Mountain torrent simulation early warning method and system based on hydrodynamic model

Abstract

The invention discloses a mountain torrent simulation early warning method and system based on a hydrodynamic model, and relates to the technical field of mountain torrent early warning, and the method comprises the steps: constructing an ice and snow kinetic energy sudden change judgment function TKE, analyzing kinetic energy disturbance caused by an ice and snow sliding speed and a snow thickness change rate in real time, and setting a sudden change threshold value F1 as a triggering condition for starting mountain torrent analysis; and accurate identification of the'non-rainfall-driven 'type mountain torrent event is realized. When the ice and snow kinetic energy sudden change judgment function TKE exceeds a threshold value, the system automatically starts an initial water content estimation model, the precipitation, the snow melting rate, the slope surface temperature change and other elements are fused, the initial water content W0 of the region is calculated, and a comprehensive inflow function Qin is further calculated. And on the basis, constructing a mountain torrent trigger function Fh, comprehensively weighing the slope response capability, the inflow growth rate and the landslide factor, dynamically outputting mountain torrent trigger intensity, and dividing risk grades.

Description

Technical Field

[0001] The present invention relates to the technical field of flash flood warning, and in particular to a flash flood simulation warning method and system based on a hydrodynamic model. Background Art

[0002] This invention relates to the field of natural disaster monitoring and early warning technology, particularly hydrodynamic modeling and disaster simulation in plateau mountainous watersheds. More specifically, it relates to a flash flood disaster prediction technology for high-altitude, cold watersheds, specifically a flash flood disaster simulation and early warning method based on a hydrodynamic model. This method, based on a three-dimensional hydrodynamic simulation platform and integrating remote sensing, the Internet of Things, edge computing, and high-precision snow and ice / hydrological data acquisition, simulates the nonlinear hydrological response of typical rivers in plateau regions under extreme meteorological conditions, focusing on the prediction, response, and risk assessment of flash flood disasters driven by the combined effects of snowmelt and sudden precipitation.

[0003] Current flash flood monitoring and early warning technologies are mostly based on rainfall-driven models, relying primarily on precipitation intensity and rainfall data from meteorological stations as triggering factors. However, in high-altitude mountainous areas, flash floods are often triggered by factors other than rainfall. This means that sudden floods can also occur due to the release of energy from solid-liquid phase transitions caused by sudden temperature rises, enhanced radiation, or rapid snowmelt on slopes. Existing early warning systems fail to consider the temporal interplay of snow and rain responses, typically simplifying the initial state of the watershed to a constant water content or static saturation, ignoring the sudden increase in surface runoff caused by the release of snow and ice water. This simplified modeling approach can lead to misjudgments or delayed responses to "combined snowmelt and heavy rain flash floods." Furthermore, traditional hydrological models are limited in spatial accuracy, response granularity, and dynamic boundary adjustment, making them incapable of supporting real-time early warning requirements in the complex terrain of the plateau. Summary of the Invention

[0004] In view of the deficiencies in the prior art, the present invention provides a flash flood simulation warning method and system based on a hydrodynamic model, which solves the problems mentioned in the background technology.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: comprising the following steps: S1. Several detection points are set up along the plateau rivers, and a collection equipment group is installed at each detection point to collect river ice and snow data in real time. A cloud server is built to transmit the river ice and snow data to the cloud server; S2. In the cloud server, a hydrodynamic model is constructed and simulated based on river ice and snow data. After the simulation, ice and snow data in the hydrodynamic model are extracted and preprocessed to obtain a standard ice and snow dataset. S3. Based on the standard ice and snow data set in the hydrodynamic model, calculate and output the ice and snow kinetic energy mutation judgment function TKE, set the mutation threshold F1 and perform preliminary comparative evaluation with the ice and snow kinetic energy mutation judgment function TKE, and perform flash flood triggering analysis based on the preliminary comparative evaluation results; S4. The flash flood analysis is performed by calculating and outputting the initial water content W0 of the current plateau river source based on the standard ice and snow data set, and calculating and outputting the comprehensive inflow function Qin based on the initial water content W0, and analyzing the soil moisture and comprehensive precipitation and snowmelt at the plateau river source; S5. Calculate and output a flash flood triggering function Fh based on the comprehensive inflow function Qin, and perform a secondary comparative evaluation based on the output result of the flash flood triggering function Fh to analyze the flash flood triggering risk level.

[0006] Preferably, said S1 includes S11 and S12; S11. Divide the source of a plateau river into three source regions, including an upstream region, a midstream region, and a downstream region; and install different detection points every 500 meters in each of the three source regions, including an upstream detection point, a midstream detection point, and a downstream detection point. Different collection equipment groups are installed at each detection point to collect river ice and snow data of the plateau river in real time; The upstream area is located in the source area of ​​the plateau mountainous area; The middle reaches of the river are located in the transition zone at the foot of the mountain; The downstream area is located below the foot of the mountain; The collection equipment installed in the upstream detection point and the midstream detection point includes LiDAR and environmental sensor groups The environmental sensor group includes a thermometer, a soil moisture meter, a rain gauge and a water level meter; The collection equipment installed in the downstream detection point includes an environmental sensor; The river ice and snow data include temperature T, ice and snow thickness Hd, soil moisture Sat, precipitation R and water level H; S12. Build a cloud server and set up a 5G wireless communication network to wirelessly communicate with the communication modules of all the collection devices in the collection device group. Connect each detection point through the Internet of Things technology loT, and upload the river ice and snow data collected by all detection points to the cloud server. At the same time, set up edge computing equipment to perform preliminary data filtering and noise suppression on the river ice and snow data.

[0007] Preferably, said S2 includes S21 and S22; S21, using the three-dimensional model TELEMAC-3D as a framework of the hydrodynamic model in the cloud server to construct the hydrodynamic model; The hydrodynamic model includes an input module, a simulation module, a flash flood analysis module and an early warning module; The input module is used to combine the real-time river ice and snow data received with remote sensing technology and drone assistance to obtain terrain data and snow layer information of plateau rivers, and input them into the hydrodynamic model, and perform spatial interpolation, data fusion, and time interpolation and data completion in the hydrodynamic model; The terrain data includes terrain, slope, width and river channel information of the plateau river; The snow layer information includes snow density, snow thickness Hd, area and humidity; The spatial difference is interpolated by using the Kriging interpolation method on terrain data with a resolution lower than 90 meters, thereby increasing the terrain data to a resolution of 5 meters; The data fusion is performed by using a GIS platform to set consistent river spatial data for river ice and snow data, terrain data and snow layer information; The time difference and data completion are carried out by temporally interpolating river ice and snow data, terrain data and snow layer information, optimizing the data continuity in the simulation process, and filling in the missing data by interpolation method; The simulation module uses the river spatial data integrated based on the GIS platform to load river ice and snow data, terrain data and snow layer information, and performs simulation based on the hydrological movement SCS-CN model and hydrodynamic simulation software HEC-RAS to simulate the dynamic response of water flow and snow melting in plateau rivers. At the same time, the point information of the detection point is loaded into the river spatial data, and the spatial grid coordinates are embedded in the river space. The spatial grid coordinates are composed of longitude and latitude, and then the longitude and latitude position (x, y) coordinates of the detection point are displayed on the river spatial data, where x represents the spatial position in the longitude direction and y represents the spatial position in the latitude direction. The water flow dynamic response is achieved by simulating how the water flow in the plateau river channel changes with time, temperature T and precipitation R. The snow melt simulation simulates snow melting based on temperature T and snow layer information, and uses temperature as a driving force. The amount of snowmelt water at each moment in the snowmelt process is simulated based on the snow layer information, and the snowmelt water volume is summed with the rainfall R as the water source input simulation for plateau rivers. S22. Based on the simulation module of the hydrodynamic model, the simulation of plateau rivers and snowmelt is carried out. During the simulation process, the ice and snow data of different regions are collected in real time, and the ice and snow data are pre-processed to obtain a standard ice and snow data set; The preprocessing includes normalization and timestamp alignment; The normalization process is performed by using the z-score standard processing method to eliminate the dimensional effects of all parameters in the ice and snow data; The timestamp alignment is performed by aligning the data extraction time with the real-time data of the real time; The standard ice and snow dataset includes the ice and snow density P at time t, the residual snow thickness hs(t) at time t, the slope function Ts at time t, the snow melt rate SRF(t) at time t, the precipitation R(t) at time t, the water level H(t) at time t, and the slope area S.

[0008] Preferably, said S3 includes S31 and S32; S31, the flash flood analysis module includes snowmelt mutation analysis and flash flood analysis; The snowmelt mutation analysis constructs an ice and snow kinetic energy mutation algorithm model in the flash flood analysis module, extracts a standard ice and snow data set and inputs it into the ice and snow kinetic energy mutation algorithm model, calculates and outputs the ice and snow kinetic energy mutation judgment function TKE, and analyzes the dynamic change of the ice and snow water tank; The ice and snow kinetic energy mutation judgment function TKE is calculated and output by the following ice and snow kinetic energy mutation algorithm model; ; Where TKE(t) represents the judgment function of the sudden change of ice and snow kinetic energy at time t, t0 represents the initial time, t1 represents the terminal time, vs(t) represents the surface sliding velocity at time t, and is dimensionless. fmelt(T, Ts) represents the melting factor function under the influence of temperature and slope surface, d represents the calculus variable, and dt represents the time calculus variable.

[0009] Preferably, in S32, a critical value of the kinetic energy of the current plateau river's kinetic energy generated by the melting of ice and snow affecting flash floods is set as a mutation threshold F1, and a preliminary comparative evaluation is performed between the mutation threshold F1 and the real-time acquired ice and snow kinetic energy mutation judgment function TKE to analyze the situation of the ice and snow kinetic energy mutation, and a flash flood triggering analysis is performed based on the preliminary comparative evaluation results. The specific evaluation content is as follows; When the snow and ice kinetic energy mutation judgment function TKE (t) at time t is greater than the mutation threshold F1, it indicates that the snow and ice melting kinetic energy mutation is abnormal, and flash flood analysis is triggered at this time; When the ice and snow kinetic energy mutation judgment function TKE (t) at time t ≤ the mutation threshold F1, it means that the ice and snow melting kinetic energy mutation is normal and continuous monitoring is required.

[0010] Preferably, said S4 includes S41 and S42; S41. After triggering flash flood analysis based on preliminary comparative assessment, extract the snow melt rate SRF(t) and precipitation R(t) at time t from the standard ice and snow dataset, calculate and output the initial soil moisture content W0 in the plateau river area, and analyze the current initial moisture state of the plateau river area; The initial water content W0 is calculated and outputted by the following algorithm formula; ; Where t-t0 represents the relative time difference, W0(x, y, t) represents the initial water content at the position (x, y) at time t, e represents the exponential function, and Sat max (x, y) represents the maximum soil moisture at position (x, y), represents the regulating factor of snowmelt on water content, It represents the coefficient of influence of temperature change on the reaction rate of water content, △t represents the time interval, dt represents the time calculus variable, Represents the slope influence function at position (x, y).

[0011] Preferably, S42, based on the obtained initial water content W0, combined with the precipitation R and snow melt rate SRF at different locations in the standard ice and snow dataset, calculate and output a comprehensive inflow function Qin, and analyze the impact of precipitation and snow melt on mountain torrents under the dual driving conditions of precipitation and snow melt; The comprehensive inflow function Qin is calculated and outputted by the following algorithm formula: ; Where Qin(x, y, t) represents the integrated inflow function at the location (x, y) at time t, R(x, y, t) represents the precipitation at the location (x, y) at time t, and fm(T) represents the temperature-driven melt function. represents the precipitation and snowmelt inflow coefficient.

[0012] Preferably, said S5 includes S51 and S52; S51, based on the real-time acquired comprehensive inflow function Qin and the slope area S, a comprehensive calculation is performed to output a flash flood triggering function Fh to measure the flash flood triggering situation; The flash flood trigger function Fh is calculated and output by the following algorithm formula: ; Where Fh(x, y, t) represents the flash flood triggering function at position (x, y) at time t, S(x, y) represents the slope area at position (x, y), and X(x, y) represents the landslide index. Represents the risk amplification factor.

[0013] Preferably, S52, performing a secondary comparative evaluation based on the output result of the flash flood trigger function Fh, and generating a flash flood trigger risk level based on the secondary comparative evaluation result, and generating relevant warnings based on the flash flood trigger risk level. The specific evaluation content is as follows; When the flash flood trigger function Fh(x, y, t) at the position (x, y) at time t is less than 0.3, it indicates that the kinetic energy of ice and snow is abnormal but there is no flash flood risk. The current position (x, y) is marked as the first-level risk level, and the detection point acquisition frequency is increased by 50%; When 0.3≤the flash flood trigger function Fh(x, y, t) at position (x, y) at time t<0.8, it indicates a critical flash flood risk and the current position (x, y) is marked as a level 2 risk. At this time, the early warning module of the hydrodynamic model issues a level 1 warning message, prompting the start of the planned opening of the midstream dam. When the flash flood trigger function Fh(x, y, t) at the location (x, y) at time t is ≥ 0.8, it indicates that there is a flash flood risk. At this time, the current location (x, y) is marked as a level 3 risk level. At this time, the early warning module of the hydrodynamic model issues a level 2 warning message, prompting the immediate opening of the floodgates to release the floodwaters and pushing an evacuation message to the current area. Among them, the first-level risk level is less than the second-level risk level and less than the third-level risk level, and the third-level risk level is the highest risk level; At the same time, the second-level warning information is the highest level warning.

[0014] A flash flood simulation and early warning system based on a hydrodynamic model, comprising a detection point setting module, a hydrodynamic modeling module, an ice and snow kinetic energy analysis module, a comprehensive inflow analysis module, and a flash flood early warning module; The detection point setting module arranges several detection points along the plateau river line, and installs a collection equipment group in each detection point to collect river ice and snow data in real time, and builds a cloud server to transmit the river ice and snow data to the cloud server; The hydrodynamic modeling module constructs a hydrodynamic model in a cloud server and simulates the hydrodynamic model based on river ice and snow data. After the simulation, the ice and snow data in the hydrodynamic model are extracted and preprocessed to obtain a standard ice and snow data set. The ice and snow kinetic energy analysis module calculates and outputs the ice and snow kinetic energy mutation judgment function TKE based on the standard ice and snow data set in the hydrodynamic model, sets the mutation threshold F1 and performs preliminary comparative evaluation with the ice and snow kinetic energy mutation judgment function TKE, and triggers flash flood analysis based on the preliminary comparative evaluation results; The comprehensive inflow analysis module calculates and outputs the initial water content W0 of the current plateau river source based on the standard ice and snow data set, and calculates and outputs the comprehensive inflow function Qin based on the initial water content W0 to analyze the soil moisture and comprehensive precipitation and snowmelt at the plateau river source; The flash flood warning module calculates and outputs a flash flood trigger function Fh based on the comprehensive inflow function Qin, and performs a secondary comparative evaluation based on the output result of the flash flood trigger function Fh to analyze the flash flood trigger risk level.

[0015] The present invention provides a flash flood simulation and early warning method and system based on a hydrodynamic model. It has the following beneficial effects: (1) This method constructs a snow and ice kinetic energy mutation judgment function TKE(t) to analyze the kinetic energy disturbance caused by the snow and ice sliding speed and snow thickness change rate in real time, and sets a mutation threshold F1 as the trigger condition for starting flash flood analysis, thereby achieving accurate identification of "non-precipitation-driven" flash flood events. When TKE exceeds the threshold, the system automatically starts the initial water content estimation model, integrates factors such as precipitation, snowmelt rate, and slope temperature change, calculates the regional water content W0(x, y, t), and further calculates the comprehensive inflow function Qin. On this basis, a flash flood trigger function Fh is constructed to comprehensively balance the slope response capacity, inflow growth rate, and landslide factor, dynamically output the flash flood trigger intensity, and divide the risk level. Compared with the existing technology that relies on sudden precipitation increase as the basis for early warning, the present invention greatly improves the ability to identify sudden snowmelt flash flood events in advance through kinetic energy mutation modeling and snowmelt-soil water composite analysis.

[0016] (2) This method embeds the spatial grid and the coordinates of the detection point locations in the hydrodynamic model and outputs the flash flood trigger function Fh (x, y, t) corresponding to each detection point in real time. By setting 0.3 and 0.8 as the critical assessment interval, a three-level risk level system is constructed: when Fh < 0.3, it is marked as a level 1 risk, prompting an increase in the acquisition frequency; when Fh is between 0.3 and 0.8, it is marked as a level 2 risk, automatically triggering a level 1 warning for the midstream dam and executing a planned flood discharge; when Fh ≥ 0.8, it is marked as a level 3 risk, immediately issuing a flood gate opening and personnel evacuation order. This mechanism not only realizes the automatic closed-loop control of risk level classification and response, but also performs iterative correction based on real-time data and dynamic model feedback. Compared with the traditional warning system based on static rainfall threshold, the present invention realizes an intelligent flash flood warning solution that integrates "perception, simulation, evaluation, and response", significantly improving the response speed, execution efficiency, and management accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a schematic diagram of the steps of a flash flood simulation and early warning method based on a hydrodynamic model of the present invention; Figure 2 This is a flow chart of a flash flood simulation and early warning system based on a hydrodynamic model according to the present invention; Figure 3 Schematic diagram of the hydrodynamic model; Figure 4 The graph below shows the trend of flash flood triggering function Fh changing over time under different risk levels. DETAILED DESCRIPTION

[0018] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0019] Example 1 See also Figure 1 、 Figure 3 and Figure 4 The present invention provides a flash flood simulation warning method based on a hydrodynamic model. To achieve the above purpose, the present invention is implemented through the following technical solutions: comprising the following steps: S1. Several detection points are set up along the plateau rivers, and a collection equipment group is installed at each detection point to collect river ice and snow data in real time. A cloud server is built to transmit the river ice and snow data to the cloud server; S2. In the cloud server, a hydrodynamic model is constructed and simulated based on river ice and snow data. After the simulation, ice and snow data in the hydrodynamic model are extracted and preprocessed to obtain a standard ice and snow dataset. S3. Based on the standard ice and snow data set in the hydrodynamic model, calculate and output the ice and snow kinetic energy mutation judgment function TKE, set the mutation threshold F1 and perform preliminary comparative evaluation with the ice and snow kinetic energy mutation judgment function TKE, and perform flash flood triggering analysis based on the preliminary comparative evaluation results; S4. Flash flood analysis: Based on the standard ice and snow dataset, the initial water content W0 of the current plateau river source is calculated and output. Based on the initial water content W0, the integrated inflow function Qin is calculated and output to analyze the soil moisture and integrated precipitation and snowmelt at the plateau river source. S5. Calculate and output a flash flood triggering function Fh based on the comprehensive inflow function Qin, and perform a secondary comparative evaluation based on the output result of the flash flood triggering function Fh to analyze the flash flood triggering risk level.

[0020] In this embodiment, the method divides the plateau river into upstream, midstream and downstream areas according to the terrain characteristics, deploys multiple groups of differentiated collection and detection points, and installs integrated collection equipment groups at each point. The collected river ice and snow data are uploaded to the cloud server after preliminary cleaning using the 5G wireless communication network and edge computing equipment; secondly, a hydrodynamic simulation platform based on the TELEMAC-3D framework is constructed in the cloud server, integrating hydrological-hydraulic model components such as SCS-CN and HEC-RAS, and loading high-resolution terrain and ice and snow data to dynamically simulate the hydrological and dynamic evolution process of the current river, and then extract the standardized output during the simulation process. The system then constructs a TKE model for determining the kinetic energy mutation of ice and snow based on the standard ice and snow dataset. This model identifies sudden ice and snow power release and compares it with the preset mutation threshold F1 in real time. When an abnormal kinetic energy mutation is detected, subsequent flash flood analysis is automatically triggered. Based on the TKE trigger signal, the initial water content of the region, W0, is further calculated. The integrated inflow function, Qin, is then integrated with real-time rainfall and snowmelt data to accurately assess the regional soil moisture status and potential surface runoff formation trends. Finally, the flash flood trigger function, Fh, is calculated using Qin and slope response parameters. Based on the Fh result, a level 1 to 3 risk rating is output, enabling the classification of flash flood events and intelligent early warning linkage control. By establishing a complete mechanism chain of "kinetic energy mutation triggering, hydrodynamic modeling, inflow coupling determination, and risk level output," the present invention significantly breaks through the traditional flash flood early warning model based on rainfall intensity. Without changing existing hardware, it achieves the early identification of compound flash floods driven by the superposition of heavy rain and snowmelt. This method effectively improves the response sensitivity and judgment accuracy when facing "non-rainfall type" or "sudden sliding type" flood events, avoids the problems of false alarms and missed reports of disasters due to the failure to capture kinetic energy changes, and thus significantly enhances the intelligence, precision and timeliness of disaster prevention and mitigation in the river basin.

[0021] Example 2 See also Figure 1 ,Specifically: S1 includes S11 and S12; S11. Divide the source of the plateau rivers into three source regions, including upstream, midstream, and downstream regions; and install different detection points every 500 meters in each of the three source regions, including upstream detection points, midstream detection points, and downstream detection points. Different collection equipment groups are installed at each detection point to collect real-time river ice and snow data on the plateau rivers; The upstream area is located in the source region of the plateau and mountainous areas; The middle reaches are located in the transition zone at the foot of the mountain; The downstream area is located below the foot of the mountain; The collection equipment installed at the upstream and midstream detection points includes LiDAR and environmental sensor groups. The environmental sensor set includes a thermometer, soil moisture meter, rain gauge, and water level gauge; The collection equipment installed in the downstream detection points includes environmental sensors; River ice and snow data include temperature T, ice and snow thickness Hd, soil moisture Sat, precipitation R and water level H; S12. Build a cloud server and set up a 5G wireless communication network to wirelessly communicate with the communication modules of all the collection devices in the collection device group. Connect each detection point through the Internet of Things technology loT, and upload the river ice and snow data collected by all detection points to the cloud server. At the same time, set up edge computing equipment to perform preliminary data filtering and noise suppression on the river ice and snow data.

[0022] In this embodiment, the method scientifically divides the source of plateau rivers into three regions: upstream, midstream, and downstream. Monitoring points are deployed every 500 meters based on the topographical characteristics and hydrological behavior of each region, and differentiated collection equipment sets are deployed according to local conditions. LiDAR and environmental sensors, including thermometers, soil moisture meters, rain gauges, and water level gauges, are installed in the upstream and midstream regions to accurately collect core ice and snow hydrological parameters such as temperature, ice and snow thickness, soil moisture, precipitation, and water level. Environmental sensors are primarily deployed in the downstream region to enhance the sensitivity of downstream flood response. In S12, a cloud server based on 5G communications is further constructed, incorporating IoT technology to achieve wireless connectivity among all monitoring points. Relying on the communication modules built into each collection device for rapid data transmission, preliminary calculations and noise filtering are deployed at the edge to enable rapid screening and preprocessing of raw ice and snow data. This zoning and intelligent sensing deployment not only enables the real-time collection and upload of dynamic, distributed ice and snow data for the entire plateau river basin, but also effectively addresses the existing challenges of difficult data collection, insufficient coverage, and slow response times in plateau regions. The implementation of this plan has greatly improved the perception accuracy and data refresh frequency of sudden snowmelt, local heavy rainfall and surface runoff at upstream sources, providing high-quality and continuous data support for the precise simulation of subsequent hydrodynamic models, thereby significantly improving the front-end perception capability, spatial response resolution and warning timeliness of the overall flash flood forecast.

[0023] Example 3 See also Figure 1 and Figure 3 , specifically: S2 includes S21 and S22; S21, using the three-dimensional model TELEMAC-3D as a framework of the hydrodynamic model in the cloud server to construct the hydrodynamic model; The hydrodynamic model includes an input module, a simulation module, a flash flood analysis module, and an early warning module; The input module is used to combine the real-time river ice and snow data received with remote sensing technology and drone assistance to obtain terrain data and snow layer information of plateau rivers, and input them into the hydrodynamic model. The model then performs spatial interpolation, data fusion, and temporal interpolation and data completion. Topographic data include topography, slope, width and channel information of plateau rivers; Snow layer information includes snow density, snow and ice thickness Hd, area and humidity; The spatial difference is interpolated by using the Kriging interpolation method to interpolate the terrain data with a resolution lower than 90 meters, and the terrain data is improved to a resolution of 5 meters; Data fusion is achieved by using a GIS platform to set up consistent river spatial data for river ice and snow data, terrain data, and snow layer information; Time difference and data completion optimize the data continuity during the simulation process by performing temporal interpolation on river ice and snow data, terrain data, and snow layer information. Missing data are filled by interpolation to ensure data integrity and accuracy. The simulation module uses the river spatial data integrated based on the GIS platform to load river ice and snow data, terrain data and snow layer information. It simulates the dynamic response of water flow and snow melting in plateau rivers based on the hydrological movement SCS-CN model and hydrodynamic simulation software HEC-RAS. At the same time, the point information of the detection point is loaded into the river spatial data, and the spatial grid coordinates are embedded in the river space. The spatial grid coordinates are constructed by longitude and latitude, and then the longitude and latitude position (x, y) coordinates of the detection point are displayed on the river spatial data, where x represents the spatial position in the longitude direction and y represents the spatial position in the latitude direction. The dynamic response of water flow is achieved by simulating how the water flow in the plateau river channel changes with time, temperature T and precipitation R; The snow melt simulation is driven by temperature T and snow layer information. The amount of snowmelt water at each moment in the snowmelt process is simulated based on the snow layer information. The snowmelt water amount is then combined with the rainfall R to simulate the water source input of the plateau rivers. S22. Based on the simulation module of the hydrodynamic model, the simulation of plateau rivers and snowmelt is carried out. During the simulation process, the ice and snow data of different regions are collected in real time, and the ice and snow data are pre-processed to obtain a standard ice and snow data set; Preprocessing includes normalization and timestamp alignment; Normalization is done by using the z-score standard processing method to eliminate the dimensional effects of all parameters in the ice and snow data; Timestamp alignment is achieved by aligning the data extraction time with the real-time data in real time; The standard snow and ice dataset includes the snow density P at time t, the residual snow thickness hs(t) at time t, the slope function Ts at time t, the snow melt rate SRF(t) at time t, the precipitation R(t) at time t, the water level H(t) at time t, and the slope area S.

[0024] In this example, a three-dimensional plateau hydrodynamic model based on TELEMAC-3D was constructed to dynamically simulate and extract standardized data for the complex hydrological and snow-ice interactions of plateau rivers. Specifically, the hydrodynamic model, built on a cloud server, comprises an input module, a simulation module, a flash flood analysis module, and an early warning module. It receives real-time river ice and snow data collected from various monitoring points and integrates remote sensing and drone-assisted data such as slope elevation, river channel width, and snow thickness. The diverse data is spatially unified and integrated using a GIS platform. Kriging interpolation is used to improve the low-resolution DEM to 5-meter accuracy, while time interpolation and missing data completion are performed to ensure the integrity and accuracy of the model input. During the simulation phase, the SCS-CN hydrological model and the HEC-RAS hydrodynamic tool are combined to simulate water level fluctuations, flow responses, and temperature-driven snow melt processes in plateau rivers in real time. Each monitoring point is embedded in a grid space, enabling precise (x, y) coordinate-based location and regional response analysis. Subsequently, key ice and snow evolution parameters are extracted in real time during the simulation process, and the extracted data are normalized and timestamp aligned, thereby constructing a standard ice and snow dataset with a unified format and unified time sequence. This dataset includes ice and snow density, snow thickness, slope factor, snowmelt rate, precipitation, water level and slope area at time t, which is the basic support for subsequent kinetic energy mutation identification and mountain torrent analysis. Through the above-mentioned implementation, the present invention realizes a mountain hydrodynamic simulation system with high spatiotemporal resolution, multi-data fusion, and high-precision three-dimensional modeling. Compared with the traditional two-dimensional static modeling method, it significantly improves the multi-source coupling modeling capability and ice and snow mutation identification capability of precipitation-snowmelt-water flow processes in high-altitude cold areas. This method can effectively support the dynamic assessment of extreme ice and snow hydrological events, basin-level response prediction and model iterative correction, greatly enhancing the practical ability of hydrodynamic models in early identification and response accuracy of mountain torrents, and laying a high-quality data foundation for subsequent risk triggering mechanisms.

[0025] Example 4 See also Figure 1 ,Specifically: S3 includes S31 and S32; S31, flash flood analysis module includes snowmelt mutation analysis and flash flood analysis; Snowmelt mutation analysis builds an ice and snow kinetic energy mutation algorithm model in the flash flood analysis module, extracts standard ice and snow data sets and inputs them into the ice and snow kinetic energy mutation algorithm model, calculates and outputs the ice and snow kinetic energy mutation judgment function TKE, and analyzes the dynamic change of the ice and snow water tank; The ice and snow kinetic energy mutation judgment function TKE is calculated and output by the following ice and snow kinetic energy mutation algorithm model; ; Where TKE(t) represents the judgment function of the sudden change of ice and snow kinetic energy at time t, t0 represents the initial time, t1 represents the terminal time, vs(t) represents the surface sliding velocity at time t, which is dimensionless and represents the flow velocity of the ice and snow layer during the melting process, and can be approximated as the water flow velocity. fmelt(T, Ts) represents the melting factor function under the influence of temperature and slope, that is, the efficiency of snow melting changes under different temperature and slope conditions, which plays a regulating role and is extracted after simulation of the hydrodynamic model. d represents the calculus variable, and dt represents the time calculus variable. It represents the kinetic energy change rate term, which indicates the change in the motion state of the unit mass of ice and snow layer from unit time t0 to time t1, that is, the kinetic energy release rate; The physical meaning of the formula is that it is used to calculate the change in kinetic energy released by the ice and snow layer during the melting process. It describes the sudden change in kinetic energy caused by the melting and sliding of the ice and snow layer during the melting process. This sudden change represents the initial driving effect of meltwater on surface water flow.

[0026] S32. A kinetic energy critical value for the impact of the kinetic energy generated by the current melting of ice and snow in plateau rivers on flash floods. For example, when the ice and snow kinetic energy mutation determination function TKE = 10, the probability of flash floods is high, and the mutation threshold F1 is set to 10. This is set as the mutation threshold F1, and a preliminary comparative evaluation is performed between the mutation threshold F1 and the real-time acquired ice and snow kinetic energy mutation determination function TKE to analyze the situation of the ice and snow kinetic energy mutation. Based on the preliminary comparative evaluation results, a flash flood triggering analysis is performed. The specific evaluation content is as follows; When the ice and snow kinetic energy mutation judgment function TKE (t) at time t is greater than the mutation threshold F1, it indicates that the ice and snow melting kinetic energy mutation is abnormal, and flash flood analysis is triggered at this time; When the ice and snow kinetic energy mutation judgment function TKE (t) at time t ≤ the mutation threshold F1, it means that the ice and snow melting kinetic energy mutation is normal and continuous monitoring is required.

[0027] In this embodiment, the method significantly improves the ability to identify and respond to "non-rainfall-type" sudden flash floods by using the potential kinetic energy mutation behavior during the melting of ice and snow as an early judgment signal for flash flood triggering for the first time. Specifically, the flash flood analysis module integrates the dual modules of snowmelt mutation analysis and flash flood analysis. By constructing an ice and snow kinetic energy mutation algorithm model, the time series snow thickness, surface sliding speed and slope temperature factors extracted from the standard ice and snow data set are input into the model, and the ice and snow kinetic energy mutation judgment function TKE is calculated and output. This function expresses the kinetic energy change rate caused by the melting and sliding of the ice and snow layer in a specified time period in a differential form, reflecting the transition behavior of ice and snow water from "static energy storage" to "flow thrust", thereby realizing real-time quantification of potential flood energy release signals. The kinetic energy mutation threshold F1 is set based on historical data and model sensitivity analysis, and the current TKE value is compared with F1 in real time. When TKE>F1, it indicates that the ice and snow layer has undergone a drastic phase change and slip, which may cause a sudden surge in runoff in the downstream river channel, and the flash flood analysis is immediately triggered. If TKE≤F1, it means that the current ice and snow state is changing slowly or within the safety threshold, and it enters the continuous monitoring state. After the implementation of this method, it is possible to identify the "abnormal energy release" process caused by the phase change of ice and snow in advance before the flash flood manifests as significant rainfall or runoff. Compared with the traditional method that relies on rainfall thresholds or flow rate accumulation, it has achieved a breakthrough improvement in the warning time window, model sensitivity and forward-looking disaster prediction. It is particularly suitable for high-altitude mountainous areas or source basins where heavy rain and snowmelt overlap, effectively reducing the misjudgment rate and warning lag risk of sudden flash floods, and further enhancing the perception and response capabilities of the complex triggering mechanism of multi-source driven flash flood events.

[0028] Example 5 See also Figure 1 , specifically: S4 includes S41 and S42; S41. After triggering flash flood analysis based on preliminary comparative assessment, extract the snow melt rate SRF(t) and precipitation R(t) at time t from the standard ice and snow dataset, calculate and output the initial soil moisture content W0 in the plateau river area, and analyze the current initial moisture state of the plateau river area; The initial water content W0 is calculated and output by the following algorithm formula; ; Where t-t0 represents the relative time difference, which is used to indicate how long the mutation process lasts, W0(x, y, t) represents the initial water content at the position (x, y) at time t, e represents the exponential function, and Sat max (x, y) represents the maximum soil moisture at position (x, y), which is the maximum amount of water that the soil can absorb per unit area. It is set by the initial construction of the hydrodynamic model according to the soil moisture characteristics. represents the regulating factor of snowmelt on water content, It represents the effect coefficient of temperature change on the water content reaction rate, and The specific value is set by the user and is dimensionless. △t represents the time interval, and dt represents the time calculus variable. The slope influence function at the position (x, y) represents the effect of the terrain on the water flow. It generally takes a dimensionless value in the range of 0.8-1.2. The larger the value, the more obvious the acceleration effect of the slope on the water flow. The physical meaning of the formula is to calculate the initial water content of an area at a certain moment, taking into account the influence of precipitation, snowmelt rate and temperature changes, to help the model determine whether the current area is close to the dangerous critical water level, and thus determine whether it has entered the flood triggering stage.

[0029] S42. Based on the obtained initial water content W0, combined with the precipitation R and snow melt rate SRF at different locations in the standard ice and snow dataset, the comprehensive inflow function Qin is calculated and output to analyze the impact of precipitation and snow melt on flash floods under the dual driving conditions of precipitation and snow melt. The comprehensive inflow function Qin is calculated and output by the following algorithm formula; ; Where Qin(x, y, t) represents the integrated inflow function at location (x, y) at time t, R(x, y, t) represents the precipitation at location (x, y) at time t, and fm(T) represents the temperature-driven melt function, which represents the effect of temperature on snowmelt flow. It is extracted through the hydrodynamic model and has a dimensionless value. represents the inflow coefficient of precipitation and snowmelt, with a value of 0.5, indicating the contribution ratio of precipitation and snowmelt to water flow; The physical meaning of the formula lies in that it integrates the influence of precipitation and snowmelt water flow, while taking into account the sensitivity of water content to water flow response. The higher the water content, the more sensitive the water flow response and the greater the probability of flash floods.

[0030] In this embodiment, the method realizes a fine-grained assessment mechanism of the water content state and inflow intensity of plateau river areas before flash floods are triggered, and constructs a calculation chain of "initial response capacity identification and inflow sensitivity response" under the dual driving conditions of precipitation and snowmelt. Specifically, when the snow kinetic energy mutation function TKE exceeds the warning threshold, the snow melt rate SRF and precipitation R at the current moment in the standard snow and ice dataset are automatically extracted, and based on the slope response function, temperature change index and maximum soil moisture saturation capacity Sat maxAn initial water content estimation model is constructed. This model dynamically depicts the saturation trend of regional soil after a sudden change in ice and snow, quantifying whether it is approaching the critical state of runoff, and providing a pre-judgment basis for whether it has entered the flash flood triggering zone. Then, based on the estimated initial water content W0, combined with the precipitation and snowmelt rate at the current location, the temperature-driven melting response function fm(T) is used to construct a comprehensive inflow function Qin. This function not only integrates external water source input but also introduces the current water content as a nonlinear adjustment factor for the water flow response. This allows the model to accurately determine whether there are hidden triggering risks such as slope runoff accumulation, waterlogging, or landslide precursors in the "low-intensity multi-source superposition" scenario. Through the implementation of the above method, the present invention effectively solves the single-cause trigger judgment mode of "only considering precipitation intensity but not soil response ability" in the existing early warning mechanism. For the first time, it coherently integrates the dynamic processes of sudden ice and snow conversion ability, soil moisture changes, and inflow enhancement, and significantly improves the mountain torrent model's perception of slope pre-saturation areas, critical catchment areas, and mixed disaster processes of slow melting and slow decline.

[0031] Example 6 See also Figure 1 and Figure 4 ,Specifically: S5 includes S51 and S52; S51, based on the real-time acquired comprehensive inflow function Qin and the slope area S, a comprehensive calculation is performed to output a flash flood triggering function Fh to measure the flash flood triggering situation; The flash flood trigger function Fh is calculated and outputted by the following algorithm formula; ; Where Fh(x, y, t) represents the flash flood triggering function at position (x, y) at time t, S(x, y) represents the slope area at position (x, y), and X(x, y) represents the landslide index, which is a comprehensive sensitivity coefficient of terrain, soil, and vegetation. It is used to reflect whether the slope is prone to water accumulation when encountering a large amount of inflow, such as the end of the valley and the depression, which are prone to landslides, loose soil and exposed rock surface, forming a barrier lake or local collapse. After geological surveys, it is set for the soil types of different three-source areas. Represents the risk amplification factor, which controls the sensitivity of the overall model to the input variables. Different weights are set by the user according to the regional disaster prevention level or policy warning strategy. represents the inflow growth rate; The physical meaning of the formula is that it integrates factors such as inflow, slope area, and landslide index to assess the risk level of flash floods. The greater the inflow, the steeper the slope, and the greater the landslide risk, the higher the probability of triggering a flash flood.

[0032] S52: Perform a secondary comparative evaluation based on the output result of the flash flood trigger function Fh, and generate a flash flood trigger risk level based on the secondary comparative evaluation result. At the same time, generate relevant warnings based on the flash flood trigger risk level. The specific evaluation content is as follows; When the flash flood trigger function Fh(x, y, t) at the position (x, y) at time t is less than 0.3, it indicates that the kinetic energy of ice and snow is abnormal but there is no flash flood risk. The current position (x, y) is marked as the first-level risk level, and the detection point acquisition frequency is increased by 50%; When 0.3≤the flash flood trigger function Fh(x, y, t) at position (x, y) at time t is less than 0.8, it indicates a critical flash flood risk and the current position (x, y) is marked as a level 2 risk. At this time, the early warning module of the hydrodynamic model issues a level 1 warning message, prompting the start of the planned opening of the midstream dam to carry out planned flood discharge to reduce the pressure in the water storage area. When the flash flood trigger function Fh(x, y, t) at the location (x, y) at time t is ≥ 0.8, it indicates that there is a flash flood risk. At this time, the current location (x, y) is marked as a level 3 risk level. At this time, the early warning module of the hydrodynamic model issues a level 2 warning message, prompting the immediate opening of the floodgates to release the floodwaters and pushing an evacuation message to the current area. Among them, the first-level risk level is less than the second-level risk level and less than the third-level risk level, and the third-level risk level is the highest risk level; At the same time, the second-level warning information is the highest level warning.

[0033] In this embodiment, the method constructs a flash flood triggering function Fh based on the comprehensive inflow function Qin obtained in the previous step, combined with the corresponding grid area's slope area S, landslide index X, and current inflow growth rate. This function measures the flash flood triggering potential of a region under specific terrain and soil conditions. This function not only considers water catchment capacity and inflow intensity, but also incorporates landslide sensitivity factors derived from geological surveys and a manually set risk amplification factor, achieving a "fixed-point, fixed-time, and quantitative" flash flood triggering intensity calculation based on multiple factors. Based on the output of the flash flood triggering function Fh, the triggering area is dynamically evaluated according to pre-set risk grading criteria. By implementing this method, the present invention implements a complete logical chain from water source input dynamics, slope response characteristics, and soil slip risk to risk level output and disaster response instructions, establishing a flash flood risk early warning mechanism with predictive, regionally adaptable, and dynamically controllable capabilities. Compared to traditional triggering methods based on static rainfall thresholds, this method possesses adaptive assessment capabilities for complex-driven flash flood events and multi-level linkage control capabilities, significantly improving the efficiency of flash flood response in mountainous areas, the timeliness of risk avoidance, and the level of system intelligence.

[0034] Example 7 See also Figure 1 and Figure 2 ,A flash flood simulation and early warning system based on a hydrodynamic model, including a detection point setting module, a hydrodynamic modeling module, an ice and snow kinetic energy analysis module, a comprehensive inflow analysis module and a flash flood early warning module; The detection point setting module arranges several detection points along the plateau river line, and installs a collection equipment group in each detection point to collect river ice and snow data in real time, and builds a cloud server to transmit the river ice and snow data to the cloud server; The hydrodynamic modeling module constructs a hydrodynamic model in a cloud server and simulates the hydrodynamic model based on river ice and snow data. After the simulation, the ice and snow data in the hydrodynamic model are extracted and preprocessed to obtain a standard ice and snow data set. The ice and snow kinetic energy analysis module calculates and outputs the ice and snow kinetic energy mutation judgment function TKE based on the standard ice and snow data set in the hydrodynamic model, sets the mutation threshold F1 and performs preliminary comparative evaluation with the ice and snow kinetic energy mutation judgment function TKE, and triggers flash flood analysis based on the preliminary comparative evaluation results; The comprehensive inflow analysis module calculates and outputs the initial water content W0 of the current plateau river source based on the standard ice and snow data set, and calculates and outputs the comprehensive inflow function Qin based on the initial water content W0 to analyze the soil moisture and comprehensive precipitation and snowmelt at the plateau river source; The flash flood warning module calculates and outputs a flash flood trigger function Fh based on the comprehensive inflow function Qin, and performs a secondary comparative evaluation based on the output result of the flash flood trigger function Fh to analyze the flash flood trigger risk level.

[0035] Specific examples: Assume that the density of ice and snow P = 1.0, Initial sliding speed vs(t0)=0.1, final sliding speed vs(t1)=0.9; The initial residual snow thickness hs(t0)=0.6, and the final residual snow thickness hs(t1)=0.5; Melting factor function under the influence of temperature and slope fmelt(T, Ts) = 1.2 Time interval Δt = 1 Initial kinetic energy change rate term: 0.1 2 *0.6=0.006, Final kinetic energy change rate term: 0.81*0.5=0.405, Difference 0.405-0.006=0.399; TKE(t)=1.0*0.399*1.2=0.4788; Extract the cumulative response of 30 spatial grid coordinates, 0.4788*30=14.36; Assuming that the mutation threshold F1 is set to 10, when the ice and snow kinetic energy mutation judgment function TKE(t) at time t is greater than the mutation threshold F1, the flash flood analysis is triggered. Assumptions: =0.8, cumulative precipitation: =0.5, SRF=0.4, △t=2.5, =0.3, =0.6, Sat max (x, y) = 1.0; e 0.75 =2.12 W0=0.8*[0.5+0.6*0.4*2.12]=0.807; Assumption: Current precipitation R=0.4, =0.8,SRF=0.4,fm(T)=1.2 Qin=0.4+(0.8*0.4*1.2) / 1+0.807=0.612 Assumptions: S=1.2, X=0.6, =1.0; Fh=1.0*(0.6122 / 1.2+0.212*0.6)=0.4393; 0.3≤0.4393<0.8, marked as the second-level risk level, and a first-level warning information is issued.

[0036] While the embodiments of the present invention have been shown and described, it will be apparent to those skilled in the art that various changes, modifications, substitutions, and alterations can be made to the embodiments without departing from the principles and spirit of the invention.

Claims

1. A flash flood simulation and early warning method based on a hydrodynamic model, characterized by: The following steps are involved: S1. Several detection points are set up along the plateau rivers, and a collection equipment group is installed at each detection point to collect river ice and snow data in real time. A cloud server is built to transmit the river ice and snow data to the cloud server; S2. In the cloud server, a hydrodynamic model is constructed and simulated based on river ice and snow data. After the simulation, ice and snow data in the hydrodynamic model are extracted and preprocessed to obtain a standard ice and snow dataset. S3. Based on the standard ice and snow data set in the hydrodynamic model, calculate and output the ice and snow kinetic energy mutation judgment function TKE, set the mutation threshold F1 and perform preliminary comparative evaluation with the ice and snow kinetic energy mutation judgment function TKE, and perform flash flood triggering analysis based on the preliminary comparative evaluation results; S4. The flash flood analysis is performed by calculating and outputting the initial water content W0 of the current plateau river source based on the standard ice and snow data set, and calculating and outputting the comprehensive inflow function Qin based on the initial water content W0, and analyzing the soil moisture and comprehensive precipitation and snowmelt at the plateau river source; S5. Calculate and output a flash flood triggering function Fh based on the comprehensive inflow function Qin, and perform a secondary comparative evaluation based on the output result of the flash flood triggering function Fh to analyze the flash flood triggering risk level.

2. The flash flood simulation and early warning method based on a hydrodynamic model according to claim 1, characterized in that: Said S1 includes S11 and S12; S11. Divide the source of the plateau rivers into three source regions, including an upstream region, a midstream region, and a downstream region; Different detection points are installed at intervals of 500 meters in each of the three source areas, including upstream detection points, midstream detection points, and downstream detection points. Different collection equipment groups are installed in the detection points to collect river ice and snow data of plateau rivers in real time; The upstream area is located in the source area of ​​the plateau mountainous area; The middle reaches of the river are located in the transition zone at the foot of the mountain; The downstream area is located below the foot of the mountain; The collection equipment installed in the upstream detection point and the midstream detection point includes LiDAR and environmental sensor groups The environmental sensor group includes a thermometer, a soil moisture meter, a rain gauge and a water level meter; The collection equipment installed in the downstream detection point includes an environmental sensor; The river ice and snow data include temperature T, ice and snow thickness Hd, soil moisture Sat, precipitation R and water level H; S12. Build a cloud server and set up a 5G wireless communication network to wirelessly communicate with the communication modules of all the collection devices in the collection device group. Connect each detection point through the Internet of Things technology loT, and upload the river ice and snow data collected by all detection points to the cloud server. At the same time, set up edge computing equipment to perform preliminary data filtering and noise suppression on the river ice and snow data.

3. The flash flood simulation and early warning method based on a hydrodynamic model according to claim 2, characterized in that: Said S2 includes S21 and S22; S21, using the three-dimensional model TELEMAC-3D as a framework of the hydrodynamic model in the cloud server to construct the hydrodynamic model; The hydrodynamic model includes an input module, a simulation module, a flash flood analysis module and an early warning module; The input module is used to combine the real-time river ice and snow data received with remote sensing technology and drone assistance to obtain terrain data and snow layer information of plateau rivers, and input them into the hydrodynamic model, and perform spatial interpolation, data fusion, and time interpolation and data completion in the hydrodynamic model; The terrain data includes terrain, slope, width and river channel information of the plateau river; The snow layer information includes snow density, snow thickness Hd, area and humidity; The spatial difference is interpolated by using the Kriging interpolation method on terrain data with a resolution lower than 90 meters, thereby increasing the terrain data to a resolution of 5 meters; The data fusion is performed by using a GIS platform to set consistent river spatial data for river ice and snow data, terrain data and snow layer information; The time difference and data completion are carried out by temporally interpolating river ice and snow data, terrain data and snow layer information, optimizing the data continuity in the simulation process, and filling in the missing data by interpolation method; The simulation module uses the river spatial data integrated based on the GIS platform to load river ice and snow data, terrain data and snow layer information, and performs simulation based on the hydrological movement SCS-CN model and hydrodynamic simulation software HEC-RAS to simulate the dynamic response of water flow and snow melting in plateau rivers. At the same time, the point information of the detection point is loaded into the river spatial data, and the spatial grid coordinates are embedded in the river space. The spatial grid coordinates are composed of longitude and latitude, and then the longitude and latitude position (x, y) coordinates of the detection point are displayed on the river spatial data, where x represents the spatial position in the longitude direction and y represents the spatial position in the latitude direction. The water flow dynamic response is achieved by simulating how the water flow in the plateau river channel changes with time, temperature T and precipitation R. The snow melt simulation simulates snow melting based on temperature T and snow layer information, and uses temperature as a driving force. The amount of snowmelt water at each moment in the snowmelt process is simulated based on the snow layer information, and the snowmelt water volume is summed with the rainfall R as the water source input simulation for the plateau rivers. S22. Based on the simulation module of the hydrodynamic model, the simulation of plateau rivers and snowmelt is carried out. During the simulation process, the ice and snow data of different regions are collected in real time, and the ice and snow data are pre-processed to obtain a standard ice and snow data set; The preprocessing includes normalization and timestamp alignment; The normalization process is performed by using the z-score standard processing method to eliminate the dimensional effects of all parameters in the ice and snow data; The timestamp alignment is performed by aligning the data extraction time with the real-time data of the real time; The standard ice and snow dataset includes the ice and snow density P at time t, the residual snow thickness hs(t) at time t, the slope function Ts at time t, the snow melt rate SRF(t) at time t, the precipitation R(t) at time t, the water level H(t) at time t, and the slope area S.

4. The flash flood simulation and early warning method based on a hydrodynamic model according to claim 3, characterized in that: Said S3 includes S31 and S32; S31, the flash flood analysis module includes snowmelt mutation analysis and flash flood analysis; The snowmelt mutation analysis constructs an ice and snow kinetic energy mutation algorithm model in the flash flood analysis module, extracts a standard ice and snow data set and inputs it into the ice and snow kinetic energy mutation algorithm model, calculates and outputs the ice and snow kinetic energy mutation judgment function TKE, and analyzes the dynamic change of the ice and snow water tank; The ice and snow kinetic energy mutation judgment function TKE is calculated and output by the following ice and snow kinetic energy mutation algorithm model; ； Where TKE(t) represents the judgment function of the sudden change of ice and snow kinetic energy at time t, t0 represents the initial time, t1 represents the terminal time, vs(t) represents the surface sliding velocity at time t, and is dimensionless. fmelt(T, Ts) represents the melting factor function under the influence of temperature and slope surface, d represents the calculus variable, and dt represents the time calculus variable.

5. The flash flood simulation and early warning method based on a hydrodynamic model according to claim 4, characterized in that: S32. The kinetic energy threshold value of the current kinetic energy generated by the melting of ice and snow in plateau rivers, which affects flash floods, is set as a mutation threshold F1. A preliminary comparative evaluation is performed between the mutation threshold F1 and the real-time obtained ice and snow kinetic energy mutation judgment function TKE to analyze the situation of ice and snow kinetic energy mutations. Based on the preliminary comparative evaluation results, a flash flood triggering analysis is performed. The specific evaluation content is as follows; When the ice and snow kinetic energy mutation judgment function TKE (t) at time t is greater than the mutation threshold F1, it indicates that the ice and snow melting kinetic energy mutation is abnormal, and flash flood analysis is triggered at this time; When the ice and snow kinetic energy mutation judgment function TKE (t) at time t ≤ the mutation threshold F1, it means that the ice and snow melting kinetic energy mutation is normal and continuous monitoring is required.

6. The flash flood simulation and early warning method based on a hydrodynamic model according to claim 5, characterized in that: Said S4 includes S41 and S42; S41. After triggering flash flood analysis based on preliminary comparative assessment, extract the snow melt rate SRF(t) and precipitation R(t) at time t from the standard ice and snow dataset, calculate and output the initial soil moisture content W0 in the plateau river area, and analyze the current initial moisture state of the plateau river area; The initial water content W0 is calculated and outputted by the following algorithm formula; ； Where t-t0 represents the relative time difference, W0(x, y, t) represents the initial water content at the position (x, y) at time t, e represents the exponential function, and Sat max (x, y) represents the maximum soil moisture at position (x, y), represents the regulating factor of snowmelt on water content, It represents the coefficient of influence of temperature change on the reaction rate of water content, △t represents the time interval, dt represents the time calculus variable, Represents the slope influence function at position (x, y).

7. The flash flood simulation and early warning method based on a hydrodynamic model according to claim 6, characterized in that: S42. Based on the obtained initial water content W0, combined with the precipitation R and snow melt rate SRF at different locations in the standard ice and snow dataset, the comprehensive inflow function Qin is calculated and output to analyze the impact of precipitation and snow melt on flash floods under the dual driving conditions of precipitation and snow melt. The comprehensive inflow function Qin is calculated and outputted by the following algorithm formula: ； Where Qin(x, y, t) represents the integrated inflow function at the location (x, y) at time t, R(x, y, t) represents the precipitation at the location (x, y) at time t, and fm(T) represents the temperature-driven melt function. represents the precipitation and snowmelt inflow coefficient.

8. The flash flood simulation and early warning method based on a hydrodynamic model according to claim 6, characterized in that: Said S5 includes S51 and S52; S51, based on the real-time acquired comprehensive inflow function Qin and the slope area S, a comprehensive calculation is performed to output a flash flood triggering function Fh to measure the flash flood triggering situation; The flash flood trigger function Fh is calculated and output by the following algorithm formula: ； Where Fh(x, y, t) represents the flash flood triggering function at position (x, y) at time t, S(x, y) represents the slope area at position (x, y), and X(x, y) represents the landslide index. Represents the risk amplification factor.

9. The flash flood simulation and early warning method based on a hydrodynamic model according to claim 8, characterized in that: S52: Perform a secondary comparative evaluation based on the output result of the flash flood trigger function Fh, and generate a flash flood trigger risk level based on the secondary comparative evaluation result. At the same time, generate relevant warnings based on the flash flood trigger risk level. The specific evaluation content is as follows; When the flash flood trigger function Fh(x, y, t) at the position (x, y) at time t is less than 0.3, it indicates that the kinetic energy of ice and snow is abnormal but there is no flash flood risk. The current position (x, y) is marked as the first-level risk level, and the detection point acquisition frequency is increased by 50%; When 0.3≤the flash flood trigger function Fh(x, y, t) at position (x, y) at time t<0.8, it indicates a critical flash flood risk and the current position (x, y) is marked as a level 2 risk. At this time, the early warning module of the hydrodynamic model issues a level 1 warning message, prompting the start of the planned opening of the midstream dam. When the flash flood trigger function Fh(x, y, t) at the location (x, y) at time t is ≥ 0.8, it indicates that there is a flash flood risk. At this time, the current location (x, y) is marked as a level 3 risk level. At this time, the early warning module of the hydrodynamic model issues a level 2 warning message, prompting the immediate opening of the floodgates to release the floodwaters and pushing an evacuation message to the current area. Among them, the first-level risk level is less than the second-level risk level and less than the third-level risk level, and the third-level risk level is the highest risk level; At the same time, the second-level warning information is the highest level warning.

10. A flash flood simulation warning system based on a hydrodynamic model, applied to a flash flood simulation warning method based on a hydrodynamic model according to any one of claims 1 to 9, characterized in that: It includes detection point setting module, hydrodynamic modeling module, ice and snow kinetic energy analysis module, comprehensive inflow analysis module and flash flood warning module; The detection point setting module arranges several detection points along the plateau river line, and installs a collection equipment group in each detection point to collect river ice and snow data in real time, and builds a cloud server to transmit the river ice and snow data to the cloud server; The hydrodynamic modeling module constructs a hydrodynamic model in a cloud server and simulates the hydrodynamic model based on river ice and snow data. After the simulation, the ice and snow data in the hydrodynamic model are extracted and preprocessed to obtain a standard ice and snow data set. The ice and snow kinetic energy analysis module calculates and outputs the ice and snow kinetic energy mutation judgment function TKE based on the standard ice and snow data set in the hydrodynamic model, sets the mutation threshold F1 and performs preliminary comparative evaluation with the ice and snow kinetic energy mutation judgment function TKE, and triggers flash flood analysis based on the preliminary comparative evaluation results; The comprehensive inflow analysis module calculates and outputs the initial water content W0 of the current plateau river source based on the standard ice and snow data set, and calculates and outputs the comprehensive inflow function Qin based on the initial water content W0 to analyze the soil moisture and comprehensive precipitation and snowmelt at the plateau river source; The flash flood warning module calculates and outputs a flash flood trigger function Fh based on the comprehensive inflow function Qin, and performs a secondary comparative evaluation based on the output result of the flash flood trigger function Fh to analyze the flash flood trigger risk level.

Images