameters at their original values.
[0053] Specifically, after determining the simulation parameters involved in the current 3D flood simulation, the simulation parameters can be divided into fixed parameters and adjustable parameters based on their change attributes during continuous simulation. Upon receiving a structured simulation command, the parameter requirements corresponding to the structured simulation command are compared with the currently used adjustable parameters to determine the adjustable parameters that need to be changed, and these are designated as the adjustable parameters to be adjusted. During this process, fixed parameters retain their current values, and adjustable parameters not identified as needing adjustment also retain their current values; only the adjustable parameters requiring adjustment are updated according to the structured simulation command. This avoids the need to repeatedly configure all simulation parameters when simulation conditions change locally, reducing the amount of parameter processing during continuous parameter tuning.
[0054] In step S140, the current scene features corresponding to the flood simulation requirements are extracted, and the scene similarity between the current scene features and the existing scene features in the simulation scene library is determined. The target simulation scene is determined based on the scene similarity.
[0055] Among them, current scene features can represent feature information extracted from the current flood simulation requirements to characterize the current simulation conditions. Existing scene features can represent feature information corresponding to simulation scenes pre-stored in the simulation scene library. Scene similarity can represent the quantitative result of the degree of similarity between current scene features and existing scene features. Target simulation scene can represent an existing scene determined from the simulation scene library based on scene similarity and used for the current three-dimensional flood simulation.
[0056] Specifically, scenario information reflecting the differences in current simulation conditions can be extracted based on the flood simulation requirements, and this extracted scenario information is converted into current scenario features. Subsequently, the features of existing scenarios corresponding to each existing scenario in the simulation scenario library are sequentially obtained, and the current scenario features are compared with the features of each existing scenario to obtain the scenario similarity between the current simulation requirements and each existing scenario. Then, based on the scenario similarity, scenarios with a high degree of matching with the current simulation requirements are identified from multiple existing scenarios and designated as target simulation scenarios. This allows existing simulation scenarios to participate in the current simulation process, reducing the need to reconstruct complete scenarios for similar simulation conditions and improving the utilization of existing simulation scenarios.
[0057] In step S150, the adjustable parameters in the target simulation scenario are corrected based on the difference between the incrementally updated adjustable parameters and the corresponding parameters in the target simulation scenario.
[0058] Here, "difference" can represent the difference in parameter values between the incrementally updated adjustable parameters and the corresponding parameters in the target simulation scenario. "Parameter correction" can represent the process of adjusting the corresponding parameters in the target simulation scenario that have parameter differences, based on the incrementally updated adjustable parameters, so that the target simulation scenario meets the current flood simulation requirements.
[0059] Specifically, after obtaining the target simulation scenario, each adjustable parameter after incremental updates can be compared with the corresponding parameters in the target simulation scenario to determine if there are any parameter differences. For parameters without differences, the original parameters in the target simulation scenario can continue to be used; for adjustable parameters with differences, the corresponding parameters in the target simulation scenario can be corrected based on the incrementally updated parameters, transforming the target simulation scenario from an existing scenario similar to the current flood simulation requirements into a scenario that meets the current simulation conditions. Thus, while reusing the existing content of the target simulation scenario, adjustments can be made only to the actual parameter differences, further reducing redundant scenario configuration and parameter processing.
[0060] In step S160, a three-dimensional flood simulation is performed based on the target simulation scenario after parameter correction to obtain the simulation results.
[0061] Among them, three-dimensional flood simulation can represent the process of numerically calculating the occurrence and evolution of floods based on the flood calculation conditions of the target area, and characterizing the corresponding flood state in three-dimensional space. The simulation results can represent the data obtained during or after the three-dimensional flood simulation, used to characterize the flood state and its changes in the target area.
[0062] Specifically, the target simulation scenario, after parameter correction, can be used as the current simulation condition. This allows relevant parameters within the target scenario to participate in flood calculations, and the changes in flood status within the target area are calculated according to the corresponding time process, yielding simulation results corresponding to the current flood simulation requirements. Because this simulation process uses a target simulation scenario that has been matched to the scenario and corrected for differences in parameters, it ensures that the current simulation requirements correspond to the actual simulation conditions used, and reduces processing overhead caused by repeatedly constructing scenarios and repeatedly configuring unchanged parameters, thereby improving the scenario adaptability and processing efficiency of 3D flood simulations.
[0063] The following is a detailed explanation of the steps described above.
[0064] In some embodiments, obtaining flood simulation requirements, performing semantic parsing on the flood simulation requirements, and generating structured simulation instructions can be achieved through the following technical steps: The first step is to obtain the requirements for flood forecasting.
[0065] The flood simulation requirements can represent the information needed to describe the simulation conditions, adjustments to those conditions, or the desired flood status when conducting flood simulations for a target area. Flood simulation requirements can be input directly using structured parameters or in natural language. When using structured parameters, the corresponding parameters and values can be directly provided by external business systems, parameter configuration interfaces, or pre-established task templates. When using natural language, flood control management personnel, dispatchers, or other staff can input the relevant requirements according to the actual simulation purpose.
[0066] For example, flood simulation requirements may include rainfall control requirements, engineering scheduling requirements, risk query requirements, and operational condition simulation requirements. For instance, one could input "conduct a flood simulation of the target area under a 5-year return period rainstorm condition"; or "determine the flow rate, water level, and surface water accumulation of the pipeline nodes in the target area under a 110mm rainfall condition"; or "keep the current basic conditions of the area unchanged, adjust the rainfall from 110mm to 130mm, and then re-simulate"; or "analyze the changes in water accumulation in the target area after adjusting the pump station operation status".
[0067] For continuous simulations, flood simulation requirements can also describe only the changes needed relative to the current simulation state. For example, if the previous simulation was completed for a 10-year return period rainstorm, the current flood simulation requirement could simply be "adjust the rainstorm recurrence period to a 20-year return period"; if the previous simulation already set the target area and engineering operation status, the current simulation requirement could simply be "increase the simulation duration by 2 hours"; and if the existing rainfall conditions and spatial range remain unchanged, the requirement could simply be "adjust the pump station start / stop status". Therefore, flood simulation requirements can be used to fully describe a new simulation task or to indicate adjustments to certain conditions during continuous simulations.
[0068] The second step is to perform semantic analysis on the flood simulation requirements to determine the corresponding rainfall parameters, river hydrological parameters, engineering scheduling parameters, pipeline hydraulic parameters, and waterlogging risk parameters.
[0069] Among these, rainfall parameters can represent parameters used to describe the rainfall process and intensity characteristics in the target simulation conditions, such as total rainfall, rainfall intensity, rainfall duration, rainstorm recurrence interval, or rainfall sequence. River hydrological parameters can represent parameters used to describe the hydrological state or boundary conditions of rivers in the target area. Engineering scheduling parameters can represent parameters used to describe the operational status of flood control and drainage projects such as gates, pumping stations, and flood diversion facilities. Pipeline hydraulic parameters can represent parameters used to describe the hydraulic state of drainage networks or hydraulic calculation conditions. Waterlogging risk parameters can represent parameters used to limit the conditions for judging waterlogging status or analyzing waterlogging risk.
[0070] Specifically, it can identify information describing rainfall conditions, river conditions, engineering operation conditions, pipeline conditions, and waterlogging conditions in flood simulation requests, and map them to corresponding parameter categories according to their semantic meaning. For example, for the request "the cumulative rainfall in the next 3 hours is 35mm," the rainfall time condition and cumulative rainfall can be identified; for the request "re-simulate after all pump stations are turned on," the corresponding engineering scheduling conditions can be identified; for the request "view the waterlogging situation in a certain area," the corresponding spatial objects and waterlogging analysis targets can be identified. When the same flood simulation request contains multiple types of conditions, information of each type can be extracted separately and corresponding parameter items can be formed.
[0071] The third step is to encapsulate the determined parameters in a structured manner according to the preset instruction format to obtain the structured deduction instruction.
[0072] The preset instruction format can represent the data structure used to uniformly organize different types of inference parameters. Structured encapsulation can represent the organization of the parsed parameters according to preset fields, field levels, and data types, so that flood inference requirements of different expressions can be converted into data that can be read by subsequent inference processes.
[0073] For example, a standardized JavaScript Object Notation (JSON) can be used to form a simulation instruction package, with rainfall parameters, river hydrological parameters, engineering scheduling parameters, pipeline hydraulic parameters, and waterlogging risk parameters written into corresponding fields. Depending on the actual simulation needs, the simulation instruction package can also include parameters such as levee conditions, underlying surface parameters, spatial simulation range, simulation duration, and result output accuracy. For unnecessary adjustment parameters not appearing in the flood simulation requirements, the current valid values can be retained or corresponding change fields can be omitted, thus ensuring that the structured simulation instructions accurately represent the actual simulation conditions involved in the current requirements.
[0074] In some embodiments, semantic parsing of flood projection requirements can be performed through the following technical steps: The first step is to obtain the flood-labeled dataset, which includes the correspondence between natural language questions and standard simulation parameters.
[0075] The flood-annotated dataset represents a set of data constructed for flood simulation operations to characterize the mapping relationship between operational language and simulation parameters. Standard simulation parameters represent parameters that have been standardized according to the parameter types, names, and data formats used in flood hydrodynamic simulations.
[0076] The flood-labeled dataset can be constructed based on typical rainstorm cases, flood control dispatch logs, hydrological and hydrodynamic simulation parameter logs, gate and pump dispatch procedures, and disaster assessment reports. It can be further supplemented using flood control consultation records, meteorological rainstorm forecasts, and flood evolution analysis data. Based on the above data, natural language statements that reflect the operational methods of flood control can be selected, and the corresponding standard simulation parameters for each statement can be determined, thereby establishing a correspondence between natural language questions and standard simulation parameters.
[0077] For example, for the statement "The cumulative rainfall will increase by 20% in the next 3 hours," a correspondence can be established between it and the parameters for adjusting the total rainfall and the simulation duration. For the question "Adjust the gate opening to the specified value and then re-simulate," a correspondence can be established between it and the engineering scheduling parameters. For "Check the maximum water accumulation in a low-lying area," a correspondence can be established between it and the spatial simulation range and the waterlogging analysis parameters. By establishing annotation relationships for different business expressions, the same parameter can be mapped to a unified parameter meaning when expressed in different natural language.
[0078] The second step is to adapt the semantic parsing model to the domain based on the flood-labeled dataset.
[0079] Among them, the semantic parsing model can represent a model capable of recognizing semantic information in flood simulation requirements and mapping the recognition results to standard simulation parameters. Domain adaptation can represent using flood domain data and flood business rules to adjust the parameters or impose input constraints on a general language model, making the model adaptable to flood terminology, simulation parameters, and engineering scheduling logic.
[0080] For example, the semantic parsing model can employ a Large Language Model (LLM). Specifically, it can use a model with a sparse Mixture of Experts (MoE) structure, long contextual information processing, and Chinese-specific semantic understanding capabilities, such as the DeepSeek series models. During domain adaptation, lightweight domain fine-tuning can be employed to enable the model to learn the correspondence between hydrological terminology, flood conditions, simulation boundaries, and engineering scheduling expressions and standard simulation parameters. During lightweight domain fine-tuning, the general basic parameters can remain unchanged, while training is performed on parameters specifically designed for flood domain adaptation, thereby enhancing the semantic parsing performance for flood-related business scenarios while retaining general language understanding capabilities.
[0081] The third step is to retrieve flood cases, engineering scheduling procedures, hydrological parameter thresholds, and simulation constraints that match the flood simulation requirements from the flood domain knowledge base, based on the needs of flood simulation.
[0082] The flood control knowledge base can be represented as a dataset used to store searchable professional information during flood simulations. This knowledge base can be built using a vectorized storage method, allowing for the retrieval of relevant content based on semantic features corresponding to the flood simulation requirements.
[0083] For example, Retrieval-Augmented Generation (RAG) can be used for professional data retrieval. First, flood case studies, engineering scheduling procedures, hydrological parameter thresholds, simulation constraints, and other flood-related data are converted into searchable representations and stored in a water resources-specific vector knowledge base. When a flood simulation request is received, relevant data can be retrieved from the knowledge base based on the semantics of the request. For example, when the request involves the operation of gate pumps, the corresponding scheduling procedures and engineering operation boundaries can be retrieved; when the request involves extreme rainfall, the corresponding rainstorm case studies and hydrological parameter ranges can be retrieved; when the request involves assessing waterlogging risk, the corresponding parameter thresholds and operating constraints can be retrieved.
[0084] The fourth step is to adapt the retrieved content to the semantic parsing model that is adapted to the flood simulation requirements, and determine the corresponding standard simulation parameters according to the output format constraints.
[0085] Output format constraints can represent rules used to limit the output fields, field types, parameter names, and parameter expressions of the semantic parsing model. In practice, the flood simulation requirements can be used as the current parsing object, and the professional information retrieved from the knowledge base can be used as the context information corresponding to the current parsing object. At the same time, the output fields and field types are set so that the semantic parsing results are generated in a data format that hydrodynamic simulation can recognize.
[0086] For example, when flood simulation requirements include total rainfall, simulation time, and pumping station operation requirements, the model can be limited to outputting rainfall parameter fields, simulation time fields, and engineering scheduling fields separately, avoiding the mixing of natural language interpretations into the parameter results. For information involving parameter units, the corresponding units can be specified in the output format or a unified conversion rule can be established to reduce data discrepancies between different expressions.
[0087] Furthermore, prompt word engineering can be used to constrain the input and output processes. For example, the prompt message can clearly specify the flood parameter types, optional fields, parameter relationships that must not conflict, and the output format that needs to be maintained. Lightweight fine-tuning, specialized data retrieval, and output format constraints work together to reduce the deviation between general semantic parsing results and hydrodynamic simulation parameter requirements, enabling flood simulation needs to be more stably converted into standard simulation parameters that can be used in subsequent simulations.
[0088] In some embodiments, based on regional rainfall frequency statistics, hydrophysical constraints, and parameter combination rules, the inference parameters are combined and filtered to generate multiple candidate inference scenarios. This can be achieved through the following technical steps: The first step is to determine the range of rainfall parameters corresponding to different rainstorm recurrence periods based on the statistical results of regional rainstorm frequency.
[0089] Among them, the regional rainstorm frequency statistics can represent statistical results obtained based on historical rainfall data of the target area or existing rainstorm frequency analysis results, used to characterize the rainfall characteristics corresponding to different rainstorm return periods. The rainfall parameter range can represent the range of rainfall parameter values allowed to be used to generate candidate simulation conditions under the corresponding rainstorm return period.
[0090] In practice, the statistical data on the frequency of heavy rainfall corresponding to the target area can be determined first, and the magnitude of conventional heavy rainfall under different return periods can be obtained based on the statistical results. For each return period, the range of values for the corresponding rainfall, rainfall intensity, or other rainfall characteristics can be read or calculated. This ensures that subsequent parameter combinations are constrained by the actual heavy rainfall patterns in the target area, rather than arbitrarily generating rainfall conditions under unbounded conditions.
[0091] The second step involves combining rainfall intensity, rainfall duration, rainstorm recurrence period, spatial distribution of rainfall patterns, engineering scheduling status, spatial extrapolation range, and simulation duration within the range of rainfall parameters to obtain multiple candidate parameter combinations.
[0092] Candidate parameter combinations can represent a set of parameters formed by multiple parameters used to describe the same flood simulation condition. Specifically, rainfall intensity, rainfall duration, and rainstorm recurrence interval can be used as parameters related to the rainfall condition; the spatial distribution of rainfall pattern can be used as parameters related to the spatial distribution of rainfall; the engineering scheduling status can be used as the operating conditions of flood control and drainage projects; the spatial simulation range can be used as the current calculation area; and the simulation duration can be used as the current simulation time condition.
[0093] When forming candidate parameter combinations, appropriate values can be selected from the allowable range or optional states of each parameter according to the current flood simulation requirements, and different combinations can be formed. For example, under the same rainstorm recurrence period, different rainfall durations and different engineering scheduling states can be configured; under the same engineering scheduling state, combinations with different rainfall intensities and different spatial simulation ranges can be formed; for continuous rainfall or local rainfall requirements, the spatial distribution of rainfall patterns and simulation duration can be changed so that the candidate parameter combinations can reflect the differences in conditions between different flood conditions.
[0094] The scenario generation process can also use regional rainfall frequency characteristics, reasonable thresholds for hydrological parameters, water conservancy project scheduling boundaries, and spatial simulation range as prior conditions. Water conservancy project scheduling boundaries can be used to limit the permissible operating states of engineering objects such as gates, pumping stations, and flood diversion facilities; spatial simulation range can be used to limit the simulation area, so that the generated parameter combinations match the current calculation area.
[0095] The third step is to verify the combination of multiple candidate parameters based on hydrophysical constraints.
[0096] Among them, hydrophysical constraints can represent constraints formed based on the physical relationships between rainfall, runoff generation, confluence, pipeline hydraulic processes, and related engineering conditions. During verification, parameters in the candidate parameter combinations can be read one by one, and it can be determined whether the current parameter combination meets the hydrological and hydrodynamic calculation requirements of the target area.
[0097] For example, it can be determined whether rainfall parameters are within the range supported by regional rainstorm statistics, whether relevant hydrological conditions correspond to the spatial extrapolation range, and whether candidate parameter combinations can serve as effective inputs for hydrological and hydrodynamic calculations. Candidate parameter combinations that clearly violate hydrophysical relationships, cannot form effective simulation boundaries, or cannot correspond to the engineering conditions of the target area can be excluded from the next processing stage.
[0098] The fourth step is to eliminate candidate parameter combinations that have any of the following problems from the verified candidate parameter combinations, based on the parameter combination rules: exceeding rainfall parameters, conflicting rainfall patterns, conflicting engineering scheduling status, or conflicting spatial boundaries.
[0099] Among them, parameter combination rules can represent the judgment rules established for the combination relationship between multiple inference parameters. Parameter combination rules differ from single parameter range judgments in that they can further determine whether the simultaneous occurrence of two or more parameters constitutes a reasonable operating condition.
[0100] Rainfall parameter exceeding limits indicates that the rainfall parameters in the candidate parameter combination exceed the range allowed for the current region or scenario. Rainfall pattern conflict indicates that the set spatial distribution of rainfall patterns cannot be consistent with other rainfall conditions or spatial conditions. Engineering scheduling status conflict indicates that a scheduling action is inconsistent with the actual allowed operating state of the corresponding engineering object. Spatial boundary conflict indicates that there is a mismatch between the current spatial simulation range and other boundary conditions.
[0101] For example, if a pumping station is in an inoperable state, but the candidate parameter combination requires the pumping station to participate in drainage according to its operating state, it can be determined that there is a conflict in the engineering scheduling state; if the simulation area does not contain a certain engineering object, but the candidate parameter combination still uses the engineering object as the scheduling condition within the current spatial range, it can be determined that there is a conflict in the spatial boundary.
[0102] The fifth step is to generate multiple candidate simulation scenarios based on the remaining candidate parameter combinations.
[0103] Specifically, each candidate parameter combination filtered through hydrophysical constraint verification and parameter combination rules can be used as a parameter description for the corresponding simulation condition, and a corresponding candidate simulation scenario can be generated accordingly. The generated candidate simulation scenario can retain the correspondence between each parameter and the relevant boundary conditions of the current area for subsequent scenario management.
[0104] Therefore, by first determining the rainfall range based on the frequency of regional rainstorms, then combining multiple parameters, and continuously implementing hydrophysical constraint verification and parameter combination rule filtering, the number of candidate simulation conditions can be increased while reducing obviously unreasonable combinations, making the generated results more consistent with the flood evolution conditions of the target area.
[0105] In some embodiments, generating multiple candidate scenarios based on the remaining candidate parameter combinations and constructing a scenario library based on the multiple candidate scenarios can be achieved through the following technical steps: The first step is to divide the candidate simulation scenarios into conventional gradient rainstorm scenarios, short-term extreme heavy rainfall scenarios, multi-condition superposition composite scenarios, and extreme flooding scenarios exceeding the standard, based on the rainfall parameters corresponding to each candidate parameter combination and the engineering operation status.
[0106] Conventional gradient rainstorm scenarios can include low-gradient, medium-gradient, and high-gradient conventional rainfall scenarios, as well as standard rainstorm scenarios corresponding to different rainstorm return periods. These scenarios can be used to characterize the flood evolution of a target area under conventional rainfall conditions and can be used for routine risk assessments or basic drainage capacity verification.
[0107] Short-duration extreme heavy rainfall scenarios can include short-duration heavy rainfall, sudden localized rainstorms, and scenarios where hourly rainfall intensity exceeds the corresponding standards. These scenarios are characterized by relatively short rainfall duration, large instantaneous rainfall intensity, and high spatial concentration, and can be used to characterize flood conditions where heavy rainfall rapidly affects a target area within a short period of time.
[0108] Multi-condition overlay scenarios can represent hypothetical scenarios where two or more flood-related conditions coexist. For example, these could include scenarios overlaying rainfall and pump station scheduling, rainfall and pipeline blockage, and continuous multi-day rainfall. By combining rainfall conditions with engineering operation status or other conditions, composite hypothetical scenarios can be formed for multi-scheme comparison, engineering scheduling analysis, and parameter sensitivity analysis.
[0109] Extreme flood scenarios exceeding standards can include once-in-a-century rainstorms, extreme rainfall pattern deviations, and engineering failures. These scenarios can be used to represent flood conditions that exceed normal design conditions or the normal operating state of engineering projects, thus enabling the simulation scenario library to not only include conventional historical conditions but also cover extreme risk conditions.
[0110] By dividing candidate simulation scenarios into different scenario categories, the scenario library can simultaneously cover different situations such as regular rainfall, short-term heavy rainfall, combined working conditions, and extreme flooding, thereby reducing the problem of insufficient coverage of extreme working conditions when constructing scenarios based solely on historical measured working conditions.
[0111] The second step is to associate each candidate simulation scenario with its corresponding standard parameter configuration, initial hydraulic conditions, boundary constraints, and linkage calculation rules.
[0112] The standard parameter configuration represents a set of standardized parameters and their values used to describe the simulation conditions corresponding to the candidate simulation scenario. The initial hydraulic condition represents the initial hydraulic state used when the candidate simulation scenario begins hydrodynamic calculations. Boundary constraints represent the conditions that limit the spatial, hydrological, or engineering operation boundaries of the current candidate simulation scenario. The linkage calculation rules represent the rules used in the current candidate simulation scenario to determine the processing relationships between calculations related to rainfall, hydraulics, waterlogging, and engineering scheduling.
[0113] For example, for a typical gradient rainstorm scenario, the corresponding return period, rainfall duration, spatial range, and current engineering status can be configured as standard parameters, while also associating the initial hydraulic state of the current area and the calculation boundary. For complex scenarios including pump station scheduling conditions, the pump station operating conditions and corresponding linkage calculation rules can be further associated. Thus, when a candidate scenario is subsequently invoked, the parameters and calculation conditions corresponding to that scenario can be obtained simultaneously, without the need to search for each piece of information separately.
[0114] The third step is to classify, encode, and structurally store the associated candidate scenarios to build a scenario library.
[0115] Classification coding can be based on scenario category, scenario number, or other coding methods that can distinguish different scenarios. Structured storage allows scenario category, standard parameter configuration, initial hydraulic conditions, boundary constraints, and linkage calculation rules to be set as corresponding data fields, and associations are established through scenario identifiers. The simulation scenario library can support reading corresponding candidate simulation scenarios according to scenario category or scenario characteristics. Thus, the originally scattered simulation conditions can be transformed into structured scenario records with associated parameters and calculation conditions, improving the efficiency of organizing and calling multiple simulation conditions.
[0116] In some embodiments, determining the adjustable parameters to be adjusted based on structured deduction instructions can be achieved through the following technical steps: The first step is to obtain the state of the previous simulation, which includes the fixed parameters and adjustable parameters used in the previous simulation.
[0117] The simulation state can represent the parameter state that characterizes the current simulation condition during a flood simulation. The previous simulation state can be the current valid state saved after the previous simulation was completed in a continuous simulation process. When the current task is the first simulation, the initialized simulation state can be used as the initial state; when the current task is the second or subsequent simulations, the state corresponding to the end of the previous simulation can be read.
[0118] To describe the state relationships in the continuous deduction process, the first... The global state of the round of deduction is represented as follows: .in, Indicates the first The global state of the round-robin simulation; Indicates the first A fixed state formed by fixed parameters in a round of simulation; Indicates the first Adjustable states formed by adjustable parameters in round-robin simulation; This indicates the current simulation round.
[0119] The above state representation method can distinguish between the parts that remain unchanged during the deduction process and the parts that can change according to the needs, so that the continuous deduction process can continue the state based on the existing state, instead of treating each round of deduction as a new task that is unrelated to each other.
[0120] The second step is to determine the target state for this round based on the structured simulation instructions.
[0121] The target state for this round can represent the parameter state that needs to be achieved in the current round according to the requirements of the current structured simulation instructions. Specifically, it can read the parameter fields in the structured simulation instructions and write the parameter values that are explicitly required to be changed in the current instructions into the corresponding state positions; for state information that the current instructions do not require to be changed, the corresponding content in the current valid state can continue to be used.
[0122] During continuous simulation, the target state for the current round can be determined based on the state of the previous round and the current structured simulation instructions, and can be expressed as: .in, Indicates the first The target status of this round of simulation; Indicates the first The state of the previous simulation that is retained after the simulation ends; Indicates the first The structured deduction instructions corresponding to the wheel; Represents the state mapping function; This indicates the current simulation round.
[0123] The third step is to compare the target state in this round with the state in the previous round of simulation to determine the corresponding parameter increments.
[0124] Parameter increments can represent the changes in parameters of the target state in the current round relative to historical states. Parameter increments can be represented by differences in the overall state or differences in various parameter dimensions.
[0125] The parameter increment between the target state and the historical state can be expressed as: .in, Indicates the first The parameter increment of the target state in this round relative to the previous round of simulation; Indicates the first Round target status; Indicates the first The simulation status; This indicates the current simulation round.
[0126] The fourth step is to use the results from the previous round of deduction when the parameter increment is zero. When the following conditions are met... If the current simulation conditions are consistent with the previous simulation conditions, the results of the previous simulation can be used without repeating the simulation for the same state.
[0127] Fifth, if the parameter increment is not zero, determine the adjustable parameter that needs to be adjusted based on the parameter increment. When the condition is met... At this time, the positions of non-zero parameters in the parameter increment can be further located, and the corresponding adjustable parameters can be determined as the adjustable parameters that need to be adjusted.
[0128] For example, when the scenario remains unchanged from the previous round and only the rainfall conditions are changed, the parameter increments mainly correspond to rainfall-related parameters; when only the project operation status is adjusted, the parameter increments mainly correspond to the corresponding project scheduling parameters. By comparing the status first and then determining the parameters that have changed, we can avoid mistakenly using parameters not covered by the current structured simulation instructions as adjustment targets, and can directly use existing simulation results when there are no changes.
[0129] In some embodiments, incremental updates to the adjustable parameters that need to be adjusted can be performed through the following technical steps: The first step is to define the pipeline topology, underlying surface properties, terrain data, and water system boundaries as fixed parameters.
[0130] Fixed parameters can represent modeling parameters that retain their current values during continuous simulations within the same batch, provided no adjustment requests are received. Pipeline topology is used to characterize nodes, pipe segments, and their connections within the drainage network; underlying surface attributes characterize the surface properties of the target area that influence rainfall generation and runoff processes; topographic data can be represented using digital elevation models (DEMs) or similar formats to provide topographic elevation information for the target area; and water system boundaries define the computational conditions corresponding to rivers, water bodies, or relevant hydrological boundaries.
[0131] The aforementioned fixed parameters typically constitute the basic modeling conditions for the current hydrodynamic model of the target area. During continuous parameter tuning, if the flood simulation requirements do not necessitate modification of these conditions, there is no need to reset the fixed parameters due to changes in adjustable parameters such as rainfall, engineering scheduling status, or simulation time.
[0132] The second step is to determine the rainfall parameters, engineering scheduling parameters, risk assessment parameters, and simulation time parameters as adjustable parameters.
[0133] To represent the compositional relationship of adjustable states, the first... The adjustable state corresponding to the round deduction can be represented as: .in, Indicates the first Adjustable states consisting of adjustable parameters in round-robin simulation; Indicates the first The rainfall parameters corresponding to each round can include the return period of the rainstorm, the total rainfall, and the rainfall sequence. Indicates the first The risk assessment parameters corresponding to the wheel can include the water accumulation risk threshold and the analysis space range; Indicates the first The engineering scheduling parameters corresponding to the round can include gate opening degree, pump station start / stop status and flood diversion facility status; Indicates the first The simulation time parameters corresponding to the round can include the total simulation duration and the calculation time step. This indicates the current simulation round.
[0134] In practice, rainfall parameters can be adjusted by increasing or decreasing rainfall, adjusting rainfall duration, switching the recurrence period of rainstorms, or adjusting the spatial distribution of rainfall intensity according to different simulation tasks; risk assessment parameters can be adjusted by adjusting the threshold or analysis range according to the current analysis requirements; engineering scheduling parameters can be adjusted according to the target operating status of gates, pumping stations, or flood diversion facilities; and simulation time parameters can be configured according to the total simulation duration or the calculation time step.
[0135] The third step is to ensure that the fixed parameters in the current simulation state remain at their corresponding values from the previous simulation state when the structured simulation instructions do not contain adjustment information for fixed parameters.
[0136] In the When the structured deduction instructions do not involve adjustments to fixed parameters, the following can be satisfied: .in, Indicates the first Fixed parameter states in round-robin simulation; Indicates the first Fixed parameter states in round-robin simulation; This indicates the current simulation round.
[0137] Therefore, when only adjustable parameters are adjusted, the pipeline topology, underlying surface properties, terrain data, and water system boundaries retain the corresponding values from the previous round, ensuring consistent basic modeling conditions between consecutive simulations and avoiding the need to reinitialize the unchanging basic parameters each time adjustable parameters are adjusted.
[0138] The fourth step is to update the adjustable parameters that need to be adjusted based on the parameter increment, and to keep the adjustable parameters that have not changed at their corresponding values from the previous round of simulation.
[0139] To describe the composition of engine parameters, the complete set of engine parameters can be represented as: .in, This represents the full set of parameters of the engines participating in the current simulation. Represents a fixed set of parameters; This represents the set of adjustable parameters.
[0140] No. After the wheel parameters are adjusted, the adjustable parameters can be updated according to the following relationship: .in, Indicates the first The set of adjustable parameters after each round of updates; Indicates the first The set of adjustable parameters used by the wheel; Indicates the first The parameter increments determined by the structured deduction instructions; This indicates a parameterized update operator; This indicates the current simulation round.
[0141] The parameterized update operator can read the parameter position corresponding to the parameter increment and update that position using the current target value or the corresponding increment. For other adjustable parameters whose parameter increments do not indicate a change, the corresponding values from the previous round are used. For example, if the current requirement is only to change the total rainfall from 110mm to 130mm, only the parameter corresponding to the total rainfall can be updated, while the engineering scheduling parameters, risk assessment parameters, and simulation time parameters continue to maintain their current valid values. If the current requirement is only to modify the simulation duration, the rainfall and engineering scheduling conditions can be retained, and only the simulation time parameter can be adjusted.
[0142] The parametric simulation process can also be organized in a modular fashion, configuring corresponding parameter interfaces for different processing functions such as rainfall-driven simulation, hydraulic calculation, surface waterlogging, engineering scheduling, and risk analysis. This allows changed parameters to be passed to the corresponding processing function without reconfiguring other unchanged parameters. This reduces the need for redundant configuration of all parameters due to changes in local conditions during multiple simulation runs.
[0143] In some embodiments, after updating the adjustable parameters to be adjusted according to the parameter increment, the following technical steps can also be used: The first step is to determine whether the updated adjustable parameters are within the corresponding allowable range by checking if they exceed the limits.
[0144] The allowable range can represent the range of parameter values determined by regional rainfall statistics, hydrological calculation conditions, model applicability, or relevant operational constraints. Parameter exceeding limits can be determined by comparing each updated adjustable parameter with its corresponding allowable range.
[0145] For example, it can be determined whether the adjusted rainfall exceeds the allowable rainfall range under the current operating conditions, whether the simulation time parameters meet the model time configuration requirements, and whether relevant threshold parameters are within the allowable range. For parameters that exceed the corresponding range, it can be determined that the current parameter update has failed the parameter over-limit judgment, thereby preventing obviously abnormal parameters from directly entering subsequent calculations.
[0146] The second step is to perform scheduling logic verification based on the project's operating status and scheduling parameters to determine whether the project scheduling parameters match the project's operating status.
[0147] The engineering operation status can represent the current operational status of a gate, pumping station, flood diversion facility, or other engineering object in the target scenario. The scheduling logic verification can determine whether the scheduling actions required by the structured simulation instructions are consistent with the current operating conditions of the corresponding engineering object.
[0148] For example, if a pumping station is set to be out of service or unavailable, and the project scheduling parameters require that pumping station to participate in drainage, it can be determined that the project scheduling parameters and the project's operating status do not match. Similarly, if the current project object does not have the corresponding scheduling status, it can also be determined that there is a conflict in the current scheduling conditions. By jointly judging the project status and scheduling parameters, the problem of failing to detect operational logic contradictions by only checking the parameter value range can be reduced.
[0149] The third step is to perform parameter coupling verification on the updated adjustable parameters and fixed parameters based on hydrophysical constraints.
[0150] Parameter coupling verification can be used to determine whether the current inference conditions, which consist of both adjustable and fixed parameters, can meet the requirements of hydrological and hydrodynamic calculations. For example, it can determine whether the updated rainfall conditions can form an effective calculation condition together with the current spatial range, underlying surface, and river system boundaries. It can also determine whether there are conflicts between the time conditions, risk assessment conditions, and the current model's basic conditions that prevent them from being linked for calculation.
[0151] The fourth step involves using the updated adjustable parameters for scenario matching and parameter correction, provided that the parameter over-limit detection, scheduling logic verification, and parameter coupling verification all pass.
[0152] If any check fails, the current update result can be omitted for subsequent processing, and the problematic parameter can be determined based on the type of check that failed. If all three types of checks pass, the updated adjustable parameters can be retained for subsequent scene matching and parameter correction in target scenario inference.
[0153] By performing layered verification based on parameter values, engineering operation logic, and parameter combination relationships, the possibility of unreasonable operating conditions arising after parameter updates can be reduced, thereby reducing calculation failures or abnormal results caused by invalid parameters entering subsequent simulation processes.
[0154] In some embodiments, determining the scene similarity between the current scene features and each existing scene feature in the scenario library can be performed through the following technical steps: The first step is to extract rainfall characteristics, spatial simulation range, engineering scheduling conditions, simulation time scale, and risk judgment threshold from the flood simulation requirements to generate a feature vector for the current scenario.
[0155] The current scenario feature vector can be represented as a vector composed of multiple features used to distinguish different simulation conditions in the current flood simulation requirements, arranged in a unified order. Rainfall features can reflect the current rainfall conditions, such as being characterized by information like rainfall intensity, rainfall duration, and rainstorm recurrence interval; the spatial simulation range can reflect the current analysis area; the engineering scheduling conditions can reflect the current operational combination of gates, pumping stations, or other engineering objects; the simulation time scale can reflect the total simulation duration or corresponding time configuration; and the risk judgment threshold can reflect the threshold conditions used for current risk judgment.
[0156] The current scene feature vector can be represented as: .in, Represents the feature vector of the current scene; This represents the rainfall characteristic value corresponding to the current scene; This represents the feature value indicating the spatial extrapolation range corresponding to the current scene; This represents the feature value of the engineering scheduling condition corresponding to the current scenario; This represents the simulation timescale feature value corresponding to the current scene; This represents the risk assessment threshold feature value corresponding to the current scenario.
[0157] When different features use different data types, they can be converted into a unified numerical representation before distance calculation. For example, continuous numerical parameters can be represented by normalized parameter values, discrete engineering scheduling conditions can be represented by pre-set state codes, and spatial extrapolation ranges can be represented by features that reflect spatial scale or range category. This allows different types of features to enter a unified distance calculation process.
[0158] The second step is to obtain the feature vectors of each existing scenario.
[0159] For the first in the simulation scenario library For each existing scenario, its scenario feature vector can be represented as: .in, Indicates the first The feature vector of each existing scenario. Indicates the first Rainfall characteristic values for existing scenarios; Indicates the first Spatial extrapolation range characteristic values of existing scenarios; Indicates the first Characteristic values of engineering scheduling conditions in existing scenarios; Indicates the first Simulation timescale characteristic values of an existing scenario; Indicates the first Risk identification threshold feature value for each existing scenario; This indicates the index of the existing scenario in the scenario library.
[0160] The current scene feature vector and the existing scene feature vector use the same feature order and representation method to ensure that there is a correspondence between the same dimensions.
[0161] The third step is to calculate the weighted feature distance between the current scene feature vector and the feature vectors of each existing scene, based on the feature weights corresponding to each feature dimension.
[0162] For example, the current scene and the first scene can be calculated using the following weighted Euclidean distance. Differences between existing scenarios: ; in, Represents the feature vector of the current scene With the Feature vector of existing scene Weighted feature distance between them; Represents the feature vector of the current scene; Indicates the first Feature vectors of existing scenarios; Represents the th element in the current scene feature vector. One eigenvalue; Indicates the first In the first existing scenario One eigenvalue; Indicates the first The feature weights corresponding to each feature; This represents the scene feature dimension index, with a value ranging from 1 to 5. This indicates an index for existing scenarios.
[0163] Feature weights can be set according to the sensitivity of different features to flood projection results. For example, when rainfall conditions have a significant impact on the projection results in a certain type of target area, the rainfall feature can be given a higher weight; when changes in engineering scheduling conditions have a significant impact on waterlogging results, the weight of the engineering scheduling condition dimension can also be increased accordingly. By using feature weights, it is possible to avoid treating features from different scenarios equally when their actual impact is significantly different.
[0164] The fourth step is to determine the scene similarity based on the weighted feature distance and the maximum feature distance.
[0165] For example, the current scene can be compared with the first... The scene similarity between existing scenes is represented as follows: ; in, Represents the feature vector of the current scene With the Feature vector of existing scene Scene similarity between them; Indicates the current scene and the first Weighted feature distance between existing scenarios; Indicates the maximum feature distance; Represents the feature vector of the current scene; Indicates the first Feature vectors of existing scenarios; This indicates an index for existing scenarios.
[0166] The maximum feature distance can be determined based on the maximum weighted feature distance that can be formed between scene features in the current scene feature space. When setting the maximum feature distance, the weighted feature distance for each existing scenario should not exceed [a certain value]. Scene similarity is between 0 and 1; the smaller the weighted feature distance, the greater the corresponding scene similarity.
[0167] In some embodiments, determining the target scenario based on scenario similarity can be achieved through the following technical steps: The first step is to sort the multiple existing scenarios according to their similarity.
[0168] Specifically, the system can iterate through all existing scenarios in the scenario library, obtain the scenario similarity between the current scenario and each existing scenario, and then sort the existing scenarios in descending order of scenario similarity. The sorting results can be used to characterize the degree of similarity between each existing scenario and the current flood simulation requirements.
[0169] The second step is to identify the existing scenes with the highest scene similarity as the target simulation scene.
[0170] The target scenario can be defined as an existing scenario that is prioritized for reuse during the current simulation. Higher scenario similarity indicates that the corresponding existing scenario is closer to the current flood simulation requirements in terms of rainfall characteristics, spatial simulation range, engineering scheduling conditions, simulation time scale, and risk assessment threshold. Therefore, the existing scenario with the highest scenario similarity can be identified as the target scenario, and this target scenario can be used as the object for subsequent parameter adjustments.
[0171] Therefore, existing scenarios that closely match the current flood simulation requirements can be selected from the simulation scenario library first, without having to rebuild the complete simulation conditions from the initial state to meet the current flood simulation requirements, thereby improving the utilization of existing simulation scenarios.
[0172] Furthermore, in some embodiments, the adjustable parameters in the target simulation scenario are corrected based on the difference between the incrementally updated adjustable parameters and the corresponding parameters in the target simulation scenario. This can be achieved through the following technical steps: The first step is to compare the adjustable parameters in the target scenario with the adjustable parameters after incremental updates to identify the adjustable parameters that differ from each other.
[0173] Specifically, each adjustable parameter in the target scenario can be compared with the incrementally updated adjustable parameters obtained in step S130, according to the parameter name or corresponding parameter identifier. For parameters with the same value, it can be determined that there is no difference between the corresponding parameters; for parameters with different values, the corresponding parameters can be identified as adjustable parameters with differences.
[0174] For example, if the spatial range, engineering scheduling status, and simulation duration of the target simulation scenario are consistent with the current flood simulation requirements, and only the total rainfall is different, then only the parameter corresponding to the total rainfall can be determined as the adjustable parameter with the difference; if the rainfall conditions are the same but the pump station start-up and shutdown status are different, then the corresponding engineering scheduling parameters can be determined as the adjustable parameters with the difference.
[0175] The second step is to keep the inference parameters that are not different in the target scenario unchanged, and to correct the adjustable parameters that are different based on the incrementally updated adjustable parameters, so as to obtain the target scenario with corrected parameters.
[0176] Specifically, for inference parameters that do not differ, the existing parameters and related calculation conditions in the target inference scenario can continue to be used. For example, the pipeline topology, terrain conditions, basic hydraulic conditions, and other parameters consistent with the current flood inference requirements in the target inference scenario can remain unchanged. For adjustable parameters that differ, the corresponding parameters in the target inference scenario can be replaced or adjusted according to the incrementally updated adjustable parameters.
[0177] For example, when the current flood simulation requirements differ from the target simulation scenario only in terms of total rainfall, the pipeline topology, terrain conditions, underlying surface properties, water system boundaries, engineering operation conditions, and other identical parameters in the target simulation scenario can be kept unchanged, and only the adjustable parameters corresponding to the total rainfall can be modified; when the current flood simulation requirements only change the start-up and shutdown status of pumping stations, the rainfall conditions and other basic conditions in the target simulation scenario can be kept unchanged, and only the corresponding engineering scheduling parameters can be modified.
[0178] In one example, the current requirement differs from the target scenario only in the total rainfall. The target scenario's network topology, terrain, underlying surface, water system boundaries, engineering operating conditions, and other identical parameters can be retained; only the rainfall amount needs to be adjusted to match the current requirement. In another example, the current requirement only necessitates changing the pump station's start / stop status. The rainfall and basic modeling conditions from the target scenario can be reused; only the corresponding engineering scheduling parameters need to be modified.
[0179] In some embodiments, before performing a three-dimensional flood simulation based on the target simulation scenario with parameter correction, the fixed parameters that remain unchanged can be combined with the adjusted parameters that have been corrected to obtain the simulation parameters corresponding to the target simulation scenario, and the three-dimensional flood simulation can be performed on the target simulation scenario with parameter correction based on the simulation parameters.
[0180] Specifically, after correcting the adjustable parameters that differ in the target simulation scenario based on the incrementally updated adjustable parameters, the fixed parameters that remain unchanged during the current simulation process can be read, and the fixed parameters and the corrected adjustable parameters can be combined according to the corresponding parameter categories and parameter identifiers to form the complete simulation parameters currently used in the target simulation scenario.
[0181] The fixed parameters can continue to use the pipeline topology, underlying surface attributes, terrain data, and water system boundaries that remain unchanged in the current simulation. The modified adjustable parameters can be rainfall parameters, engineering scheduling parameters, risk assessment parameters, and simulation time parameters that have been incrementally updated and corrected according to the parameter differences in the target simulation scenario. Through parameter categories or parameter identifiers, each fixed parameter and modified adjustable parameter can be mapped to the corresponding parameter position in the target simulation scenario.
[0182] For example, if the current flood simulation requirement only changes the total rainfall relative to the target simulation scenario, the fixed parameters corresponding to the pipeline topology, underlying surface properties, terrain data, and water system boundaries can be kept unchanged. The corrected total rainfall and other adjustable parameters that retain their current valid values can be combined with the fixed parameters to form the simulation parameters corresponding to the current target simulation scenario. As another example, if the current flood simulation requirement only changes the pump station start / stop status, the fixed parameters and other unchanged adjustable parameters can be kept unchanged, and the corrected engineering scheduling parameters can be combined with the remaining parameters.
[0183] In this way, fixed parameters can continue to participate in 3D flood simulation after the adjustable parameters are incrementally updated and the target simulation scenario parameters are corrected. This allows the basic modeling conditions that have not changed during the current simulation to be reused, and the target simulation scenario after parameter correction has complete simulation parameters corresponding to the current flood simulation requirements, thereby reducing the redundant configuration of unchanged parameters.
[0184] In some embodiments, reference Figure 2 As shown, the three-dimensional flood simulation based on the target scenario after parameter correction can be performed to obtain the simulation results through the following technical steps: Step S210: Generate the corresponding rainfall process based on the rainfall parameters in the target scenario after parameter correction.
[0185] The rainfall process can be represented as the rainfall input that varies with time and space within the current simulation time range. Specifically, the rainfall-driven control module can read parameters such as total rainfall, rainfall intensity, rainfall duration, rainstorm recurrence period, and spatial distribution of rainfall patterns in the target simulation scenario to form the rainfall boundary used in the three-dimensional flood simulation.
[0186] When rainfall parameters are incrementally adjusted, the current rainfall process can be updated according to the corrected rainfall parameters without reconfiguring other unchanged conditions in the target simulation scenario. For example, when only the total rainfall is adjusted, the adjusted rainfall input can be generated based on the current rainfall pattern and rainfall duration; when the return period of the rainstorm changes, the rainfall condition parameters for the corresponding return period can be used to form a new rainfall process.
[0187] Step S220: Apply the corresponding engineering scheduling conditions according to the engineering scheduling parameters.
[0188] Among them, engineering scheduling conditions can represent the operating conditions applied to engineering objects such as gates, pumping stations, and flood diversion facilities during the three-dimensional flood simulation. Specifically, the engineering scheduling and control module can read the corrected engineering scheduling parameters to determine the operating status of each engineering object, and use the corresponding operating status as the engineering operating conditions for subsequent pipeline hydraulic calculations.
[0189] For example, engineering scheduling parameters can represent gate opening, pump station start / stop status, or flood diversion facility operation status. When engineering scheduling parameters in the target simulation scenario are modified, the modified operation status can be used in the corresponding calculation time period; for engineering objects that have not changed, the existing operation status in the target simulation scenario can be maintained. Thus, engineering scheduling conditions can participate in subsequent pipeline hydraulic calculations and reflect the impact of different engineering operation statuses on the drainage process.
[0190] Step S230: Perform hydraulic calculations on the pipeline network based on the rainfall process and engineering scheduling conditions to obtain the flow rate and water level at the pipeline nodes.
[0191] Among them, pipeline hydraulic calculation can be described as the process of numerically calculating the flow state in a drainage pipeline network based on rainfall input, engineering scheduling conditions, pipeline structure, and corresponding hydraulic boundaries. Specifically, the pipeline hydraulic calculation module can read the rainfall process, engineering scheduling conditions, and current pipeline-related calculation conditions to calculate the hydraulic state of each pipeline node and pipe segment, thereby obtaining the pipeline node flow rate and node water level.
[0192] For example, when a pumping station is started, the gate opening is adjusted, or the operating status of flood diversion facilities changes, the corresponding engineering scheduling conditions can be incorporated into the hydraulic calculation process of the pipeline network. This allows different engineering operating states to be reflected in the calculation results of node flow and node water level. Since the basic conditions such as pipeline topology can maintain their current effective state during continuous simulation, when the current demand only changes rainfall conditions or engineering scheduling conditions, the corresponding hydraulic calculations can be performed directly on the existing pipeline network structure, thereby reducing the duplication of pipeline network basic conditions.
[0193] Step S240: Calculate the evolution of surface waterlogging based on the hydraulic calculation results of the pipeline network to obtain the surface water depth and inundation range.
[0194] Among them, the surface waterlogging evolution calculation can represent the calculation of the formation and change of surface water in a target area based on the drainage status of the pipe network, topography, and underlying surface conditions. Specifically, the surface waterlogging evolution module can receive the flow rate and water level of the pipe network nodes obtained from the hydraulic calculation of the pipe network, and combine them with the topography and underlying surface conditions in the target simulation scenario to calculate the surface water depth and corresponding inundation range at different locations.
[0195] In localized low-lying areas, surface water can accumulate and expand during the simulation process if the drainage capacity of the pipe network is insufficient or if the water level at key points rises. As rainfall decreases and drainage continues, the area of water accumulation can gradually shrink. Since the results of the pipe network hydraulic calculations already reflect the corresponding engineering scheduling conditions, changes in the engineering scheduling status can be further reflected in the changes in surface water depth and inundation range through the pipe network hydraulic status.
[0196] Step S250: Risk assessment is performed based on pipeline node flow rate, node water level, surface water depth, and flooding range to obtain the simulation results.
[0197] Specifically, the risk discrimination analysis module can receive the flow rate and water level of pipeline nodes obtained from pipeline hydraulic calculations, as well as the surface water depth and inundation range obtained from surface waterlogging evolution calculations. Based on the current risk discrimination conditions, it determines the corresponding flood risk status and generates a projection result. Thus, rainfall processes and engineering scheduling conditions jointly participate in pipeline hydraulic calculations. The results of these calculations are further used for surface waterlogging evolution calculations, and risk discrimination is performed based on the resulting pipeline hydraulic and surface waterlogging states. This allows rainfall-driven processes, engineering scheduling, pipeline hydraulic calculations, surface waterlogging evolution, and risk discrimination to form a continuous three-dimensional flood projection process within the same target projection scenario.
[0198] In some implementations, the parametric simulation engine can be divided into a rainfall-driven control module, a pipeline hydraulic calculation module, a surface waterlogging evolution module, an engineering scheduling and control module, and a risk discrimination analysis module. Each module can read parameters related to its own calculations, but the calculation results can be passed to subsequent processing modules according to the flood evolution relationship. The parametric simulation engine can also support operations such as increasing or decreasing rainfall, adjusting rainfall duration, switching the return period of rainstorms, adjusting the spatial distribution of rainfall intensity, adjusting water volume thresholds, and configuring simulation duration.
[0199] In some embodiments, risk assessment is performed based on pipeline node flow rate, node water level, surface water depth, and inundation range to obtain projection results. This can be achieved through the following technical steps: The first step is to correlate the flow rate, water level, surface water depth, and flooding range of the pipeline nodes according to the same time point to obtain time-series flood status data corresponding to multiple time points.
[0200] In this context, time-series flood status data can represent data formed by associating multiple types of flood calculation results corresponding to the same time point according to the simulation time sequence. The same time point can be determined by the calculation time step used in the three-dimensional flood simulation. For example, when the simulation process generates calculation results at fixed time intervals, the corresponding timestamp can be used as the association identifier for each type of result.
[0201] In practice, the flow rate sequence, water level sequence, surface water depth sequence, and inundation range sequence of pipeline nodes can be obtained separately. Data corresponding to the same time node are then linked into a set of time-series flood status data according to timestamps or time step numbers. In this way, each set of time-series flood status data can simultaneously reflect the hydraulic state of the pipeline network and the surface waterlogging state at the current moment. During the three-dimensional hydrodynamic calculation, the corresponding rainfall process can also be saved simultaneously, allowing rainfall, pipeline node flow rate, node water level, and surface waterlogging state to be analyzed along a unified time axis, thereby avoiding misalignment of peak times for different indicators.
[0202] The second step is to assess the risk of flooding at each time point based on the preset risk assessment threshold, thereby obtaining the flood risk status corresponding to each time point.
[0203] The preset risk threshold represents the pre-defined criteria used to distinguish different flood states. The flood risk state can be represented by the state result obtained by comparing the hydraulic and waterlogging data at the current time point with the corresponding risk conditions.
[0204] Specifically, for each time point, corresponding time-series flood status data can be read and compared with the current preset risk assessment threshold. If the corresponding indicator at the current time point does not meet the risk conditions, the corresponding normal or low-risk state can be obtained; if the corresponding indicator meets or exceeds the risk conditions, the corresponding risk state can be determined. For the entire simulation process, a flood risk state change sequence can be formed according to the time sequence, thereby reflecting the change of risk state as the flood evolves.
[0205] The third step is to determine the peak flood time based on multiple time-series flood status data, and obtain the peak flow, highest water level, maximum water depth and inundation range of the node corresponding to the peak flood time.
[0206] The peak flood time represents the point in time when the hydraulic state of flooding reaches its peak during the current simulation. In practice, the peak time can be determined based on the flow time series of nodes that characterize the flood peak process, or a unified peak time node can be determined based on the flood peak determination method corresponding to the current simulation object. After determining the peak flood time, instead of independently selecting the maximum values of each indicator from different time points, the corresponding network node flow, node water level, surface water depth, and inundation range are read at the same peak time node. This yields linked results such as the node peak flow, node highest water level, maximum regional water depth, and inundation range corresponding to the flood peak time. For implementations requiring analysis of the overall average water depth, the overall average water depth can be calculated based on the regional water accumulation data corresponding to the same flood peak time. By acquiring multiple indicators at a unified time node, the time inconsistency problem caused by extracting independent maximum values of each indicator can be reduced, enabling corresponding analysis of node hydraulic state and surface waterlogging state during the same flood peak phase.
[0207] The fourth step is to obtain the projection results based on the flood risk status corresponding to each time node and the peak flow, highest water level, maximum water depth and inundation range of the node corresponding to the peak flood time.
[0208] The simulation results can simultaneously include the overall risk status and key flood indicators corresponding to the peak stage. The overall risk status reflects the change of flood risk over time, while the peak stage indicators characterize the more significant hydraulic and waterlogging conditions during this simulation. The two together constitute the simulation results, avoiding the inability to reflect process changes by outputting only a single static peak value, and also avoiding the lack of key moment indicators when only a large amount of time-series data is output.
[0209] The results of the three-dimensional flood simulation can also be dynamically displayed through a three-dimensional geographic information system (GIS), which maps the inundation range and risk status at different time points to the three-dimensional spatial scene of the target area, so that the process of water accumulation range and flood status changing over time can be continuously presented.
[0210] Figure 3 This illustration schematically depicts the flood evolution process under a 5-year return period rainfall scenario according to some embodiments of the present disclosure. (Reference) Figure 3 As shown, under a 5-year return period rainfall scenario, continuous flood simulation can be performed on the target area. The simulation scheme is "5a[124.9163]", and the complete simulation process includes 15 frames with a frame interval of 1800s, a start and end time of 0h to 7h, and a playback magnification of 1x. For example, the simulation time corresponding to the 5th / 15th frame is 2h, at which time the maximum water depth is 4.72m and the flooded area is 35.744km².
[0211] Flood simulations can select either a risk surface or a water depth surface as the layer type. When using a risk surface, different areas can be displayed as low-risk, medium-risk, relatively high-risk, and high-risk areas. Low-risk corresponds to a water depth of less than 0.27m, medium-risk to 0.27–0.40m, relatively high to 0.40–0.60m, and high-risk to greater than 0.60m. The intelligent question-and-answer assistant can generate corresponding simulation results based on flood simulation requirements such as "5-year return period rainfall and inundation situation." The evolution analysis area can simultaneously display the scheme name, total number of frames, frame interval, playback magnification, as well as water depth evolution curves and inundation area change curves, thereby characterizing the changes in water depth, inundation area, and risk status at different time points.
[0212] Figure 4 The diagram schematically illustrates the peak value projection results for a 110mm rainfall condition according to some embodiments of this disclosure. (Reference) Figure 4 As shown, under a rainfall condition of 110mm, a flood simulation result named "Result_110" can be generated. The complete simulation process consists of 29 frames with a frame interval of 900s and a playback magnification of 1x. For example, the simulation time corresponding to the 15th / 29th frame is 3 hours and 30 minutes, at which time the maximum water depth is 4.58m and the inundated area is 15.505km².
[0213] Furthermore, peak flow analysis can be performed on the pipeline network nodes. Specifically, the critical node N0301WS0099 has a nodeIdx of 3825.00, a peakFlow of 4.60, and a peakTimeSec of 4200.00, meaning the peak flow rate is approximately 4.60 m³ / s, occurring about 4200 seconds after the start of rainfall. The evolution analysis area can simultaneously display water depth evolution curves and inundation area change curves, and, combined with the risk surface or water depth surface in three-dimensional space, analyze the hydraulic state of critical nodes and the waterlogging changes in the target area.
[0214] Figure 5 The illustration schematically depicts the water receding process under 130mm rainfall conditions according to some embodiments of the present disclosure. (Reference) Figure 5 As shown, the water receding phase under a 130mm rainfall condition can be analyzed. For example, the water receding phase includes 7 time points, covering a receding process of approximately 1.5 hours; the maximum water depth is approximately 4.959m, and the water area gradually decreases from approximately 11.089km² to approximately 9.316km², with a cumulative reduction of approximately 1.77km².
[0215] In the flood simulation area, the complete simulation result consists of 29 frames with a frame interval of 900 seconds and a playback magnification of 1x. For example, the simulation time corresponding to frame 29 / 29 is 7 hours, at which point the maximum water depth is 4.59 meters and the inundated area is 8.843 km². The evolution analysis area can simultaneously display the water depth evolution curve and the inundated area change curve. The water depth increases initially and then gradually stabilizes, while the inundated area reaches a large value and then gradually decreases over time, thus characterizing the water depth changes and the shrinking of the inundated area during the flood receding phase.
[0216] In some embodiments, risk assessment is performed on flood status data for each time series based on a preset risk assessment threshold to obtain the flood risk status corresponding to each time node. This can be achieved through the following technical steps: The first step is to obtain the flow threshold, water level threshold, and risk classification threshold corresponding to the water accumulation depth for each pipeline node.
[0217] Among them, the flow threshold can represent a critical value used to determine whether the flow rate at a pipeline node or corresponding pipe section exceeds the limit; for example, it can be set based on the pipeline network's design drainage capacity or relevant calculation standards. The water level threshold can represent a critical value used to determine whether the water level at a pipeline node has reached an unfavorable operating state; for example, it can be set based on the safe water level conditions within the pipe. The risk classification threshold can represent a set of thresholds used to determine different water accumulation risk levels based on the depth of surface water accumulation.
[0218] In practice, corresponding thresholds can be set for different pipeline nodes, different pipe sections, or different spatial areas. For the risk classification threshold corresponding to the water depth, multiple threshold ranges can be set according to the flood risk classification requirements adopted for the target area to distinguish different levels of water accumulation risk.
[0219] The second step involves comparing the flow rate at each time node with the corresponding flow rate threshold, the water level at each node with the corresponding water level threshold, and the surface water depth with the corresponding risk classification threshold to obtain the flow rate exceeding limit status, water level exceeding limit status, and water accumulation risk level for each time node.
[0220] Among them, the flow rate exceeding the limit status can indicate whether the flow rate of the pipeline node at the current time point has reached or exceeded the corresponding flow rate threshold; the water level exceeding the limit status can indicate whether the water level of the node at the current time point has reached or exceeded the corresponding water level threshold; the water accumulation risk level can be determined based on the risk classification threshold range in which the surface water accumulation depth at the current time point falls.
[0221] For example, when the flow rate at a certain pipeline node reaches or exceeds the corresponding flow rate threshold, the flow rate exceeding the limit at that time point can be determined; when the water level at a node reaches or exceeds the corresponding water level threshold, the water level exceeding the limit at that time point can be determined; when the surface water depth in a certain area falls into the corresponding risk classification threshold range, the water accumulation risk level of that area can be determined.
[0222] The third step is to determine the flood risk status corresponding to each time point based on the flow rate exceeding the limit, the water level exceeding the limit, and the water accumulation risk level.
[0223] Specifically, for the same time point, the states of exceeding flow limits, exceeding water levels, and waterlogging risk levels can be correlated, so that the flood risk status at the current time point simultaneously reflects the pipeline flow, node water level, and surface waterlogging. For example, if there are flow or water level exceeding limits at a certain time point, and the corresponding area has a high waterlogging risk level, the relevant information can be combined as components of the flood risk status at that time point. By sequentially assessing the risk at multiple time points, a sequence of flood risk statuses that changes over time can be formed.
[0224] Furthermore, in some embodiments, the projection results are obtained based on the flood risk status corresponding to each time point and the peak flow, highest water level, maximum water depth, and inundation range of the node corresponding to the flood peak time. This can be achieved through the following technical steps: The first step is to determine the duration of the flow exceeding the limit based on the time points when the flow exceeds the limit continuously, and to determine the duration of the water level exceeding the limit based on the time points when the water level exceeds the limit continuously.
[0225] Among them, the duration of flow exceeding the limit can represent the length of time that a pipeline node is continuously in a state of flow exceeding the limit, and the duration of water level exceeding the limit can represent the length of time that a pipeline node is continuously in a state of water level exceeding the limit. When the simulation results are generated according to a fixed time step, the corresponding duration can be determined based on the number of time nodes that are continuously in the corresponding exceeding state and the time interval between adjacent time nodes; when each time node has a timestamp, the corresponding duration can also be determined based on the start and end times of the continuous exceeding interval.
[0226] For example, if a pipeline node is in a state of excessive flow for multiple consecutive time points starting from the first time point, until the flow returns to within the threshold at the next time point, the duration of the excessive flow can be determined based on the time when it first entered the excessive flow state and the time when it returned to within the threshold. The duration of water level exceeding the limit can be determined in the same way based on the time points when it is continuously in the excessive water level state.
[0227] The second step is to determine the extent of flow exceeding the limit based on the peak flow of the node and the corresponding flow threshold, and to determine the extent of water level exceeding the limit based on the highest water level of the node and the corresponding water level threshold.
[0228] Among them, the flow rate exceeding the limit can represent the degree to which the peak flow of a node exceeds the corresponding flow rate threshold, for example, it can be determined based on the difference between the peak flow of a node and the corresponding flow rate threshold; the water level exceeding the limit can represent the degree to which the highest water level of a node exceeds the corresponding water level threshold, for example, it can be determined based on the difference between the highest water level of a node and the corresponding water level threshold.
[0229] The third step is to generate projection results based on the flood risk status, duration of flow exceeding limits, duration of water level exceeding limits, magnitude of flow exceeding limits, magnitude of water level exceeding limits, maximum water depth in the region, and inundation range corresponding to each time node.
[0230] Specifically, the flood risk status corresponding to each time node can be organized according to the time sequence of the simulation, and the duration of flow exceeding the limit, the duration of water level exceeding the limit, the magnitude of flow exceeding the limit, and the magnitude of water level exceeding the limit can be associated with the corresponding pipe network nodes. At the same time, the maximum water accumulation depth and the inundation range of the region can be associated with the corresponding spatial region, thereby forming a simulation result that can reflect the hydraulic status of the pipe network and the surface waterlogging status.
[0231] For example, the simulation results can include the flood risk status at different time points, the pipeline nodes where flow or water level exceeds limits, the duration and magnitude of the exceedances, and the maximum water depth and inundation range in the region corresponding to the peak flood stage. By correlating the temporal risk status, the exceedance characteristics of pipeline nodes, and the surface waterlogging characteristics, the current flood simulation process can be characterized from multiple aspects, including temporal changes, pipeline hydraulic status, and spatial waterlogging status.
[0232] In some embodiments, after obtaining the simulation results, the scene features, simulation parameters, and simulation results corresponding to this three-dimensional flood simulation are associated to form a simulation scene record, and the simulation scene record is stored in the simulation scene library. This can be done through the following technical steps: The first step is to associate the scene features, simulation parameters, and simulation results corresponding to this three-dimensional flood simulation to form a simulation scene record.
[0233] A scenario simulation record represents a data record used to fully describe a completed 3D flood simulation. Scenario features describe the characteristics of the scenario corresponding to this simulation; simulation parameters are used to reconstruct or identify the specific operating conditions used in this simulation; and simulation results are used to save the hydraulic, waterlogging, and risk calculation results generated under the corresponding operating conditions. Specifically, a scenario record identifier can be generated for each completed simulation, and the current scenario features, fixed parameters, adjustable parameters, and simulation results can be associated according to this identifier. Therefore, when the scenario record is read subsequently, the simulation operating conditions that produced the corresponding results can be obtained simultaneously, rather than only saving the result data for which parameter conditions cannot be confirmed.
[0234] The second step is to store the simulation scenario records in the simulation scenario library, and record the corresponding simulation rounds and parameter differences.
[0235] The simulation round number represents the current simulation round within a continuous simulation process. Parameter differences represent the parameter changes in the current simulation relative to the previous round or corresponding historical scenarios. In actual storage, each simulation round can be assigned a unique version identifier, and the corresponding simulation round number, current parameters, parameter differences from the previous round, and the simulation result of this round can be saved. Relevant time information can also be saved, allowing multiple continuous simulation scenarios to be organized according to the simulation order. For unchanged basic conditions, they can be retrieved through the association between the current scenario record and the corresponding basic conditions, without needing to indiscriminately store all information in each simulation record.
[0236] In one alternative implementation, a hierarchical time-series storage approach can be adopted, where the basic state can be continuously used, while each round of dynamic state can separately save the corresponding version tag, current parameters, and inference metrics for this round. This method allows for recording parameter and result changes between different rounds and supports subsequent comparisons of different inference versions.
[0237] The third step is to record the simulation scenarios as existing scenarios when new flood simulation needs are obtained in the future, and then use these simulation scenarios to participate in the scenario similarity calculation.
[0238] When new flood simulation needs arise, corresponding scene features can be extracted from the newly added simulation scenario records and combined with other scenarios in the original simulation scenario library as existing scenarios. Then, following the scenario similarity determination method in step S140, the new current scenario features can be compared with these existing scenarios respectively.
[0239] Therefore, each time a new 3D flood simulation is completed, its actual scenario features, simulation parameters, and simulation results can be rewritten into the simulation scenario library. As the number of actual simulations increases, the simulation scenario library can be continuously expanded. When subsequent needs are similar to a newly added scenario, that new scenario can be matched and utilized again as a historical approximation of the operating conditions.
[0240] For example, after the initial simulation of a 130mm rainfall scenario, the 130mm rainfall scenario, corresponding parameters, and results can be written into the simulation scenario library. When subsequent requests are received that are similar to this scenario in terms of rainfall conditions, spatial range, engineering scheduling conditions, and simulation timescale, this record can be used for similarity comparison and may be identified as the target simulation scenario. Thus, the scenario library is no longer solely composed of initial preset scenarios, but can gradually accumulate completed scenarios as the actual simulation process progresses, reducing the number of times similar scenarios need to be reconstructed later.
[0241] In some embodiments, after obtaining the deduction results, associating the deduction results with knowledge objects in the flood knowledge graph, and further obtaining flood risk analysis results, can be done through the following technical steps: The first step is to associate the simulation results with the knowledge objects in the flood knowledge graph.
[0242] The Flood Knowledge Graph can represent a data structure that organizes flood-related professional knowledge in the form of entities, attributes, and relationships. Knowledge objects can represent entities, relationships, or rules in the Flood Knowledge Graph that can establish a corresponding relationship with the simulation scenario, simulation parameters, or simulation results.
[0243] A flood knowledge graph can pre-store professional knowledge such as rainstorm-causing factors, hydraulic characteristics of pipe networks, water system relationships, underlying surface properties, engineering scheduling rules, and risk transmission mechanisms. For example, rainstorm-causing factors can be related to rainfall intensity, rainfall duration, or extreme rainfall conditions; pipe network hydraulic characteristics can be related to pipe network nodes, flow rates, water levels, and pipe network load status; water system relationships can represent the hydraulic connections between different rivers, water bodies, or drainage facilities; underlying surface properties can reflect local surface features; engineering scheduling rules can represent the operating conditions of facilities such as pumping stations and gates; and risk transmission mechanisms can be used to describe the influence relationships between different flood factors.
[0244] Specifically, based on information such as nodes, spatial regions, flood conditions, parameter categories, and engineering objects in the simulation results, corresponding knowledge objects can be identified from the flood knowledge graph, and a correspondence between the result data and the knowledge objects can be established. For example, a high-flow-rate pipeline node can be associated with the corresponding pipeline node entity in the flood knowledge graph, a waterlogged area can be associated with the corresponding underlying surface or spatial region entity, and the current pumping station operating status can be associated with the corresponding engineering scheduling rules.
[0245] The second step involves semantic mapping of the simulation scenario, simulation parameters, and entities, relationships, and rules in the flood knowledge graph, as well as calibrating the corresponding parameter thresholds and resolving conflicting information.
[0246] Semantic mapping can represent the transformation of the meanings of objects and parameters in the inference process into entities, relations, or rules in the flood knowledge graph, resulting in a unified semantic representation that corresponds to each other. Entity mapping can determine the correspondence between nodes, regions, and engineering objects in the inference data and entities in the flood knowledge graph; relation mapping can determine the correspondence between the relationships between objects in the inference results and the relations in the flood knowledge graph; rule mapping can determine the correspondence between the current inference conditions and the professional rules in the flood knowledge graph.
[0247] Threshold calibration refers to adjusting the risk thresholds used in the simulation parameters to a unified comparison condition when there are differences in the units of expression, applicable objects, or level definitions between the professional thresholds in the flood knowledge graph. Conflict information resolution refers to determining the information that can be used for subsequent analysis based on preset rules when there are inconsistencies between the simulation results, scenario parameters, and flood knowledge graph knowledge.
[0248] In practical implementation, a semantic alignment layer can be set up to map entities, relationships, and rules in the flood knowledge graph, and further perform threshold calibration and conflict resolution. For example, if a node in the simulation results uses an internal node number, while the flood knowledge graph uses an engineering object name, the mapping relationship between the node number and the engineering object can be used to identify the same object; if the water accumulation risk threshold in the simulation parameters and the corresponding risk rule in the flood knowledge graph use different units, the units and threshold ranges can be unified first, and then risk rule matching can be performed.
[0249] By using semantic mapping, threshold calibration, and conflict information resolution, the simulation scenarios, simulation parameters, and professional knowledge in the flood knowledge graph can be made to be in a semantically related state, avoiding the inability to perform subsequent rule matching simply because of differences in object names or data representation methods.
[0250] The third step is to constrain the simulation results based on the disaster-causing mechanisms, risk transmission paths, engineering scheduling boundaries, and parameter thresholds in the flood knowledge graph.
[0251] Disaster-causing mechanisms can represent the professional rules describing the relationship between rainfall, hydrology, hydraulics, topography, and engineering status and flood formation in the flood knowledge graph; risk transmission paths can represent the associated paths of flood risk propagation from one object or state to other objects or states; engineering scheduling boundaries can represent the operating conditions and scheduling range that engineering objects can adopt; parameter thresholds can represent the professional critical values used for flood status judgment.
[0252] Specifically, based on the mapping results obtained in the second step, rules and associated paths related to the current scenario can be read from the flood knowledge graph, and these rules can be used to constrain the inference results. For example, when the inference results show a significant increase in water accumulation in a certain area, knowledge relationships related to the underlying surface of the current area, adjacent pipe network nodes, and the operating status of pumping stations can be read; when the inference results show that the water level at a node exceeds the corresponding safety conditions, the hydraulic characteristics and parameter thresholds of the corresponding pipe network can be combined to determine whether the current state meets the corresponding risk rules. The engineering scheduling boundary can also be used to check whether the engineering state interpretation formed during semantic reasoning exceeds the allowed operating conditions of the corresponding engineering object, so that subsequent analysis is constrained by existing flood knowledge, rather than relying solely on language associations to generate results.
[0253] The fourth step is to input the constrained inference results into the semantic reasoning model to conduct flood causation analysis, process correlation analysis, and risk transmission analysis, and obtain flood risk analysis results.
[0254] The semantic reasoning model can be represented as a model capable of semantic analysis of the relationships between flood events by combining structured inference results and professional constraints. The semantic reasoning model can be implemented using a large language model adapted for the flood domain.
[0255] During the joint processing, the flood knowledge graph can provide deterministic professional constraints such as disaster-causing mechanisms, risk transmission paths, scheduling boundaries, and parameter thresholds, while the semantic reasoning model can analyze the relationships between different pieces of information by combining current simulation requirements, simulation scenarios, and multi-round simulation contexts. In this way, the structured rules in the flood knowledge graph and the contextual analysis function of the semantic reasoning model can work together to contribute to the three-dimensional flood simulation results.
[0256] Flood causation analysis can identify the relevant factors causing water accumulation, water level rise, or changes in pipeline load based on current projections and related knowledge. Process correlation analysis can analyze the correlation between different flood states according to the flood evolution time and the connections between related objects. Risk transmission analysis can analyze the subsequent risk objects or risk areas that may be associated with a certain flood state based on the risk transmission paths in the flood knowledge graph.
[0257] For example, when a scenario predicts a high water depth in a certain area, we can correlate it with factors such as high-intensity rainfall, localized low-lying surfaces, high load on adjacent pipe networks, and pump station scheduling. We can then analyze the correlation between these factors and the regional water accumulation based on the relationship paths in the flood knowledge graph. If the pump station operating status is changed in another scenario, we can further analyze the correspondence between changes in engineering scheduling and changes in flood evolution by combining the differences between the two scenarios.
[0258] For example, when high-intensity rainfall continues to flow into the pipeline network, the flow at the nodes increases, followed by rising water levels and localized surface waterlogging. The temporal correspondence and hydraulic association rules in the flood knowledge graph can be used to conduct a process correlation analysis of rainfall conditions, pipeline hydraulic status, and surface waterlogging. In cases where adjacent areas have connections through water systems or surface runoff, the spatial objects involved in risk propagation can also be determined based on these relationships.
[0259] The final flood risk analysis results can include flood causal analysis results, process correlation analysis results, and risk transmission analysis results. Compared with giving the risk status based on only a single numerical indicator, by mapping the numerical results generated by the three-dimensional flood simulation to the professional relationships and rules in the flood knowledge graph, and then analyzing them under professional constraints by the semantic reasoning model, the simulation results of nodal flow, water level, water depth, and inundation range can be further linked to the flood formation process and related engineering conditions.
[0260] In one specific implementation, a joint reasoning approach combining a flood knowledge graph and a large language model can be adopted. The flood knowledge graph stores disaster-causing factors from heavy rainfall, hydraulic characteristics of the pipe network, water system relationships, underlying surface attributes, engineering scheduling rules, and risk transmission mechanisms. It utilizes entity mapping, relation mapping, rule mapping, threshold calibration, and conflict resolution to achieve semantic correspondence. The semantic reasoning model receives the inference results constrained by rules and contextual information relevant to the current needs, performing semantic reasoning on the current flood situation. Through the above processing, comprehensive results of flood causal analysis, process correlation analysis, and risk transmission analysis can be generated.
[0261] Furthermore, in the exemplary embodiments of this disclosure, a three-dimensional flood simulation system based on scene generation and parameter control is also provided. (Refer to...) Figure 6 As shown, the flood 3D simulation system 600 based on scene generation and parameter control includes: a demand analysis module 610, a scene generation module 620, a parameter update module 630, a scene matching module 640, a parameter correction module 650, and a 3D simulation module 660. Wherein: The requirement parsing module 610 can be used to obtain flood simulation requirements, perform semantic parsing on the flood simulation requirements, and generate structured simulation instructions. The scenario generation module 620 can be used to combine and filter the inference parameters based on the regional rainstorm frequency statistics, hydrophysical constraints and parameter combination rules, generate multiple candidate inference scenarios, and build an inference scenario library based on the multiple candidate inference scenarios. The parameter update module 630 can be used to divide the inference parameters into fixed parameters and adjustable parameters. While the fixed parameters remain unchanged, the adjustable parameters that need to be adjusted are determined according to the structured inference instructions, and the adjustable parameters that need to be adjusted are incrementally updated. The scene matching module 640 can be used to extract the current scene features corresponding to the flood simulation requirements, determine the scene similarity between the current scene features and each existing scene feature in the simulation scene library, and determine the target simulation scene based on the scene similarity. The parameter correction module 650 can be used to correct the adjustable parameters in the target simulation scenario based on the difference between the incrementally updated adjustable parameters and the corresponding parameters in the target simulation scenario; The 3D simulation module 660 can be used to perform 3D flood simulation based on the target simulation scenario after parameter correction, and obtain the simulation results.
[0262] The specific details of each module in the above-mentioned flood 3D simulation system based on scene generation and parameter control have been described in detail in the corresponding flood 3D simulation method based on scene generation and parameter control, so they will not be repeated here.
[0263] It should be noted that although several modules or units of the flood three-dimensional simulation system based on scene generation and parameter control have been mentioned in the detailed description above, this division is not mandatory. In fact, according to the embodiments of this disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.
[0264] Furthermore, in an exemplary embodiment of this disclosure, an electronic device is also provided that can implement the above-described flood three-dimensional simulation method based on scene generation and parameter control.
[0265] Those skilled in the art will understand that various aspects of this disclosure can be implemented as a system, method, or program product. Therefore, various aspects of this disclosure can be embodied in the following forms: a completely hardware embodiment, a completely software embodiment (including firmware, microcode, etc.), or an embodiment combining hardware and software aspects, collectively referred to herein as a "circuit," "module," or "system."
[0266] The following reference Figure 7 To describe an electronic device 700 according to such an embodiment of the present disclosure. Figure 7 The electronic device 700 shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments disclosed herein.
[0267] like Figure 7 As shown, the electronic device 700 is presented in the form of a general-purpose computing device. The components of the electronic device 700 may include, but are not limited to: at least one processing unit 710, at least one storage unit 720, a bus 730 connecting different system components (including storage unit 720 and processing unit 710), and a display unit 740.
[0268] The storage unit stores program code that can be executed by the processing unit 710, causing the processing unit 710 to perform the steps described in the "Detailed Description" section of this specification according to various exemplary embodiments of this disclosure. The storage unit 720 may include a readable medium in the form of a volatile storage unit, such as a random access memory (RAM) unit 721 and / or a cache memory unit 722, and may further include a read-only memory unit (ROM) unit 723.
[0269] The storage unit 720 may also include a program / utility 724 having a set (at least one) of program modules 725, including but not limited to: an operating system, one or more application programs, other program modules, and program data, each or some combination of these examples may include an implementation of a network environment.
[0270] Bus 730 can represent one or more of several types of bus structures, including a memory cell bus or memory cell controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any of the various bus structures.
[0271] Electronic device 700 can also communicate with one or more external devices 770 (e.g., keyboard, pointing device, Bluetooth device, etc.), and with one or more devices that enable a user to interact with electronic device 700, and / or with any device that enables electronic device 700 to communicate with one or more other computing devices (e.g., router, modem, etc.). This communication can be performed via input / output (I / O) interface 750. Furthermore, electronic device 700 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 760. As shown, network adapter 760 communicates with other modules of electronic device 700 via bus 730. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with electronic device 700, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.
[0272] Through the description of the above embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, terminal system, or network device, etc.) to execute the methods according to the embodiments of this disclosure.
[0273] In exemplary embodiments of this disclosure, a computer-readable storage medium is also provided, on which a program product capable of implementing the methods described above is stored. In some possible embodiments, various aspects of this disclosure may also be implemented as a program product including program code, which, when the program product is run on a terminal device, causes the terminal device to perform the steps according to various exemplary embodiments of this disclosure described in the "Detailed Description" section above.
[0274] The program product for implementing the above-described flood three-dimensional simulation method based on scene generation and parameter control according to embodiments of this disclosure can employ a portable compact disk read-only memory (CD-ROM) and include program code, and can run on a terminal device, such as a personal computer. However, the program product of this disclosure is not limited thereto. In this disclosure, the readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.
[0275] The program product may employ any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.
[0276] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A readable signal medium may also be any readable medium other than a readable storage medium, capable of sending, propagating, or transmitting programs for use by or in conjunction with an instruction execution system, apparatus, or device.
[0277] The program code contained on the readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination thereof.
[0278] Program code for performing the operations of this disclosure can be written in any combination of one or more programming languages, including object-oriented programming languages such as Java and C++, and conventional procedural programming languages such as C or similar languages. The program code can execute entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).
[0279] Furthermore, the above figures are merely illustrative of the processes included in the method according to exemplary embodiments of this disclosure and are not intended to be limiting. It is readily understood that the processes shown in the above figures do not indicate or limit the temporal order of these processes. Additionally, it is readily understood that these processes may be executed synchronously or asynchronously, for example, in multiple modules.
[0280] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the claims.
[0281] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.
Claims
1. A method for three-dimensional flood simulation based on scene generation and parameter control, characterized in that, include: Obtain flood simulation requirements, perform semantic parsing on the flood simulation requirements, and generate structured simulation instructions; Based on the statistical results of regional rainstorm frequency, hydrophysical constraints, and parameter combination rules, the inference parameters are combined and filtered to generate multiple candidate inference scenarios, and an inference scenario library is constructed based on the multiple candidate inference scenarios. The inference parameters are divided into fixed parameters and adjustable parameters. While the fixed parameters remain unchanged, the adjustable parameters that need to be adjusted are determined according to the structured inference instructions, and the adjustable parameters that need to be adjusted are incrementally updated. Extract the current scene features corresponding to the flood simulation requirements, determine the scene similarity between the current scene features and each existing scene feature in the simulation scene library, and determine the target simulation scene based on the scene similarity; Based on the difference between the incrementally updated adjustable parameters and the corresponding parameters in the target simulation scenario, the adjustable parameters in the target simulation scenario are corrected; A three-dimensional flood simulation was performed based on the target scenario after parameter correction, and the simulation results were obtained.
2. The flood three-dimensional simulation method based on scene generation and parameter control according to claim 1, characterized in that, The process of obtaining flood forecasting requirements, performing semantic parsing on the flood forecasting requirements, and generating structured forecasting instructions includes: Obtain the aforementioned flood simulation requirements; Semantic parsing is performed on the flood simulation requirements to determine the corresponding rainfall parameters, river hydrological parameters, engineering scheduling parameters, pipeline hydraulic parameters, and waterlogging risk parameters; According to the preset instruction format, the determined parameters are encapsulated in a structured manner to obtain the structured deduction instruction.
3. The flood three-dimensional simulation method based on scene generation and parameter control according to claim 2, characterized in that, The semantic parsing of the flood projection requirements includes: Obtain a flood-labeled dataset, which includes the correspondence between natural language questions and standard simulation parameters; Domain adaptation of the semantic parsing model is performed based on the aforementioned flood-annotated dataset; Based on the flood simulation requirements, flood cases, engineering scheduling procedures, hydrological parameter thresholds, and simulation constraints that match the flood simulation requirements are retrieved from the flood domain knowledge base. The retrieved content is adapted to the semantic parsing model of the flood simulation requirement input domain, and the corresponding standard simulation parameters are determined according to the output format constraints.
4. The flood three-dimensional simulation method based on scene generation and parameter control according to claim 1, characterized in that, Based on regional heavy rainfall frequency statistics, hydrophysical constraints, and parameter combination rules, the inference parameters are combined and filtered to generate multiple candidate inference scenarios, including: The range of rainfall parameters corresponding to different rainstorm recurrence periods is determined based on the statistical results of the regional rainstorm frequency. Within the range of rainfall parameters, multiple candidate parameter combinations are obtained by combining rainfall intensity, rainfall duration, rainstorm recurrence period, spatial distribution of rainfall pattern, engineering scheduling status, spatial extrapolation range, and simulation duration. The candidate parameter combinations are verified based on the hydrophysical constraints. According to the parameter combination rules, candidate parameter combinations that have any of the following problems are removed from the candidate parameter combinations that have passed the verification: exceeding the limit of rainfall parameters, conflicting rainfall patterns, conflicting engineering scheduling status, or conflicting spatial boundaries. Multiple candidate scenarios are generated based on the remaining candidate parameter combinations.
5. The three-dimensional flood simulation method based on scene generation and parameter control according to claim 1, characterized in that, The step of determining the adjustable parameters to be adjusted according to the structured deduction instructions includes: Obtain the state of the previous simulation, which includes the fixed parameters and adjustable parameters used in the previous simulation; The target state for this round is determined based on the structured deduction instructions; Compare the current target state with the previous simulation state to determine the corresponding parameter increment; If the parameter increment is zero, the results of the previous round of deduction will be used. If the parameter increment is not zero, the adjustable parameter to be adjusted is determined based on the parameter increment.
6. The flood three-dimensional simulation method based on scene generation and parameter control according to claim 1, characterized in that, The step of determining the scene similarity between the current scene features and each existing scene feature in the scenario database includes: From the flood simulation requirements, rainfall characteristics, spatial simulation range, engineering scheduling conditions, simulation time scale, and risk judgment threshold are extracted to generate a feature vector for the current scene. Obtain the feature vector of each existing scenario; Based on the feature weights corresponding to each feature dimension, calculate the weighted feature distance between the current scene feature vector and each of the existing scene feature vectors; The scene similarity is determined based on the weighted feature distance and the maximum feature distance.
7. The flood three-dimensional simulation method based on scene generation and parameter control according to claim 1, characterized in that, The three-dimensional flood simulation is performed on the target scenario based on the parameter correction to obtain the simulation results, including: Generate the corresponding rainfall process based on the rainfall parameters in the target simulation scenario after parameter correction; Apply the corresponding engineering scheduling conditions according to the engineering scheduling parameters; Based on the rainfall process and the engineering scheduling conditions, hydraulic calculations of the pipeline network are performed to obtain the flow rate and water level at the pipeline nodes. Based on the hydraulic calculation results of the pipeline network, the evolution of surface waterlogging is calculated to obtain the surface water depth and inundation range; The risk assessment is performed based on the flow rate of the pipeline nodes, the water level of the nodes, the depth of surface water accumulation, and the inundation range to obtain the simulation results.
8. The three-dimensional flood simulation method based on scene generation and parameter control according to claim 7, characterized in that, The risk assessment based on the flow rate of the pipeline nodes, the water level of the nodes, the surface water depth, and the inundation range, to obtain the projection results, includes: By associating the flow rate of the pipeline nodes, the water level of the nodes, the depth of surface water accumulation, and the flooding range at the same time point, time-series flood status data corresponding to multiple time points are obtained. Based on a preset risk discrimination threshold, risk discrimination is performed on each of the time-series flood status data to obtain the flood risk status corresponding to each time node; The peak flood time is determined based on multiple time-series flood status data, and the peak flow rate, highest water level, maximum water depth, and inundation range of the node corresponding to the peak flood time are obtained. The projection results are obtained based on the flood risk status corresponding to each time node and the peak flow, highest water level, maximum water depth and inundation range of the node corresponding to the peak flood time.
9. The three-dimensional flood simulation method based on scene generation and parameter control according to claim 8, characterized in that, The step of performing risk assessment on each of the time-series flood status data according to a preset risk assessment threshold to obtain the flood risk status corresponding to each time node includes: Obtain the flow threshold, water level threshold, and risk classification threshold corresponding to the water accumulation depth for each pipeline node; For each of the aforementioned time points, the flow rate of the pipeline node is compared with the corresponding flow rate threshold, the water level of the node is compared with the corresponding water level threshold, and the surface water depth is compared with the corresponding risk classification threshold to obtain the flow rate exceeding limit status, water level exceeding limit status, and water accumulation risk level corresponding to each of the aforementioned time points; Based on the above-mentioned flow rate exceedance status, water level exceedance status, and water accumulation risk level, the flood risk status corresponding to each of the above-mentioned time nodes is determined; The extrapolation results are obtained based on the flood risk status corresponding to each time node and the peak flow, highest water level, maximum water depth, and inundation range of the node corresponding to the flood peak time, including: The duration of the flow rate exceeding the limit is determined based on the time nodes when the flow rate exceeds the limit continuously, and the duration of the water level exceeding the limit is determined based on the time nodes when the water level exceeds the limit continuously. The flow rate exceeding the limit is determined based on the peak flow rate of the node and the corresponding flow rate threshold, and the water level exceeding the limit is determined based on the highest water level of the node and the corresponding water level threshold. The projection results are generated based on the flood risk status corresponding to each time node, the duration of the flow exceeding the limit, the duration of the water level exceeding the limit, the magnitude of the flow exceeding the limit, the magnitude of the water level exceeding the limit, the maximum water depth in the area, and the inundation range.
10. A three-dimensional flood simulation system based on scene generation and parameter control, used to implement the three-dimensional flood simulation method based on scene generation and parameter control as described in any one of claims 1 to 9, characterized in that, The system includes: The requirement parsing module is used to obtain flood simulation requirements, perform semantic parsing on the flood simulation requirements, and generate structured simulation instructions; The scenario generation module is used to combine and filter the inference parameters based on the regional rainstorm frequency statistics, hydrophysical constraints and parameter combination rules, generate multiple candidate inference scenarios, and build an inference scenario library based on the multiple candidate inference scenarios. The parameter update module is used to divide the inference parameters into fixed parameters and adjustable parameters. While keeping the fixed parameters unchanged, it determines the adjustable parameters that need to be adjusted according to the structured inference instructions and performs incremental updates on the adjustable parameters that need to be adjusted. The scene matching module is used to extract the current scene features corresponding to the flood simulation requirements, determine the scene similarity between the current scene features and each existing scene feature in the simulation scene library, and determine the target simulation scene based on the scene similarity. The parameter correction module is used to correct the adjustable parameters in the target simulation scenario based on the difference between the incrementally updated adjustable parameters and the corresponding parameters in the target simulation scenario. The three-dimensional simulation module is used to perform three-dimensional flood simulation based on the target simulation scenario after parameter correction, and obtain the simulation results.
