A breach loss evaluation method dynamically quantifying breach width

Abstract

A breach loss evaluation method quantifies breach width dynamically, and belongs to the field of dike breach disaster loss evaluation. The dike breach and flood evolution model is used to simulate the flood evolution process of the dike breach, so as to obtain the characteristics of the dike breach flood. A dike breach flood disaster life-economy loss comprehensive evaluation model based on a dynamic Bayesian network is constructed, the downstream disaster loss situation caused by the dike breach flood disaster is quantified, and a breach width and dike breach loss quantitative representation model is linearly fitted through multi-scenario and multi-scale dike breach working condition simulation. The dike breach evolution law is disclosed, the dike breach flood disaster loss situation can be quickly and effectively evaluated, and the systematicness of the dike breach flood disaster loss evaluation method and the timeliness and scientificity of decision making are improved.

Description

Technical Field

[0001] This invention belongs to the field of water conservancy engineering technology and relates to a method for assessing the loss of dikes in case of breach. Background Technology

[0002] Dike engineering is a crucial safeguard for the lives and property of people in my country's lake areas. Although my country's flood control system based on dike engineering is becoming increasingly sophisticated, the properties of dike materials remain highly uncertain due to the long service life of most dikes and limitations imposed by construction techniques and historical conditions at the time. Under high flood levels and prolonged immersion and erosion during the flood season, dikes are prone to problems such as piping, seepage, landslides, and erosion. Dike breaches, as typical low-probability, high-risk, and high-loss social disasters, pose a serious threat to the lives and property of the people. Reasonable and reliable risk assessment methods are essential for risk assessment personnel to formulate emergency decisions and responses to dike breach floods.

[0003] Current research generally neglects the physical mechanisms of water-sand coupling and layer-by-layer erosion during levee breaches, failing to reveal the fundamental laws governing breach evolution and making it difficult to accurately capture the characteristics of levee breach floods. Furthermore, levee breach flood loss assessment is influenced by numerous and complex factors. Besides flood characteristics such as inundation depth and flow velocity, the subjective evacuation behavior of people during levee breaches is also a significant factor affecting loss assessment. The uncertainty of these multiple factors leads to low accuracy in levee breach flood loss assessment. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a method for assessing levee breach losses by dynamically quantifying breach width. This method constructs a water-soil coupled flood-breach evolution model to obtain the characteristics of breach morphology evolution and flood-breach evolution. Based on this, a Bayesian network is used to dynamically assess flood-related loss of life and economic losses during the flood evolution process. Finally, a functional relationship between breach width and total levee breach loss is established, realizing full-chain technical support from flood evolution to loss control. This helps risk personnel to respond quickly and efficiently to emergencies and provides technical support for their flood risk assessment and post-disaster recovery work.

[0005] The present invention provides a method for assessing levee breach loss by dynamically quantifying breach width, comprising the following steps:

[0006] Step S1: Intelligent Data Collection

[0007] (1) System collects and organizes multi-source monitoring data: Using satellite, UAV and unmanned boat and other integrated air-space-ground monitoring methods, meteorological data, GIS data, hydrological data and other multi-source monitoring data are collected in an all-round way. The system organizes the geometric parameters of the dike body, the process line of water level change in front of the dike, local remote sensing impact data, digital elevation data, population distribution and economic development status required for model calculation.

[0008] (2) Intelligent query and acquisition of specific data: By inputting a specific year and region, the data acquisition system can be filtered to obtain the raw data that meets the requirements under specific working conditions.

[0009] (3) Data preprocessing: Check the various types of raw data obtained, identify missing and duplicate values, and calibrate the geometric and physical parameters of the embankment. At the same time, perform format conversion and processing on the GIS data to ensure the accuracy of the data.

[0010] Step S2: Construct a water-soil coupled model of the entire flood evolution process of levee breach.

[0011] (1) Constructing a water-soil coupled embankment breach evolution model: Based on the breach evolution rate, the embankment breach process is divided into two stages: the rapid breach stage and the slow evolution stage. The following formulas are used to calculate the process line of breach flow change and the pattern of breach morphology change during the embankment breach process.

[0012]

[0013]

[0014] Where Q is the breach flow rate, m 3 / s;k sm is the tailrace flooding coefficient; constants c1 and c2 are 1.7 and 1.3 respectively; m is the breach slope ratio (horizontal height / vertical height); b is the breach bottom width, m; H is the water depth at the breach location, m; z s The reservoir water level is measured in meters (m) and z. t The tailwater level is m; g is the acceleration due to gravity, m / s². 2 n is the Manning roughness coefficient; L is the breach length along the flow direction, in meters; R is the hydraulic radius, in meters; λ en and λ ex A is the head loss at the inlet and outlet of the breach, in meters; A is the cross-sectional area of ​​the breach, in square meters. 2 B is the width of the breached water surface, in meters; ρ is the water density, in kilograms per cubic meter of water. 3 ;ρ a air density, kg / m³ 3 C d U is the drag coefficient; win Wind speed, m / s; θ win θ is the angle between the breach centerline and the wind direction (°); |Q| is the absolute value of the breach flow rate (m). 3 / s.

[0015] (2) Constructing a model for levee breach and flood evolution: The breach flow and breach width are calculated using the levee breach evolution model. Based on geographic elevation data and combined with local land use data, a two-dimensional hydrodynamic model for levee breach is constructed using the Navier-Stokes equation to obtain the characteristics of the breach flood (including inundation depth, flood velocity, etc.).

[0016] Step S3: Construct a module for assessing life and economic losses from levee breach floods.

[0017] (1) Data preprocessing required for life loss assessment: Based on the calculation results of the levee breach flood evolution model, extract the characteristics of the levee breach flood evolution (inundation depth, flood velocity, etc.). Combined with population and building spatial distribution data, the study area is divided into multiple areas with different geographical locations according to the distance of buildings from the dam site and the distance required for residents to evacuate. At the same time, the economic development status of the study area is collected, and the total pre-disaster economy of the study area is calculated.

[0018] (2) Constructing a module for assessing life loss in levee breach floods: Based on a large amount of historical data, it is assumed that the mortality rate of people in safe areas is 0, while the mortality rate of people in high-risk areas is 0.91. For people in low- and medium-risk areas, a log-normal distribution is used to construct a function of mortality rate and water depth, as shown in the following formula:

[0019]

[0020] Among them, F D (h) is a function of mortality rate and flood depth h; It is the standard normal distribution function; and These are the mean and standard deviation, respectively, representing the low-risk area. , medium-risk areas , .

[0021] (3) Constructing a module for assessing economic losses from levee breaches: To differentiate between various economic sectors, multiple economic loss calculation models are used to assess the economic losses from levee breaches. The length-ratio model is used for infrastructure such as pipelines, roads, and bridges, while the economic loss caused by the shutdown of industrial and commercial production is assessed using the economic activity interruption time model.

[0022] The flood losses for other economic sectors are calculated using a constructed function of water depth and loss rate, as shown in the following formula:

[0023]

[0024] Based on this, the modified loss rate method is applied to different economic industries to obtain the final loss rate calculation formula as shown in the figure:

[0025]

[0026] Among them, R Loe (i,h) represents the economic loss rate of the i-th type of economy at a water depth of h, where the loss rate is R when the water depth h > 3.5 m. Loe (i, 3.5); Gr ij R represents the proportion of the j-th economic type within the i-th economic type. Loe (i,j,h) represents the economic loss rate of the j-th economic type at water depth h; f v f is the flow rate correction factor; T R' is the correction factor for the warning time. Loe (i,h) represents the final economic loss rate of the i-th type of economy at a water depth of h.

[0027] Step S4: Construct a dynamic Bayesian dam failure loss assessment model

[0028] (1) Constructing a Bayesian network: Integrating the economic loss assessment module and the life loss assessment module, constructing a Bayesian network for assessing flood damage caused by levee breaches, and using Bayes' formula and the total probability formula to perform calculation and reasoning on the Bayesian network.

[0029] (2) Improvement of Bayesian network time series: Add "time nodes" to the Bayesian network to divide the flood evolution process into several continuous time segments. Based on the Markov assumption and the assumption of constant transition probability, establish a probability transition matrix to obtain the probability distribution of flood depth and flood velocity in each time segment during the flood evolution process, and finally realize the dynamic assessment of flood loss due to dike breach.

[0030] Step S5: Construct a quantitative characterization model of breach width and levee failure loss:

[0031] (1) A flood evolution model of the entire process of levee breach was used to simulate levee breach conditions in multiple scenarios and at multiple scales. The flood evolution was collected under different initial breach widths and different levee soil parameters (such as different cohesion and internal friction angles). The economic loss was calculated by a dynamic Bayesian levee breach loss assessment model. For loss of life, the human resource method was used to estimate the value of loss of life. The time-varying curves of breach width and the corresponding total economic loss under each scenario were obtained simultaneously.

[0032] (2) The Levenberg-Marquardt algorithm was used for nonlinear regression fitting, and a piecewise nonlinear function was constructed as a quantitative characterization model of the breach width and the total loss of the embankment. The difference in the loss variation law between different breach width ranges was taken into account. The function fit degree R²=0.95, which meets the accuracy requirements of engineering applications.

[0033]

[0034] Where L represents economic loss, B represents breach width, a1 and a2 are loss correction coefficients, and b1, b2, and b3 are exponential coefficients.

[0035] The beneficial effects of this invention are as follows:

[0036] (1) The levee breach flood evolution model constructed in this invention not only scientifically characterizes the nonlinear evolution process of levee breaches caused by water erosion, such as lateral widening and vertical downward shearing, but also accurately quantifies the changes in the breach flow process curve and breach profile, thus obtaining a more accurate flood evolution process. Compared with the method of deriving the flow curve of the entire levee breach process by fitting measured flow data, the levee breach numerical simulation method proposed in this invention can still simulate and predict the evolution process of levee breach floods relatively accurately even in the absence of measured data. For flood loss assessment, the constructed dynamic Bayesian levee breach loss assessment model not only considers the logical relationship between various factors and their uncertainties, but also considers the risk evacuation behavior of at-risk populations during the flood evolution process. It combines the spatial distribution of population and buildings to ensure the accuracy of assessment data and comprehensively assesses the loss of life and economic losses in levee breach floods. Finally, the functional relationship between breach width and total levee breach loss is constructed, which helps at-risk personnel to respond quickly and efficiently to emergencies.

[0037] (2) This invention helps risk personnel to respond quickly and efficiently to floods caused by levee breaches, and provides technical support for their flood risk assessment and post-disaster recovery work. Attached Figure Description

[0038] Figure 1 This is a flowchart of the present invention.

[0039] Figure 2 This is a diagram showing the change in flow rate at the breach.

[0040] Figure 3 The diagram shows the process of change in the geometric morphology of the ulcer. Among them, (a) is the process line of the lateral expansion of the ulcer, and (b) is the process line of the longitudinal expansion of the ulcer.

[0041] Figure 4 This is a diagram illustrating the evolution of a flood caused by a levee breach.

[0042] Figure 5 This is a map showing the division of the Sanjiaolianwei area.

[0043] Figure 6 This is a dynamic assessment chart of the losses caused by the levee breach and flooding. Detailed Implementation

[0044] To more clearly illustrate the purpose, technical solution, and advantages of this invention, the following will provide an in-depth analysis of the invention in conjunction with detailed accompanying drawings and specific embodiments. It should be clarified that these specific embodiments are merely auxiliary means for understanding the connotation of this invention, and not intended to limit its scope or boundaries.

[0045] Example

[0046] This example focuses on the 2020 breach of the Sanjiaolianwei dike in Jiangxi Province, located at the border of Jiujiang City and Nanchang City.

[0047] Step S1: Intelligent Data Collection

[0048] Data obtained from automated drone patrols shows that the dike is 33.57 km long, with an elevation of 22.50 m, a width of 6.00 m, and a slope ratio of 1:3. The protected area of ​​the dike is 56.28 km². 2 The project covers and protects 50,300 mu of farmland and a population of approximately 23,400. Geometric and material parameters of the dike were collected through monitoring equipment and on-site surveys. Elevation data of the triangular embankment, local imagery, and real-time water level data from unmanned aerial vehicles were also accessed through an established data platform.

[0049] Step S2: Construct a water-soil coupled model of the entire flood evolution process of levee breach.

[0050] The levee breach evolution model uses the real-time water levels inside and outside the levee as the main input parameters, and the remaining model parameters are shown in Table 1. The Sanjiao Lianwei began to breach at 19:30, at which time the water level outside the levee was 21.02 m and the levee foundation elevation was 17.00 m. The levee breach evolution model was then established.

[0051] Table 1. Parameter values ​​for the dike breach evolution model.

[0052]

[0053] A flood evolution model for a breached levee was developed, with the entire levee area covered by the triangular levee. The flood inflow boundary was defined by the levee's diversion point, and all other boundaries were closed. The elevation coordinate system was WGS84, and the digital elevation model had a resolution of 8 m. Irregular triangular meshes from the MIKE 21 model were selected, with control mesh side lengths ranging from 30 to 50 m, resulting in a total of 81,534 mesh elements. The values ​​of physical parameters involved in the model, such as computation time, boundary conditions, and eddy viscosity coefficient, are shown in Table 2.

[0054] Table 2 MIKE 21 Model Parameter Values

[0055]

[0056] By constructing a model of levee breach and flood evolution, the breach flow hydrograph, breach width expansion hydrograph, and breach bottom elevation change curve were obtained. Figure 2 , Figure 3 As shown, the calculated peak flow rate of the breach is 1282 m³ / s. 3 / s, the peak time was calculated to be 12.9 h after the breach, and the maximum breach width was calculated to be 177.0 m. The inundation depth distribution at some typical moments is shown below. Figure 4 As shown, 100 hours after the breach, the polder area was almost completely covered by floodwaters, with the maximum flooded area reaching 55.92 km². 2 In particular, the northern part of the polder area, due to its lower elevation, was flooded to a depth of more than 4.5 meters, while the southern part, with its higher elevation, was flooded to a depth of less than 4.5 meters in most areas. The entire polder area was severely affected by the breach flood.

[0057] Step S3: Construct a module for assessing life and economic losses from levee breach floods.

[0058] Based on the characteristics of the breach flood (inundation depth, flood velocity, etc.) obtained in step S2, the main factors affecting the assessment of breach flood losses are collected and organized. The formulas for calculating life loss and quantifying economic losses of various economic industries are adopted. The conditional probabilities of the factors affecting life and economic losses of breach floods are calculated using the Monte Carlo simulation method in Python. The conditional probability distribution tables of life loss and economic loss modules under different conditions are obtained.

[0059] Step S4: Construct a dynamic Bayesian dam failure loss assessment model

[0060] A visualized Bayesian network was built using Bayesian network inference software (GeNIe). The conditional probability distribution tables of the life loss and economic loss modules under different conditions, obtained in step S3, were imported to quantify the probability distribution of the Bayesian network nodes. The Bayesian network was then improved by adding "time nodes" to divide the flood evolution process into several continuous time segments. Based on the Markov assumption and the assumption of invariant transition probabilities, a probability transition matrix was established to obtain the probability distribution of inundation depth and flood velocity in each time segment during the flood evolution process.

[0061] Then, the distribution of at-risk populations and the pre-disaster economic distribution of various industries in the study area were collected and organized. Given the large area of ​​the study region, to accurately delineate the states of parent nodes in the Bayesian network, the triangular embankment was divided into five areas with different geographical locations: R1, R2, R3, R4, and R5, based on the distance of buildings from the dam site and the distance required for resident evacuation. Figure 5 As shown.

[0062] To achieve a reasonable division and quantification of the at-risk population, this invention, based on the spatial distribution characteristics of buildings in remote sensing images, assumes a linear correlation between the population of each area and the proportion of the building area in that area to the total building area of ​​the triangular dike. The 23,411 permanent residents of the dike area are then distributed across each area, completing a preliminary spatial distribution of the population. Based on this, to further clarify the number of at-risk individuals directly threatened by floods, the number of submerged building pixels is counted, combined with the flood inundation range data obtained from the previous simulation, thus determining the number of at-risk individuals. The specific distribution of the at-risk population in each area is shown in Table 3.

[0063] Table 3. Distribution of at-risk population

[0064]

[0065] Based on the actual conditions of the Sanjiaolianwei area, and considering the damage to infrastructure such as roads and pipelines, the repair cost M for damaged roads is set at 1.2 million yuan / km. For losses incurred by industry and commerce due to production stoppages or reductions caused by the flood, referring to the statistical data in the "Yongxiu County 2020 National Economic and Social Development Statistical Bulletin," the hourly net industrial and commercial output value D of Sanjiaolianwei is determined to be 966,000 yuan / h. For other economic sectors, the value of agriculture and fisheries is determined to be 14.95 yuan / m². 2 The value of forestry and animal husbandry is 2.80 yuan / m². 2 The value of the tertiary industry is 275.69 yuan / m². 2 The value of residents' property is 237.43 yuan / m². 2 Based on the land use data of the triangular embankment, the pre-disaster economic distribution of the area can be further estimated. Taking the R5 area as an example, the model calculates the changes at each node during the flood's evolution, based on the flood evolution results. Figure 6 As shown in Tables 4 and 5, the final life casualties and economic losses are statistically analyzed.

[0066] Table 4. Assessment Results of Life Losses Due to the Collapse of the Triangle Joint Embankment

[0067]

[0068] Table 5. Economic Loss Assessment Results of the Triangular Joint Embankment Breach

[0069]

[0070] The life loss assessment results indicate that the vast majority of at-risk individuals were successfully evacuated to safe areas or moved indoors, effectively reducing the danger posed by the flood. Calculations based on different combinations of parent node states yielded life loss rates of 0.06%, 0.04%, 0.09%, 0.50%, and 1.02% for each area of ​​the triangular embankment. Combining this with the number of at-risk individuals in each area, the potential death toll was further calculated. The total potential death toll was 29, with an average life loss rate of 0.12% for the region. Area R4, with its large at-risk population and high flood risk, has a potential death toll of 13, thus requiring focused attention and intensified rescue and relief efforts.

[0071] The economic losses caused by the breach of the Sanjiao Lianwei dike amounted to approximately 582 million yuan, which is highly consistent with the actual economic losses of 600 million yuan. Of this, direct economic losses were approximately 404 million yuan, and indirect economic losses were approximately 178 million yuan. Furthermore, the R2 area, with its densely packed buildings and high land value, suffered the largest economic losses in this flood, totaling approximately 113 million yuan. Further analysis of the losses across various economic sectors reveals that agriculture and fisheries, being highly sensitive to floods, suffered the greatest losses. As the main industries in the Sanjiao Lianwei area, agriculture and fisheries accounted for a large proportion of the land area affected, resulting in the most severe economic losses for this sector, totaling approximately 207 million yuan. Economic losses due to the shutdown of industrial and commercial production caused by the flood ranked second, at approximately 180 million yuan.

[0072] Step S5: Construct a quantitative characterization model of breach width and levee failure loss

[0073] By employing the dynamic Bayesian levee breach loss method constructed in step S4, and considering multiple scenarios and scales of levee breach conditions, the flood loss under different conditions was calculated. According to the formula obtained from nonlinear regression fitting, the breach width is 177m, falling within the third segment of the function (B≥100 m). Substituting this into the third segment function, where L... max The value is 620 million yuan, L B2 With a value of 450 million yuan and b3 of 0.03, the total loss was calculated to be 603 million yuan. The results show that the quantitative characterization model of breach width and levee breach loss constructed in this invention can accurately quantify the actual economic loss of levee breaches under different watershed scales by adjusting the parameters.

[0074] In summary, the Bayesian network method can accurately assess the losses during the evolution of a levee breach flood, providing a scientific basis for emergency response and post-disaster compensation for levee breach disasters.

[0075] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent modifications made based on the content of the present invention specification and drawings, or direct or indirect applications in related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for assessing levee breach loss by dynamically quantifying breach width, characterized in that, Includes the following steps: Step S1: Intelligent Data Collection (1) System collects and organizes multi-source monitoring data: Using satellite, UAV and unmanned boat integrated air-ground monitoring method, meteorological data, GIS data and hydrological data are collected in all aspects. The system organizes the geometric parameters of the dike body, the process line of water level change in front of the dike, local remote sensing impact data, digital elevation data, population distribution and economic development status required for model calculation. (2) Intelligent query and acquisition of specific data: By inputting a specific year and region, the data acquisition system can be filtered to obtain the raw data that meets the requirements under specific working conditions; (3) Data preprocessing: Check the various types of raw data obtained, identify missing and duplicate values, calibrate the geometric and physical parameters of the embankment, and convert and process the GIS data to ensure the accuracy of the data. Step S2: Construct a water-soil coupled model of the entire flood evolution process of levee breach: (1) Constructing a water-soil coupled embankment breach evolution model: Based on the breach evolution rate, the embankment breach process is divided into two stages: a rapid breach stage and a slow evolution stage. The following formulas are used to calculate the process line of breach flow change and the pattern of breach morphology change during the embankment breach process: Where Q is the breach flow rate, m 3 / s;k sm denoted as the tailrace inundation coefficient; constants c1 and c2 are taken as 1.7 and 1.3 respectively; m is the breach slope ratio, horizontal height / vertical height; b is the breach bottom width, in meters; H is the water depth at the breach location, in meters; z s The reservoir water level is measured in meters (m) and z. t The tailwater level is m; g is the acceleration due to gravity, m / s². 2 n is the Manning roughness coefficient; L is the breach length along the flow direction, in meters; R is the hydraulic radius, in meters; λ en and λ ex A is the head loss at the inlet and outlet of the breach, in meters; A is the cross-sectional area of ​​the breach, in square meters. 2 B is the width of the breached water surface, in meters; ρ is the water density, in kilograms per cubic meter of water. 3 ;ρ a air density, kg / m³ 3 C d U is the drag coefficient; win Wind speed, m / s; θ win Let |Q| be the angle between the breach centerline and the wind direction; |Q| is the absolute value of the breach flow rate, m. 3 / s; (2) Constructing a model for levee breach and flood evolution: The levee breach evolution model is used to calculate the breach flow and breach width. Based on geographic elevation data and combined with local land use data, a two-dimensional hydrodynamic model for levee breach is constructed using the Navier-Stokes equation to obtain the characteristics of the breach flood, including the inundation depth and flood velocity information. Step S3: Construct a module for assessing life and economic losses from levee breach floods: (1) Data preprocessing required for life loss assessment: Based on the calculation results of the levee breach flood evolution model, extract the evolution characteristics of the levee breach flood, including the inundation depth and flood velocity information; Combine the population and building spatial distribution data, and divide the study area into multiple areas with different geographical locations based on the distance of buildings from the dam site and the distance required for residents to evacuate; At the same time, collect the economic development status of the study area and calculate the total pre-disaster economy of the study area. (2) Constructing a module for assessing life loss in levee breach floods: Based on a large amount of historical data, it is assumed that the mortality rate of people in safe areas is 0, while the mortality rate of people in high-risk areas is 0.

91. For people in low- and medium-risk areas, a log-normal distribution is used to construct a function of mortality rate and water depth, as shown in the following formula: Among them, F D (h) is a function of mortality rate and flood depth h; It is the standard normal distribution function; and These are the mean and standard deviation, respectively, representing the low-risk area. , medium-risk areas , ; (3) Construct a module for assessing economic losses from dike breaches: In order to differentiate between various economic industries, multiple economic loss calculation models are used to assess the economic losses from dike breaches; the length ratio model is used for pipelines, roads and bridges, and the economic loss caused by the shutdown of industrial and commercial production is used for the economic activity interruption time model. The flood losses for other economic sectors are calculated using a constructed function of water depth and loss rate, as shown in the following formula: Based on this, the adjusted loss rate method is applied to different economic sectors to obtain the final loss rate calculation formula: Among them, R Loe (i,h) represents the economic loss rate of the i-th type of economy at a water depth of h, where the loss rate is R when the water depth h > 3.5 m. Loe (i, 3.5); Gr ij R represents the proportion of the j-th economic type within the i-th economic type. Loe (i,j,h) represents the economic loss rate of the j-th economic type at water depth h; f v f is the flow rate correction factor; T R' is the correction factor for the warning time. Loe (i,h) represents the final economic loss rate of the i-th type of economy at a water depth of h; Step S4: Construct a dynamic Bayesian dam failure loss assessment model: (1) Constructing a Bayesian network: Integrating the economic loss assessment module and the life loss assessment module, constructing a Bayesian network for assessing flood damage caused by levee breaches, and using Bayes' formula and the total probability formula to perform calculation and reasoning on the Bayesian network; (2) Improvement of Bayesian network time series: Add time nodes to the Bayesian network to divide the flood evolution process into several continuous time segments. Based on the Markov assumption and the assumption of constant transition probability, establish a probability transition matrix to obtain the probability distribution of flood depth and flood velocity in each time segment during the flood evolution process, and finally realize the dynamic assessment of flood loss due to dike breach. Step S5: Construct a quantitative characterization model of breach width and levee failure loss: (1) The flood evolution model of the entire process of levee breach was used to simulate the levee breach conditions in multiple scenarios and at multiple scales. The flood evolution was collected under different initial breach widths and different levee soil parameters, including different cohesion and internal friction angles. The economic loss was calculated by the dynamic Bayesian levee breach loss assessment model. For loss of life, the human resource method was used to estimate the value of loss of life. The time-varying curves of breach width and the corresponding total economic loss under each scenario were obtained simultaneously. (2) The Levenberg-Marquardt algorithm was used for nonlinear regression fitting to construct a piecewise nonlinear function as a quantitative characterization model of the breach width and the total loss of the levee. This model takes into account the differences in the loss variation patterns in different breach width ranges. The function fit R² = 0.95, which meets the accuracy requirements for engineering applications. Where L represents economic loss, B represents breach width, a1 and a2 are loss correction coefficients, and b1, b2, and b3 are exponential coefficients.

Images