CN-122022391-A - Flood risk dynamic assessment method and system based on digital twinning

CN122022391ACN 122022391 ACN122022391 ACN 122022391ACN-122022391-A

Abstract

The invention discloses a flood risk dynamic assessment method and system based on digital twinning, and belongs to the technical field of water conservancy informatization. The method comprises the steps of constructing a digital twin body of an urban river basin based on GIS and BIM, establishing real-time mapping with physical sensing equipment, performing minute-scale rolling prediction on a flood evolution process by using a space-time diagram neural network simulation model introducing physical conservation constraint, generating a dynamic risk thermodynamic diagram by combining disaster-bearing body attributes and population exposure models, and generating a control instruction set for a drainage pump station and a traffic gate and issuing and executing when the risk exceeds a threshold value. The invention realizes the minute-level dynamic simulation of flood risk and closed-loop scheduling of physical facilities, and remarkably improves the response speed and the active defense capability of urban flood control and disaster reduction.

Inventors

ZHANG YANG
HAN QING
ZENG JIAN
ZHU DENGBIN
JIANG YUANZHONG
LIU XIN
XU LIMING
FU JUNJIE
WANG ZIMING
CHEN MINGEN

Assignees

杭州定川信息技术有限公司

Dates

Publication Date: 20260512
Application Date: 20260409

Claims (9)

1. The flood risk dynamic assessment method based on digital twinning is characterized by comprising the following steps of: Constructing a digital twin body of the urban river basin, namely constructing a virtual mapping model of the urban river basin based on a Geographic Information System (GIS) and a Building Information Model (BIM), and establishing a real-time data mapping relation between the virtual mapping model and physical sensing equipment; Acquiring real-time meteorological hydrologic data as input characteristics, and inputting the real-time meteorological hydrologic data into a pre-trained physical information enhanced space-time diagram neural network simulation model, wherein the simulation model is based on a heterogeneous diagram structure of a city river, trains through a composite loss function fused with physical conservation constraints, performs minute-level rolling prediction on the flood evolution process, and outputs predicted water depth and flow velocity vectors of all hydraulic topological nodes; Generating a dynamic risk level, namely performing space superposition analysis on the predicted water depth and flow velocity vector and a preset disaster-bearing body physical attribute library, calculating a dynamic risk index by combining a population exposure model, and generating a risk thermodynamic diagram containing the risk level and the submerged range; And (3) performing closed-loop scheduling on the physical facilities, namely generating a control instruction set aiming at a drainage pump station, a traffic gate and emergency broadcasting equipment in the physical world when the dynamic risk index exceeds a preset threshold value, and issuing and executing the control instruction set to change the running state of the physical facilities.
2. The digital twinning-based flood risk dynamic assessment method according to claim 1, wherein the real-time meteorological hydrologic data comprises radar rainfall extrapolation data, river channel water level monitoring data and pipe network liquid level data; The establishing of the real-time data mapping relation between the virtual mapping model and the physical sensing equipment specifically comprises the steps of cleaning multi-source heterogeneous data by utilizing edge computing nodes, aligning time and space, and mapping monitoring data under a physical coordinate system to grid nodes corresponding to the virtual mapping model in real time.
3. The flood risk dynamic assessment method based on digital twinning according to claim 1, wherein the training method of the space-time diagram neural network simulation model comprises the following steps: constructing a training set containing historical flood event samples, wherein the samples contain input meteorological hydrologic feature vectors and output hydrodynamic state labels; physical conservation constraint conditions are introduced in the model training process, and model parameters are optimized by using a back propagation algorithm, so that a predicted result output by the model meets the historical data rule and the hydrodynamic physical rule simultaneously.
4. The flood risk dynamic assessment method based on digital twinning according to claim 1, wherein the calculating the dynamic risk index by combining the population exposure model specifically comprises: Acquiring real-time communication signaling data or video monitoring data, and analyzing population density distribution characteristics of a target area; And calculating risk weights of different areas by adopting a weighted fusion algorithm based on the predicted water depth, the flow velocity vector and population density distribution characteristics, and generating a grading early warning signal.
5. The digital twinning-based flood risk dynamic assessment method according to claim 1, wherein the generating the control instruction set for the drainage pump station, the traffic gate and the emergency broadcasting equipment in the physical world specifically comprises: when the water depth of the tunnel node is predicted to exceed the safety threshold, a traffic signal lamp switching instruction and a barrier gate closing instruction are generated; and when the accumulated water in the predicted area exceeds the water discharge capacity of the pump station, generating a pump station group joint debugging instruction, and changing the operation frequency of the water pump and the opening degree of the gate.
6. A digital twinning-based flood risk dynamic assessment system, applied to the method of any one of claims 1-5, the system comprising: the twin construction unit is used for constructing a virtual mapping model of the urban river basin based on GIS and BIM technologies and establishing a data mapping channel with the physical sensing equipment; The simulation deduction unit is used for loading a trained space-time diagram neural network model, carrying out minute-level simulation on the flood evolution process based on real-time input data, and outputting a predicted water depth and flow velocity vector; The risk assessment unit is used for carrying out superposition analysis on the simulation result, the physical attribute of the disaster-bearing body and the population exposure, calculating a dynamic risk index and generating a risk thermodynamic diagram; The scheduling execution unit is used for generating control instructions for the drainage facility, the traffic facility and the emergency facility according to the risk thermodynamic diagram and issuing the control instructions to the physical execution end.
7. The digital twinning-based flood risk dynamic assessment system according to claim 6, wherein the simulation deduction unit comprises: the diagram structure construction module is used for dividing the urban river basin into a plurality of space grids, and constructing diagram structure data of which the nodes represent hydrologic characteristics and the edges represent hydraulic connection relations; the space-time characteristic extraction module is used for extracting space-dependent characteristics through the graph convolution layer and extracting time sequence characteristics through the time gating layer; And the physical constraint module is used for embedding constraint conditions of a hydrodynamic conservation equation in the model reasoning process and correcting the prediction output.
8. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the steps of the method according to any of claims 1-5 when the program is executed.
9. A computer readable storage medium having stored thereon a computer program, which when executed by a processor performs the steps of the method according to any of claims 1-5.

Description

Flood risk dynamic assessment method and system based on digital twinning Technical Field The invention relates to the technical field of water conservancy informatization and digital twinning, in particular to a flood risk dynamic assessment method and system based on digital twinning. Background With the acceleration of global climate change and urbanization process, extreme rainfall events are frequent, and urban waterlogging is increasingly serious. Accurately and timely assessing flood risk and taking effective scheduling measures are key to reducing disaster loss. Traditional flood risk assessment methods rely primarily on historical data statistics, offline hydraulic models based on physical equations (e.g., SWMM, MIKE) or empirical formulas. These methods suffer from the following inherent drawbacks: static and hysteresis evaluation is generally based on design of rain types or historical extremum, real-time weather and hydrologic monitoring data cannot be integrated, prediction results are seriously delayed from disaster development, and emergency response of 'minute level' is difficult to support. The accuracy and efficiency are contradictory, the high-accuracy physical model has high computational complexity and long time consumption, and cannot meet the requirement of online rolling prediction, the simplified model sacrifices accuracy, and the production convergence and pipe network coupling process under the under-cushion of the complex city is difficult to accurately simulate. The existing method is mostly remained on the 'risk early warning' level, most of output is static risk diagram or grade report, and the existing method is lack of direct linkage control mechanism of physical facilities such as drainage pump station, traffic control, emergency broadcast and the like, so that a visible and invisible information island is formed, and a complete disaster prevention and reduction closed loop cannot be formed. The digital twin technology provides a new solution to the above problems. However, the existing digital twin application focuses on three-dimensional visualization and data integration, and in the core 'simulation deduction' link, either a traditional low-efficiency physical model is still relied on, or a pure data-driven AI model (such as a general cyclic neural network RNN and a convolutional neural network CNN) is adopted, and the model lacks of embedding hydrodynamic physical rules (such as mass conservation and momentum conservation), so that generalization capability is poor under the condition of data scarcity or extreme situations, and a prediction result may violate physical common sense, so that decision basis is unreliable. Therefore, how to construct a closed loop system which can fuse multi-source data in real time, simulate flood evolution rapidly and highly accurately and automatically trigger physical facility control, and realize the spanning from 'static evaluation' to 'dynamic scheduling' becomes a technical problem to be solved in the field. Disclosure of Invention The invention aims to overcome the defects of the prior art and provides a flood risk dynamic assessment method and system based on digital twinning. Aiming at the problems of slow calculation, poor physical consistency of a pure data driving model and disjoint assessment and scheduling of a traditional physical model, the invention constructs a space-time diagram neural network simulation model fused with physical conservation constraint, and realizes minute-level high-precision rolling prediction of flood evolution and automatic closed-loop scheduling of physical facilities. In order to achieve the above purpose, the present invention adopts the following technical scheme: in a first aspect, the present invention provides a flood risk dynamic assessment method based on digital twinning, comprising the steps of: And step S101, constructing a digital twin body of the urban river basin, namely constructing a high-precision three-dimensional virtual mapping model of the urban river basin based on a geographic information system GIS and a building information model BIM and by fusing data of terrains, landforms, buildings, pipe networks and river channels. Meanwhile, a real-time data mapping relation between the virtual mapping model and equipment (such as a rain gauge, a water level gauge, a camera and an Internet of things terminal) deployed in the physical world is established, and real-time synchronization from the physical world state to the virtual space is realized. Step S102, online simulation of a flood evolution process, namely acquiring real-time meteorological hydrologic data (such as radar rainfall extrapolation, river channel water level and pipe network liquid level), inputting the real-time meteorological hydrologic data into a preset space-time diagram neural network simulation model for minute-level rolling prediction, and outputting predicted water depth and flow velocity vectors of all hyd