Intelligent preplan method and system for urban flood control

Abstract

The application discloses a kind of wisdom preplan method and system for urban flood control, the method includes: obtaining the real-time monitoring data of target area flood control and the resource configuration data of current execution preplan;Reflecting the scene demand characteristics of disaster evolution demand based on monitoring data is constructed, and preplan supply characteristics reflecting emergency ability supply are extracted based on resource configuration data;The matching relationship of scene demand characteristics and preplan supply characteristics in target time window is calculated comprehensively, and the preplan adaptation state is obtained;In response to preplan adaptation state satisfying preset dynamic adjustment trigger condition, preplan adjustment instruction is generated.The application introduces target time window and supply cumulative model, and expands static evaluation into time evolution game, can carry out forward-looking trigger based on the adjustment lead time of predicted failure time, effectively solve the problem of traditional preplan response lag and resource scheduling timeliness deficiency.

Description

Technical Field

[0001] This invention belongs to the field of urban flood control plans, and in particular, a smart plan method and system for urban flood control. Background Technology

[0002] Urban flood control plans are a core technological support for responding to extreme weather disasters and ensuring the safe operation of cities. With the increasing complexity of urban infrastructure and the frequency of extreme weather events, building an intelligent plan system that can accurately quantify risks, scientifically allocate resources, and respond to changes in disaster situations in real time has significant technical research value for improving the accuracy of emergency response and the timeliness of resource allocation.

[0003] Current urban flood control plan development technology mainly relies on a static process of risk assessment, plan development, and drill optimization. The typical approach involves pre-compiling static text containing organizational command structures, emergency response grading standards, and resource allocation plans based on historical hydrological data and typical scenario assumptions. In practical applications, the system typically triggers corresponding level of plan activation or termination commands passively based on whether real-time monitored water levels or rainfall data reach preset warning thresholds.

[0004] However, existing technologies suffer from technical problems such as supply-demand mismatch and delayed response triggering when dealing with complex, dynamically evolving disasters. Specifically, existing methods mostly compare supply and demand based on static snapshots of the current moment, ignoring the time-varying relationship between the dynamic changes in demand during flood evolution and the gradual accumulation of emergency resource mobilization, resulting in an inability to accurately predict future supply-demand gaps. Furthermore, existing adjustment mechanisms often employ passive triggering logic, initiating adjustments only after monitoring indicators have exceeded thresholds, failing to consider the physical time lag inherent in resource scheduling itself. This leads to situations where, after adjustment instructions are issued, on-site resources often cannot be deployed within the effective time window before the contingency plan expires, causing a time lag in emergency response. Summary of the Invention

[0005] The purpose of this invention is to provide a smart contingency plan method and system for urban flood control, so as to solve the above-mentioned problems existing in the prior art.

[0006] Technical solutions for smart emergency response plans in urban flood control include:

[0007] Obtain real-time flood control monitoring data and resource allocation data for the current implementation plan for the target area;

[0008] Based on real-time flood control monitoring data, scenario demand characteristics reflecting the needs of disaster evolution are constructed, and contingency plan supply characteristics reflecting the supply of emergency response capabilities are extracted based on resource allocation data.

[0009] By comprehensively calculating the matching relationship between the demand characteristics of the calculation scenario and the supply characteristics of the contingency plan within a preset target time window, the contingency plan adaptation status is obtained.

[0010] In response to the pre-set dynamic adjustment triggering conditions being met in the contingency plan adaptation status, a contingency plan adjustment instruction is generated.

[0011] According to one aspect of this application, a smart contingency planning system for urban flood control includes:

[0012] The data acquisition module is used to acquire real-time flood control monitoring data and resource configuration data of the current implementation plan for the target area;

[0013] The feature construction module is used to construct scenario demand features that reflect the needs of disaster evolution based on real-time flood monitoring data, and to extract plan supply features that reflect the supply of emergency response capabilities based on resource allocation data.

[0014] The adaptation assessment module is used to comprehensively calculate the matching relationship between the scenario demand characteristics and the contingency plan supply characteristics within a preset target time window to obtain the contingency plan adaptation status.

[0015] The dynamic adjustment module is used to generate a plan adjustment instruction in response to the pre-set dynamic adjustment trigger conditions being met in the plan adaptation state.

[0016] Beneficial effects: By introducing a target time window and a supply accumulation model, this invention extends static evaluation into a time-series evolution game, enabling proactive triggering based on the adjustment lead time of the predicted failure time, effectively solving the problems of delayed response and insufficient timeliness of resource scheduling in traditional contingency plans. Attached Figure Description

[0017] Figure 1 A flowchart illustrating the steps of a smart emergency response plan method for urban flood control provided in this application embodiment.

[0018] Figure 2 A flowchart illustrating the steps for responding to a pre-defined dynamic adjustment triggering condition in a pre-defined adaptation state, as provided in this application embodiment.

[0019] Figure 3 A flowchart illustrating the steps for calculating one-dimensional fitness in an embodiment of this application.

[0020] Figure 4 A flowchart illustrating the steps for obtaining a plan adaptation state as provided in this application embodiment. Detailed Implementation

[0021] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0022] It should be noted that the terms include and have, and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to such process, method, product, or device.

[0023] like Figure 1 As shown, a smart emergency response plan method for urban flood control includes the following steps:

[0024] Obtain real-time flood control monitoring data for the target area and resource allocation data for the current implementation plan.

[0025] In this embodiment, a complete data perception foundation needs to be established. Real-time flood control monitoring data refers to a multi-source dynamic data set that reflects the current and short-term flood risk status of a city. Specifically, real-time flood control monitoring data can be obtained from urban meteorological departments, water conservancy and water resources departments, and municipal monitoring networks, and may include: real-time rainfall data, such as the cumulative rainfall over the past 1 hour, 3 hours, and 24 hours; river water level data, covering the comparison between real-time water levels and warning water levels at key sections of major flood-prone rivers; urban flooding point monitoring data, such as the water depth in low-lying areas like underpasses and underground passages; and reservoir dam operation data, such as current reservoir capacity and discharge flow. Furthermore, real-time flood control monitoring data can also include meteorological radar echo maps and short-term rainfall forecasts for the next 0 to 6 hours, providing a foundation for subsequent time-series evolution analysis. Preferably, the data collection cycle for real-time flood control monitoring data can be dynamically adjusted according to the disaster level, for example, collecting data every 30 minutes under normal conditions, every 10 minutes under warning conditions, and every 5 minutes under response conditions.

[0026] Meanwhile, resource allocation data refers to structured information about emergency resource deployment and dispatch capabilities extracted from the digital contingency plan database, from plans currently being implemented or prepared for implementation. Specifically, resource allocation data originates from a pre-built hierarchical and categorized intelligent contingency plan system. This system classifies urban flood control risks into three levels: high, medium, and low, and develops corresponding digital contingency plans for each level. Resource allocation data specifically includes: human resource allocation information, such as the location, number of personnel, and professional skill types of rescue teams; material and equipment allocation information, such as the inventory quantity and storage warehouse coordinates of flood control sandbags, inflatable boats, and large drainage pump trucks; and engineering facility dispatch rules, such as the start and stop thresholds for drainage pumping stations and the opening control strategies for sluice gates. In actual implementation, the static attributes and dynamic availability status of the above-mentioned resources can be retrieved in real time through a database query interface based on the currently activated contingency plan level ID, such as a Level II response plan.

[0027] Based on real-time flood control monitoring data, scenario demand characteristics reflecting the needs of disaster evolution are constructed, and contingency plan supply characteristics reflecting the supply of emergency response capabilities are extracted based on resource allocation data.

[0028] In this embodiment, the raw monitoring data and resource data are transformed into feature quantities that can be calculated by mathematical models. Scenario demand features are quantitative representations of the objective demand for various emergency measures required by the city to cope with floods under current or future disaster scenarios. Specifically, the construction process of scenario demand features involves mapping real-time flood control monitoring data to specific business demand dimensions. For example, based on real-time water level and topographic elevation data, the area affected by flooding is calculated using Geographic Information System (GIS) spatial analysis technology; based on the resident population density data within the area, the number of people requiring emergency evacuation is calculated, i.e., personnel evacuation demand; based on the difference between the predicted peak flood flow and the safe discharge capacity of the river channel, the volume of flood storage and detention areas or the drainage capacity of pumping stations that need to be dispatched is calculated, i.e., engineering dispatch demand. Scenario demand features can be static values ​​reflecting the current state, such as the material demand corresponding to the current water depth, or dynamic sequences that change over time, such as the predicted personnel evacuation demand every 30 minutes over the next 4 hours.

[0029] Correspondingly, the supply characteristics of a contingency plan are a quantitative representation of the emergency response capabilities that the current plan can provide. Specifically, the extraction process of the supply characteristics involves analyzing the availability and timeliness of emergency resources. For example, the total number of personnel in all rescue teams currently mobilized by the plan is used as a supply characteristic of personnel transfer capacity; the rated total flow of all available drainage equipment is summarized as a supply characteristic of engineering dispatch capacity. Similar to scenario demand characteristics, the supply characteristics of a contingency plan can be either static indicators reflecting the total amount of resources or dynamic supply capacity sequences that accumulate over time after considering factors such as resource mobilization lags and transportation time. Through characteristic processing, complex flood control operational problems are transformed into standard supply and demand matching mathematical problems.

[0030] By comprehensively calculating the matching relationship between the demand characteristics of the scenario and the supply characteristics of the contingency plan within the preset target time window, the contingency plan adaptation status is obtained.

[0031] Specifically, a target time window is introduced to unify the handling of both static and dynamic adaptation modes. The target time window refers to the time span T for supply and demand matching analysis. In some implementations, when only the emergency status at the current moment is considered, the target time window can be set to the current moment, i.e., T approaches 0. In this case, the contingency plan adaptation status is specifically reflected in the static comparison value between the supply characteristic value and the demand characteristic value at the current moment, such as the supply-demand ratio or the adaptation score at a single moment. In other preferred implementations, in order to achieve forward-looking early warning, the target time window can be set to a prediction window that includes the current moment to a preset future period, such as the next 4 hours. In this case, the system will perform dynamic matching calculations on a moment-by-moment basis between the time-series demand sequence and the time-series supply sequence within this window. The contingency plan adaptation status is specifically reflected in the time-series adaptation curve A(t) that evolves over time, which can intuitively reflect the profit and loss trend of the contingency plan's response capability over a future period.

[0032] In response to the pre-set dynamic adjustment triggering conditions being met in the contingency plan adaptation status, a contingency plan adjustment instruction is generated.

[0033] In this embodiment, the system automatically determines whether intervention is needed based on the calculated adaptation status. The dynamic adjustment trigger condition is a pre-defined logical rule used to determine whether the current contingency plan is invalid or about to become invalid. In static mode, the dynamic adjustment trigger condition can be an adaptation score below a preset threshold, such as 0.6; in dynamic mode, the dynamic adjustment trigger condition can be that the time-series adaptation curve falls below the validity threshold at some future time, or the adjustment lead time calculated based on the failure time is less than the safety buffer time. Once the trigger condition is met, the system generates a corresponding contingency plan adjustment instruction. The specific content of the contingency plan adjustment instruction depends on the degree of adaptation deviation and may include fine-tuning the existing contingency plan execution parameters, such as increasing the number of pump stations activated; supplementing emergency resources, such as urgently calling upon backup warehouse materials; or directly switching to a higher-level contingency plan, such as upgrading from a level-two response to a level-one response. Finally, the contingency plan adjustment instruction will be pushed to the emergency command platform to assist decision-makers in quickly adjusting action plans and achieving dynamic closed-loop optimization of the flood control contingency plan.

[0034] This embodiment achieves proactive dynamic closed-loop management from risk perception to contingency plan execution by constructing a dual-modal adaptation mechanism that combines static and dynamic approaches.

[0035] In one possible implementation, the scenario demand characteristics are quantified into a multi-dimensional demand vector that includes at least the dimensions of personnel transfer demand, engineering scheduling demand, and material support demand; the contingency plan supply characteristics are quantified into a multi-dimensional supply vector corresponding to the dimensions of the multi-dimensional demand vector.

[0036] In this embodiment, a multi-dimensional vector space is used to describe the supply and demand relationship in order to comprehensively evaluate the flood control emergency system. The multi-dimensional demand vector is denoted as R, with its components corresponding to different business dimensions. Specifically, the multi-dimensional demand vector can be represented as R = (R... per R eng R mat ); where R per The relocation needs of representatives are expressed in units of people; for example, based on the GIS analysis results of the current flooded area, the threatened population is calculated to be 1000 people. eng Represents engineering scheduling needs, and the unit can be flow rate (cubic meters per second) or volume (cubic meters), for example, the additional drainage capacity required to lower the water level is 500 cubic meters per second; R mat This represents the demand for material supplies, measured in pieces or tons. For example, 2000 sandbags are needed to reinforce a dike. Correspondingly, the multidimensional supply vector is denoted as P and expressed as P = (P... per P eng P mat Each component represents the upper limit of the corresponding capacity supply configured in the current contingency plan. For example, P perTo maximize the transfer capacity of the rescue teams mobilized under the contingency plan, P eng To activate the pumping station's total drainage capacity, P mat This refers to the reserve of pre-positioned materials.

[0037] In some extended implementations, the multidimensional demand vector can further include an information dissemination demand dimension and a response timeliness demand dimension, forming a five-dimensional vector structure. Among these, the information dissemination demand R... inf The response time requirement R can be calculated based on the population of the affected area, the coverage of communication base stations, and the number and types of early warning information that need to be issued; tim Calculations can be made based on the flood's evolution speed, the minimum time required for safe evacuation, and the permissible inundation duration for critical facilities. Correspondingly, the contingency plan's supply vector is also expanded into a five-dimensional structure, with each dimension's supply characteristics corresponding one-to-one with its demand characteristics.

[0038] In a further embodiment, the target time window is the current moment; the matching relationship between the scenario demand characteristics and the contingency plan supply characteristics within the preset target time window is comprehensively calculated to obtain the contingency plan adaptation status, including:

[0039] Calculate the ratio of the multidimensional supply vector to the multidimensional demand vector at the current moment, and determine the calculated ratio as the plan adaptation state.

[0040] In other words, the ratio of the multidimensional supply vector to the multidimensional demand vector at the current moment is calculated, and the calculated static comparison value is determined as the plan adaptation state.

[0041] Specifically, the system focuses on the current time point t. now State assessment. For each dimension k in the vector, calculate the supply-demand ratio for that dimension. k =P k / R k , where P k R is the supply vector of the k-th dimension; k Let R be the demand vector of the k-th dimension. The supply-demand ratio intuitively reflects the current sufficiency of resources. For example, if the demand for personnel transfer is R... per For 1000 people, and supply capacity P per If the number of people is 1200, the supply-demand ratio is 1.2, indicating that the capacity is sufficient and there is a slight surplus.

[0042] like Figure 3 As shown, in a preferred embodiment, the matching relationship between the comprehensive calculation scenario demand characteristics and the contingency plan supply characteristics within a preset target time window includes calculating the one-dimensional fit degree, specifically:

[0043] For each dimension in the multidimensional demand vector and the multidimensional supply vector, calculate the supply-demand ratio of the supply characteristic value and the demand characteristic value;

[0044] The supply-demand ratio is mapped to a one-dimensional fit degree using a pre-defined non-linear fit function.

[0045] The mapping rule of the nonlinear adaptation function is configured as follows: when the supply-demand ratio is within the preset supply-demand balance range, the single-dimensional adaptation degree takes the maximum value; when the supply-demand ratio exceeds the upper limit of the supply-demand balance range, the single-dimensional adaptation degree decreases as the supply-demand ratio increases, so as to reflect the scheduling redundancy risk caused by resource surplus.

[0046] In this embodiment, based on practical experience in flood control and disaster relief—that more resources are not necessarily better, and excessive resource accumulation can lead to chaotic on-site command, traffic congestion, and resource waste—a nonlinear evaluation logic with an oversaturation penalty mechanism is introduced. Specifically, for the k-th dimension, the nonlinear adaptation function φ(x) is used to adjust the supply-demand ratio x=Ratio. k Mapped to one-dimensional fitness A k The specific mathematical expression is as follows:

[0047] When 0 ≤ x < 1.0, φ(x) = x;

[0048] When 1.0 ≤ x ≤ 1.5, φ(x) = 1.0;

[0049] When x > 1.5, φ(x) = max(0.2, 1.0 - β*(x - 1.5));

[0050] Where x is the supply-demand ratio P k / R k β is the excess penalty coefficient, typically set between 0.2 and 0.4. The max function ensures that the fit is not lower than the preset lower limit of 0.2. When supply is less than demand (x < 1.0), the fit increases linearly with the increase in supply, reflecting the improvement in capacity. When supply is slightly greater than demand (1.0 ≤ x ≤ 1.5), it is in the golden range of supply and demand balance, and the fit reaches its maximum value of 1.0, indicating that both demand is met and reasonable safety redundancy is achieved. When supply is excessive (x > 1.5), the fit decreases linearly with the increase in supply, reflecting the penalty for over-allocation of resources. For example, if a region urgently needs 10 assault boats, but actually dispatches 30 (x = 3.0), with β set to 0.3, the fit will drop to 1.0 - 0.3 * (3.0 - 1.5) = 0.55, indicating that dispatchers should reduce resource deployment in that region to avoid congestion.

[0051] like Figure 4 As shown, in a further implementation, the pre-plan adaptation state is obtained, including:

[0052] Obtain the weight parameters corresponding to each dimension;

[0053] Based on the weight parameters, a weighted geometric average is calculated for the single-dimensional fit of each dimension to obtain the comprehensive fit. The comprehensive fit is then determined as the pre-plan fit state, so that the comprehensive fit is zero when the single-dimensional fit of any dimension is zero.

[0054] In this embodiment, to reflect the "weakest link" effect in emergency management—that is, the absence of any key dimension may lead to the failure of the overall rescue—a weighted geometric mean method is used to calculate the comprehensive score. Specifically, the weight parameter w for each dimension is determined according to the stage of disaster evolution. k In a preferred embodiment, the system presets a weight vector table for multiple stages: in the early warning stage, the information release dimension has a higher weight, set as W1=(0.15, 0.10, 0.15, 0.35, 0.25); in the response stage, the personnel transfer dimension has an increased weight, set as W2=(0.35, 0.25, 0.25, 0.10, 0.05). After the system identifies that it is currently in the response stage, it calls W2 as the weight parameter. The overall fit degree A is calculated using the following formula. total :

[0055] A total =∏(A k wk );

[0056] Where ∏ represents a multiplication operation, A k For the single-dimensional fit of the k-th dimension, w k Let be the weight of the k-th dimension, and satisfy Σw k =1. Compared to the traditional arithmetic mean, the geometric mean has a veto power. For example, if in a certain assessment, the personnel relocation fit is 0.9, the material support fit is 0.8, but the information dissemination fit is 0 due to the damage to the communication base station, then no matter how high the scores of other dimensions are, the overall fit A will be 0. total All of these will be directly reduced to 0. Managers of mandatory contingency plans must pay attention to all shortcomings to ensure a balanced match of capabilities across all dimensions.

[0057] This embodiment is applicable to scenarios in the early stages of disaster development or when data conditions are limited. It uses quantitative indicators to quickly evaluate the effectiveness of the contingency plan at the current moment.

[0058] In another possible implementation, the target time window is a prediction window that includes the current moment to a preset future period; the scenario demand characteristics are constructed as a time-series demand sequence that dynamically changes with the evolution of the flood.

[0059] Specifically, the system extends its perspective to the future. The prediction window length is set to T. windowFor example, from the current time t0 to t0+4 hours after the flood peak passes. Within this window, demand is no longer static, but a dynamic sequence that fluctuates with the evolution of the flood. The time-series demand sequence is denoted as R(t), where t takes the discrete time steps within the prediction window, for example, one point every 30 minutes.

[0060] In a preferred implementation, the scenario demand characteristics are constructed as a time-series demand sequence that dynamically changes with the evolution of the flood, specifically:

[0061] The distribution of flood inundation depth and flood velocity at different times within the prediction window is extrapolated using a hydrodynamic model based on real-time flood monitoring data.

[0062] In this embodiment, to obtain accurate physical parameters, the system invokes a hydrodynamic simulation engine. In some preferred embodiments, a two-dimensional hydrodynamic model based on the Saint-Venant equations can be used. A high-precision urban digital elevation model (DEM) and real-time rainfall and upstream water boundary conditions are employed to perform numerical integration calculations, outputting a gridded water depth field h(x, y, t) and a flow velocity field v(x, y, t). In other basic embodiments requiring higher computational speed, a simplified volume balance method can be used for simulations in relatively enclosed low-lying areas. The specific calculation formulas are as follows:

[0063] h i (t)=h i (t0)+∫(Q in_i (τ)-Q out_i (τ)) / Area i *dτ;

[0064] Where h i (t) represents the average flooding depth of the i-th region at time t; h i (t0) represents the measured water depth at the initial moment; Q in_i (τ) and Q out_i (τ) represent the inflow and outflow of the i-th region at time τ, respectively; Area i Let be the horizontal projected area of ​​the i-th region; the interval of the integral ∫ is from t0 to t; d represents the differential sign. The inundation depth h of each risk area at each time point within the prediction window can be obtained using any of the above methods. i (t) and flow velocity v i (t) data.

[0065] For the dimension of personnel relocation demand, a preset transfer coefficient function is used to process the distribution of flood inundation depth and flood flow velocity, calculate the number of people to be relocated at each time, and generate a time-series demand sequence for the dimension of personnel relocation demand; among which, the transfer coefficient function defines the relocation ratio corresponding to the risk level under different combinations of water depth and flow velocity.

[0066] In this embodiment, service requirements are calculated based on physical parameters. A transfer coefficient function η is introduced. i (h, v) is used to establish a quantitative mapping relationship between physical risks and personnel relocation. The transfer coefficient function classifies risks into levels based on the combination of water depth and current velocity, and specifies the corresponding relocation ratio. Its specific piecewise function expression is as follows:

[0067] When h i <0.3 and v i When η < 0.5, i =0;

[0068] When (0.3≤h) i <0.5) or (0.5≤v) i When η < 1.0), i =0.3;

[0069] When (0.5≤h) i <1.0) or (1.0≤v) i When η < 2.0), i =0.7;

[0070] When h i ≥1.0 or v i When ≥2.0, η i =1.0;

[0071] Where the submergence depth h i The unit is meters, and the flow velocity is v. i The unit is meters per second. When the water depth is very shallow and the current is very slow, the risk is manageable and no evacuation is necessary (coefficient is 0). As the water depth or current velocity increases, the risk level increases, and the proportion of the population requiring evacuation gradually increases to 30% and 70%. When the water depth exceeds 1 meter or the current velocity exceeds 2 meters per second, it poses a life-threatening threat to pedestrians, and everyone must be evacuated (coefficient is 1.0). Based on the evacuation coefficient, the evacuation demand R for the i-th region at time t is calculated. per_i (t):

[0072] R per_i (t)=Pop i *η i (h i (t), v i (t));

[0073] Pop i Let be the resident population of region i. Summing over all risk regions yields the total population transfer demand sequence R for the entire city at time t. per (t)=ΣR per_i (t).

[0074] For the engineering scheduling demand dimension, the flood peak flow prediction data is processed using a preset scheduling demand function to calculate the engineering scheduling demand at each time point and generate a time-series demand sequence for the engineering scheduling demand dimension.

[0075] In this embodiment, the engineering scheduling requirement reflects the number of flood storage and detention areas or pumping stations needed to be activated to absorb floods exceeding the standard. The specific calculation formula is as follows:

[0076] R eng (t) = max(0, (Q) peak (t) - Q safe ) * T dur * 3600 / V reg );

[0077] Where R eng (t) represents the engineering scheduling demand at time t, which is dimensionless and indicates the number of engineering scheduling units that need to be activated; Q peak (t) represents the peak flood discharge at time t predicted based on the hydrological model; Q safe The safe discharge capacity of the river channel is determined by the river channel design standards; T dur The duration of the flood; 3600 is the conversion factor between hours and seconds; V reg The storage capacity of the unit project scheduling unit; the max function ensures that the demand is zero when the peak flow is less than the safe discharge capacity.

[0078] When the predicted peak flood flow exceeds the safe discharge capacity of the river channel, the excess flood volume needs to be absorbed through engineering measures such as activating flood storage and detention areas or increasing drainage pumping stations. Multiplying the excess flow by the duration gives the total flood volume to be absorbed, and dividing by the unit engineering capacity gives the required number of engineering units.

[0079] For the dimension of material support demand, the total material demand at each time point is calculated using a preset material demand growth function, generating a time-series demand sequence for the dimension of material support demand.

[0080] Specifically, the demand for material support reflects the quantity of various emergency supplies needed for flood control and disaster relief, and takes into account the amplification effect of demand caused by time uncertainty. The specific calculation formula is as follows:

[0081] R mat (t) =Σ j=1 m (D j * (1 +α* (t - t start )));

[0082] Where R mat (t) represents the total material support demand at time t; D jThe baseline demand for category j materials can be calculated based on the disaster-stricken area and the affected population; α is the time uncertainty factor, ranging from 0.05 to 0.15, reflecting the characteristic that the longer the preset time, the greater the fluctuation in actual demand; t is the current calculation time; t start This represents the activation time of the contingency plan; m represents the total number of material categories, covering flood control sandbags, woven bags, life jackets, emergency lighting, etc.

[0083] The actual demand for emergency supplies gradually increases as the duration of a disaster lengthens. This is partly due to consumption and loss of supplies during use, and partly due to the uncertainty of initial estimates; as the situation on the ground becomes clearer, additional supplies are often needed. Therefore, a time-related increment is added to the baseline demand to ensure sufficient supply reserves.

[0084] The contingency plan supply characteristics are constructed as a time-series supply sequence that accumulates gradually with response time.

[0085] In a preferred implementation, the contingency supply characteristics are constructed as a time-series supply sequence that accumulates progressively with response time, including:

[0086] Obtain the theoretical total allocation of emergency resources and mobilization response parameters;

[0087] Using a pre-defined resource supply accumulation function, the cumulative amount of resources in place from the start of the plan to each moment within the prediction window is calculated based on the theoretical total allocation and mobilization response parameters, resulting in a time-series supply sequence that reflects the nonlinear growth process of resource supply capacity over time.

[0088] In this embodiment, operations research modeling is used to describe the dynamic ramp-up process of supply capacity. Unlike the static model which assumes instantaneous resource availability, this model fully considers the time lags of mobilization, assembly, and transportation. The time-series supply sequence is denoted as P(t). Different accumulation models are used for different types of resources. For continuous resources such as personnel and general supplies, the preferred exponential accumulation model is used. The exponential accumulation model assumes that the resource availability rate gradually slows down and tends to saturate over time. The specific calculation formula is as follows:

[0089] P per (t)=P per_max *(1-exp(-λ*(tt start )));

[0090] Where P per (t) represents the accumulated transfer capacity at time t; P per_max The total capacity limit configured for the contingency plan; t start Let be the activation time of the contingency plan; exp be the natural exponential function; and λ be the mobilization rate parameter. Furthermore, the mobilization rate parameter λ is related to the average scheduling distance D of the resources.avg The correlation satisfies the relationship λ=λ0*exp(-μ*D avg ), where λ0 is the base rate and μ is the distance attenuation coefficient. This indicates that the greater the distance, the slower the mobilization rate and the more delayed the resource delivery.

[0091] For discrete resources such as large-scale engineering machinery and drainage pump trucks, a step-cumulative model can be used. The step-cumulative model describes the process of resources arriving in batches and at specific times. The specific calculation formula is as follows:

[0092] P eng (t)=Σ(P eng_j *u(tt j ));

[0093] Where P eng_j Let t be the capacity value of the j-th scheduling unit, such as the j-th pump truck; j Let u(t) be the estimated arrival and start time of the j-th scheduling unit; u(t) is a unit step function, which takes the value 1 when t ≥ 0, and 0 otherwise; P eng (t) represents the total transfer capacity of discrete resources accumulated and available at time t. The system constructs a demand sequence R(t) reflecting the future evolution of the disaster and a supply sequence P(t) reflecting the dynamic availability of resources, laying a solid data foundation for subsequent time series matching.

[0094] Furthermore, by comprehensively calculating the matching relationship between the demand characteristics of the computing scenario and the supply characteristics of the contingency plan within a preset target time window, the contingency plan adaptation status is obtained, including:

[0095] At each time step within the prediction window, the dynamic matching degree between the time-series supply sequence and the time-series demand sequence is calculated, a time-series fit degree curve reflecting the evolution trend of the fit degree over time is generated, and the time-series fit degree curve is determined as the plan fit state.

[0096] In this embodiment, the system performs a convolutional point-by-point comparison between the time-series demand sequence R(t) and the time-series supply sequence P(t). Specifically, for each discrete time t within the prediction window, the system calls the nonlinear fitting function φ(x) and the geometric mean formula to calculate the instantaneous comprehensive fit A at that time. total (t). By connecting the calculation results of all time points in chronological order, a continuous time-series fit curve, denoted as A(t), can be generated. The time-series fit curve visually illustrates how the contingency plan's response capability changes as the flood evolves and resources gradually arrive. For example, the curve may show a pattern of first rising and then falling: initially, as resources arrive rapidly, the fit increases; later, as the flood inundation area expands beyond the resource coverage limit, the fit begins to decline. This evolutionary view of the entire process provides commanders with richer information than a single-moment snapshot.

[0097] According to one aspect of this application, the comprehensive calculation of the matching relationship between scenario demand characteristics and contingency plan supply characteristics within a preset target time window also includes marking the failure time, specifically:

[0098] Traverse the generated time series fit curves, identify the time point when the time series fit curve first falls below the preset validity threshold within the prediction window, and mark the identified time point as the failure time.

[0099] Specifically, risk thresholds are identified through mathematical analysis of the curve's shape. A preset effectiveness threshold A is used. threshold This is the baseline standard for judging whether a contingency plan is still qualified, usually set at 0.6 or 0.7. The basis for setting the effectiveness threshold is: when the overall adaptability is below 0.6, it means that the supply capacity of at least one key dimension is seriously insufficient or there are moderate gaps in multiple dimensions. At this time, the overall effectiveness of the contingency plan cannot be guaranteed. The specific value can be adjusted according to the risk preference of urban emergency management: risk-averse cities can raise the threshold to 0.75, and risk-neutral cities can set it at 0.65.

[0100] The system starts from the current time t0 and scans the curve A(t) along the time axis in the positive direction. When it detects that the fitness value A(t) at a certain time t is less than A... threshold When the fit value at the previous moment is greater than or equal to the validity threshold, the curve's downward crossing point is captured. This time point is marked as the failure time, denoted as T. fail If the curve remains above the threshold throughout the entire prediction window, the failure moment is marked as infinity, indicating that the contingency plan is safe within the current prediction range. Furthermore, the system can simultaneously calculate the maximum gap depth G. max , defined as 1.0 minus the lowest value of the fit curve within the prediction window, is used to help determine the severity of the failure.

[0101] like Figure 2 As shown, in a preferred embodiment, in response to the preset dynamic adjustment triggering condition being met in the plan adaptation state, the following includes:

[0102] Calculate the time difference between the failure time and the current time to obtain the adjustment lead; determine whether the adjustment lead is less than the preset safety buffer threshold; if so, determine that the dynamic adjustment trigger condition is met.

[0103] In this embodiment, abstract curve features are transformed into specific trigger signals. The system calculates the failure time T. fail With current system time t current The difference between them is the adjustment lead L. adjust This indicates how much time the command system has to make remedial and adjustment measures before the contingency plan completely fails. The specific calculation formula is as follows:

[0104] L adjust =T fail -t current ;

[0105] Where L adjust To adjust for lead time, the unit is hours; T fail t represents the predicted failure time. current This refers to the current moment.

[0106] The system will adjust the lead time and the preset safety buffer threshold T. buffer A comparison is made. The safety buffer threshold is set based on the minimum response period for resource scheduling, such as 2 hours or 4 hours. Specifically, the setting of the safety buffer threshold needs to consider the following factors: first, the average travel time of the rescue team from its base to the risk area; second, the loading, unloading, and transportation time of materials and equipment; and third, the mobilization and activation time of the new plan after the plan switch. For example, based on actual survey data from a certain city, the buffer threshold for a yellow alert is set at 6 hours, the buffer threshold for an orange alert at 4 hours, and the buffer threshold for a red alert at 2 hours. If L... adjust Less than T buffer This means that the time allotted for system adjustments is extremely limited, and a dynamic adjustment process must be initiated immediately. For example, assuming the current time is 10:00, and the forecast indicates that the fit will drop below 0.6 at 14:00, the lead time is 4 hours. If the set yellow warning buffer threshold is 6 hours, since 4 hours is less than 6 hours, the system will determine that the triggering condition is met and activate the corresponding warning level. By using a mechanism that triggers current actions based on future failures, the traditional passive response mode has been changed.

[0107] Optionally, the warning level can be determined based on the adjusted lead time and the maximum gap depth.

[0108] In this embodiment, the system comprehensively considers both time and severity dimensions for tiered early warning. The time dimension is represented by the adjustment lead time, reflecting how much time is available for adjustment; the severity dimension is represented by the maximum gap depth, reflecting the peak intensity of the supply-demand mismatch. The specific early warning level classification rules are as follows:

[0109] When the adjustment lead time is greater than or equal to 6 hours and the maximum gap depth is less than 0.2, it is judged as a green warning, indicating that the contingency plan is well adapted, there is sufficient time for adjustment, and no intervention is required;

[0110] When the adjustment lead time is between 4 and 6 hours, or the maximum gap depth is between 0.2 and 0.4, it is judged as a yellow warning, indicating that attention should be paid to the evolution trend of the supply and demand gap and preparations should be made for parameter fine-tuning.

[0111] When the adjustment lead time is between 2 and 4 hours, or the maximum gap depth is between 0.4 and 0.6, an orange alert is issued, indicating that the adjustment time is tight and supplementary measures need to be initiated immediately.

[0112] When the adjustment lead time is less than 2 hours, or the maximum gap depth is greater than or equal to 0.6, a red alert is issued, indicating that the adjustment time is seriously insufficient or the supply-demand gap is too large, requiring an emergency switch to the contingency plan or the activation of a combination of multiple contingency plans.

[0113] The above rules use an OR logic connection between time conditions and gap conditions, reflecting the emergency management principle of erring on the side of false alarms rather than false alarms. Even if the current adjustment lead time is still acceptable, if the predicted maximum gap depth is large, the warning level should be raised to allow more adjustment time.

[0114] This embodiment solves the problem that traditional methods cannot predict when a contingency plan will fail by constructing a timing fit curve and extracting key failure features, and realizes a forward-looking early warning based on time lead.

[0115] In one possible embodiment, the plan adjustment instruction includes at least one of the following:

[0116] Parameter fine-tuning instruction: While keeping the current contingency plan level unchanged, make incremental corrections to the resource scheduling intensity or evacuation range parameters in the current execution plan;

[0117] Supplementary measures instruction: Based on the gap analysis between the characteristics of scenario demand and the characteristics of contingency plan supply, reserve resources are called from the pre-stored emergency resource pool to fill the supply gap;

[0118] Contingency plan switching command: Terminate the currently executing contingency plan and activate a more suitable alternative contingency plan or a combination of multiple contingency plans.

[0119] In this embodiment, the system is based on the lead time L adjust Length and gap depth G max The size of the force determines the adjustment strategy for different strengths.

[0120] In some implementations, when a yellow alert is triggered, for example, with a lead time of 4 to 6 hours, a parameter fine-tuning instruction is generated. At this time, the main framework of the contingency plan remains in effect, requiring only parameter adjustments. Specifically, the system identifies the dimension k that will fail first and calculates the supply-demand gap Gap for that dimension at the time of failure. k And calculate the required supply increment. New supply configuration P new The calculation formula is as follows:

[0121] P new =P old *(1+γ*(1-A kfail ));

[0122] Where P new For the adjusted supply, P old The original supply quantity is given by γ, which is the adjustment intensity coefficient, and A is the original supply quantity. kfail This represents the fit of this dimension at the time of failure.

[0123] In other implementations, when an orange alert is triggered, for example, with a lead time of 2 to 4 hours, a supplementary measure instruction is generated. At this point, parameter fine-tuning alone is insufficient to resolve the issue, and backup resources must be invoked. The system calculates the cumulative gap integral for each dimension within the prediction window, using the following formula:

[0124] CumGap k =∫max(0,R) k (t)-P k (t))*dt;

[0125] CumGap k Let be the cumulative gap in the k-th dimension, and let the integration interval be from the current time to the end of the prediction window. The system sorts each dimension according to the size of the cumulative gap and retrieves available backup resources from the emergency resource pool, such as social rescue forces and support from neighboring cities, and calls them in order of priority to fill the gap.

[0126] In a more urgent implementation, when a red alert is triggered, such as when the lead time is less than 2 hours, a contingency plan switching command is generated. At this point, the current contingency plan is completely ineffective and must be switched. The system traverses the candidate contingency plans in the contingency plan library, calculates the timing fit curve between each candidate contingency plan and the current scenario, and selects the contingency plan with the largest adjustment lead time as the new contingency plan.

[0127] In a further embodiment, activating a combination of multiple contingency plans includes:

[0128] For each dimension of the scenario requirements, calculate the single-dimensional adaptability of each plan in the pre-stored alternative plan library on that dimension;

[0129] The measure with the highest single-dimensional adaptability in each plan is extracted as the execution measure for that dimension;

[0130] The optimal implementation measures extracted from each dimension are recombined to construct a temporary combined plan for execution.

[0131] In this embodiment, when a single contingency plan cannot meet the needs of extreme disasters, an optimal contingency plan is constructed using an optimal contingency splicing method. Specifically, the system treats each contingency plan in the contingency plan library as a set of measures. Assume there are two alternative contingency plans, A and B, and three demand dimensions: personnel evacuation, material support, and engineering scheduling. The system calculates the single-dimensional fit of plans A and B in these three dimensions. Assume plan A scores highly in the personnel evacuation dimension, while plan B scores highly in the engineering scheduling dimension. The system extracts the personnel evacuation measures module from plan A, the engineering scheduling measures module from plan B, and the material support module from the one with the higher material support score. The extracted optimal modules are logically reorganized, and after conflict detection (e.g., checking whether the same road resources have been requisitioned), they are integrated into a new temporary combined contingency plan. This fully utilizes the strengths of different contingency plans to construct a super contingency plan adapted to the current complex scenario.

[0132] It should be noted that in extreme disaster scenarios, a single contingency plan often cannot meet all the needs simultaneously. For example, Contingency Plan A, designed for regular rainstorms, may have strong personnel evacuation capabilities, equipped with a large number of inflatable boats and rescue personnel, but its engineering dispatch capabilities are limited; while Contingency Plan B, designed for dam breaches, has strong engineering dispatch capabilities, pre-positioned high-powered pump trucks and emergency rescue machinery, but its personnel evacuation coverage is limited.

[0133] This embodiment of multi-plan combination treats different plans as a collection of detachable measure modules, rather than an indivisible whole. By calculating the single-dimensional adaptability of each plan across different dimensions, the measure modules of each plan in their respective advantageous dimensions are extracted and reassembled into a temporary combined plan. This can break through the capability boundaries of a single plan, fully leverage the strengths of different plans, and construct a super emergency plan adapted to the current complex scenario.

[0134] Before actual implementation, the system needs to perform conflict detection between modules. For example, it checks whether the same road is simultaneously planned as a material transportation route and a personnel evacuation route. If so, the route needs to be adjusted or staggered scheduling needs to be implemented to avoid execution conflicts caused by resource competition. The combined plan after constraint verification can be issued for execution to achieve optimal allocation of emergency resources.

[0135] In one embodiment of this application, the method further includes offline correction of mobilization response parameters, specifically:

[0136] Obtain data on the actual arrival time of resources in historical emergency events;

[0137] Calculate the deviation between the theoretical arrival time predicted using mobilization response parameters and the actual arrival time;

[0138] The mobilization response parameters are corrected and updated based on the deviation, and the updated mobilization response parameters are stored for use in the construction of time-series supply sequences.

[0139] In this embodiment, real-world feedback data is used to improve the model's prediction accuracy. After each emergency event, the system extracts the actual arrival time T of the resources from the logs. actual Simultaneously, the reading model utilizes the mobilization rate parameter λ when the event occurs. old Predicted theoretical arrival time T predicted Calculate the deviation between the two: Error=T actual -T predicted If the deviation is positive, it indicates that the actual mobilization is slower than predicted, and λ needs to be decreased; conversely, if the deviation is negative, λ should be increased. The parameter update uses the following formula:

[0140] λ new =λ old -ε*(Error / T predicted );

[0141] Where λ new The updated mobilization rate parameter is ε, where ε is the learning rate coefficient, typically set to a small value such as 0.01 to ensure the stability of parameter evolution. The updated mobilization rate parameter will be stored in the database and retrieved the next time-series supply sequence P(t) is generated. This endows the system with the ability to self-evolve, making it increasingly accurate with use.

[0142] Alternatively, parameter updates can also be performed using the following formula:

[0143] λ new =λ old (1-ε*(Error / T predicted )).

[0144] This embodiment achieves precision and intelligence in adjustment actions through a hierarchical strategy and a multi-plan combination algorithm.

[0145] According to one aspect of this application, a hardware system architecture for implementing the methods described in any of the above embodiments is also provided. The system composition can be described from the perspective of functional modules, which can be implemented wholly or partially by software programs running on a computer device. Specifically, a smart contingency plan adjustment system for urban flood control includes a data acquisition module, a feature construction module, an adaptation evaluation module, and a dynamic adjustment module.

[0146] The data acquisition module is used to acquire real-time flood control monitoring data and resource configuration data of the current implementation plan for the target area.

[0147] Specifically, the data acquisition module can include multiple communication interface sub-units, which connect to front-end sensors such as water level gauges and rain gauges via Internet of Things (IoT) protocols, and connect to the meteorological bureau's data center and emergency management resource database via database interfaces (ODBC / JDBC). The data acquisition module is responsible for data collection, cleaning, format conversion, and time synchronization to ensure the quality of data in subsequent processing.

[0148] The feature construction module is used to construct scenario demand features that reflect the needs of disaster evolution based on real-time flood monitoring data, and to extract plan supply features that reflect the supply of emergency response capabilities based on resource allocation data.

[0149] In its implementation, the feature construction module integrates a GIS spatial analysis engine and a hydrodynamic calculation unit. It can rapidly perform simulations using a pre-built hydrodynamic model library and overlay the simulation results with population raster data to calculate a time-series demand sequence. Simultaneously, the feature construction module also includes a built-in operations research calculation unit, utilizing a resource mobilization cumulative function to generate a time-series supply sequence.

[0150] The adaptation assessment module is used to comprehensively calculate the matching relationship between the scenario demand characteristics and the contingency plan supply characteristics within a preset target time window to obtain the contingency plan adaptation status.

[0151] In this embodiment, the fit evaluation module is the core computing engine of the system, which incorporates a nonlinear fit function library and a geometric mean algorithm library. It supports high-concurrency matrix operations and can quickly complete supply and demand matching calculations for hundreds or thousands of time steps within the future prediction window, generating a visualized time-series fit curve.

[0152] The dynamic adjustment module is used to generate a plan adjustment instruction in response to the pre-set dynamic adjustment trigger conditions being met in the plan adaptation state.

[0153] Specifically, the dynamic adjustment module is essentially an intelligent decision support system. It continuously monitors the curve characteristics output by the adaptation and evaluation module. Once it detects that the lead time is less than a threshold, it immediately triggers the built-in rule engine or optimization algorithm, such as a multi-plan combination algorithm, to generate specific scheduling instructions. The dynamic adjustment module can also connect to a large-screen display terminal to present red / orange / yellow warning signals to command personnel in the form of pop-ups or audible and visual alarms.

[0154] At the hardware level, the aforementioned modules can be deployed on one or more high-performance servers. Server configuration includes, but is not limited to: a central processing unit (CPU) for executing data processing and logical judgment instructions; a graphics processing unit (GPU) for accelerating grid computing and deep learning algorithms for hydrodynamic models; memory for storing massive amounts of historical monitoring data, contingency plan libraries, and system program code; and a network communication interface for data interaction with external sensor networks and command terminals. Furthermore, the system may include input / output devices such as displays, keyboards, and mice for administrators to configure parameters and perform manual intervention.

[0155] In one possible embodiment, the process of generating a plan adjustment instruction in response to the plan adaptation state meeting a preset dynamic adjustment trigger condition is as follows: After the system initially generates an adjustment plan, such as a parameter fine-tuning plan or a measure supplement plan, it does not directly issue an execution order, but instead inputs it as a plan to be verified into the simulation engine for rapid verification. Specifically, the verification process includes two core steps: constraint verification and effect pre-evaluation.

[0156] In the constraint verification phase, the system checks whether the solution to be verified meets the following three types of hard constraints: First, resource constraints. The system verifies the adjusted supply quantity P in a certain dimension. k_new Is it less than or equal to the maximum available resource P for this dimension? k_max That is, satisfying relation P. k_new ≤P k_max For example, if calculations indicate that 500 additional rescue personnel are needed, the system needs to verify whether there are indeed 500 available personnel in the backup database. Secondly, there is a time constraint. The system verifies the deployment time T of the resources. deploy Is it less than or equal to the allowable time window T for disaster evolution? available That is, satisfying relation T. deploy ≤T available Among them, T deploy Calculated based on the physical location of the resource and the transportation speed, T available The inundation time is calculated based on the flood evolution model. If the area is already flooded and inaccessible by the time resources arrive, the plan is invalid. Third, conflict constraints. The system checks whether there are spatial or logical conflicts between the adjustment measures in each dimension, such as checking whether the material transportation route overlaps with the personnel evacuation route, causing traffic jams.

[0157] In the preliminary effect evaluation section, the system will substitute the constraint-tested solutions into the fitness calculation model to recalculate the future overall fitness. The specific calculation formula is as follows:

[0158] A total_new =f(R(S predict ), P new );

[0159] Where A total_new The adjusted expected overall fit; R(S) predict P represents a short-term scenario demand forecast; new The adjusted plan provides the supply configuration; f() is the comprehensive suitability calculation function.

[0160] The system sets a pre-evaluation pass threshold, for example, 0.7. If A total_new If A is ≥0.7, the adjustment plan is deemed valid, and a formal adjustment instruction is generated and issued; if A total_new If the value is less than 0.7, the solution is deemed feasible but ineffective. The system will automatically trigger an iterative optimization process, readjusting parameters or increasing resource allocation until the threshold requirement is met or all resources have been exhausted. This closed-loop mechanism of generation, verification, optimization, and release ensures the practical value of the instructions.

[0161] In an optional implementation, an anomaly handling mechanism is also included. When missing sensor data is detected, the system can interpolate and fill in the missing data using the most recent valid data, or call up backup sensor data. When the data loss time exceeds a preset threshold, such as three consecutive sampling periods, the system triggers a data anomaly alarm and switches to a conservative mode, i.e., using the statistical average of historical data from the same period as a substitute input. When a communication interruption is detected, the system can enable a locally cached contingency plan library for offline decision-making and automatically synchronize the data after communication is restored. When the cumulative error of the prediction model exceeds a preset limit, the system automatically triggers a model recalibration process, and during calibration, the triggering conditions are adjusted to a more conservative threshold setting.

[0162] This invention employs a temporal evolutionary game theory mechanism. By introducing a hydrodynamic model to deduce the future dynamic demand sequence, and combining it with a resource mobilization accumulation model to construct a supply sequence reflecting the nonlinear growth process of resources, the evaluation dimension is expanded from static comparison at a single moment to continuous curve fitting within a future prediction window. This achieves precise temporal matching between the dynamic demand of flood evolution and the gradual arrival of resources, ensuring that the evaluation results truly reflect the interaction between supply and demand over time. A forward-looking triggering mechanism based on adjustment lead time is adopted. By generating a temporal fit curve, the theoretical failure moment of the future plan is accurately located, and the time difference between that moment and the current moment is calculated as the adjustment lead time. The system no longer waits for monitoring indicators to exceed limits; instead, adjustments are initiated immediately when the lead time is less than the safety buffer threshold. This provides sufficient time windows for the physical scheduling and transportation of resources, effectively overcoming the time difference risk caused by mobilization delays and improving the predictability and timeliness of emergency response.

[0163] The preferred embodiments of the present invention have been described in detail above. However, the present invention is not limited to the specific details of the above embodiments. Within the scope of the technical concept of the present invention, various equivalent transformations can be made to the technical solutions of the present invention, and these equivalent transformations all fall within the protection scope of the present invention.

Claims

1. A smart emergency response plan method for urban flood control, characterized in that, include: Obtain real-time flood control monitoring data and resource allocation data for the current implementation plan for the target area; Based on real-time flood control monitoring data, scenario demand characteristics reflecting the needs of disaster evolution are constructed, and contingency plan supply characteristics reflecting the supply of emergency response capabilities are extracted based on resource allocation data. By comprehensively calculating the matching relationship between the demand characteristics of the comprehensive calculation scenario and the supply characteristics of the contingency plan within a preset target time window, the contingency plan adaptation status is obtained. In response to the pre-set dynamic adjustment triggering conditions being met in the contingency plan adaptation status, a contingency plan adjustment instruction is generated.

2. The method according to claim 1, characterized in that, The scenario demand characteristics are quantified into a multi-dimensional demand vector that includes at least the dimensions of personnel transfer demand, engineering scheduling demand, and material support demand; the contingency plan supply characteristics are quantified into a multi-dimensional supply vector corresponding to the dimensions of the multi-dimensional demand vector.

3. The method according to claim 2, characterized in that, The target time window is the current moment; The contingency plan adaptation status has been obtained, including: Calculate the ratio of the multidimensional supply vector to the multidimensional demand vector at the current moment, and determine the calculated ratio as the plan adaptation state.

4. The method according to claim 1, characterized in that, The target time window is a forecast window that includes the current time to a preset future time period; The scenario demand characteristics are constructed as a time-series demand sequence that dynamically changes with the evolution of the flood, and the contingency plan supply characteristics are constructed as a time-series supply sequence that gradually accumulates with the response time. The contingency plan adaptation status has been obtained, including: At each time step within the prediction window, the dynamic matching degree between the time-series supply sequence and the time-series demand sequence is calculated, a time-series fit degree curve reflecting the evolution trend of the fit degree over time is generated, and the time-series fit degree curve is determined as the plan fit state.

5. The method according to claim 4, characterized in that, The contingency plan's supply characteristics are constructed as a time-series supply sequence that accumulates progressively over response time, including: Obtain the theoretical total allocation of emergency resources and mobilization response parameters; Using a pre-defined resource supply accumulation function, the cumulative amount of resources in place from the start of the plan to each moment within the prediction window is calculated based on the theoretical total allocation and mobilization response parameters, resulting in a time-series supply sequence that reflects the nonlinear growth process of resource supply capacity over time.

6. The method according to claim 4, characterized in that, The matching relationship between the demand characteristics of the comprehensive computing scenario and the supply characteristics of the contingency plan within the preset target time window also includes marking the failure time, specifically: Traverse the generated time series fit curves, identify the time point when the time series fit curve first falls below the preset validity threshold within the prediction window, and mark the identified time point as the failure time.

7. The method according to claim 6, characterized in that, In response to the pre-set dynamic adjustment trigger conditions being met in the contingency plan adaptation state, including: Calculate the time difference between the failure time and the current time to obtain the adjustment lead; Determine whether the adjustment lead time is less than the preset safety buffer threshold; If so, then the dynamic adjustment trigger condition is met.

8. The method according to claim 2, characterized in that, The matching relationship between the demand characteristics of the comprehensive computing scenario and the supply characteristics of the contingency plan within a preset target time window is calculated, including the calculation of one-dimensional fit, specifically: For each dimension in the multidimensional demand vector and the multidimensional supply vector, calculate the supply-demand ratio of the supply characteristic value and the demand characteristic value; The supply-demand ratio is mapped to a one-dimensional fit degree using a pre-defined non-linear fit function. The mapping rule of the nonlinear fitting function is configured as follows: when the supply-demand ratio is within the preset supply-demand balance range, the single-dimensional fitting degree takes the maximum value; when the supply-demand ratio exceeds the upper limit of the supply-demand balance range, the single-dimensional fitting degree decreases as the supply-demand ratio increases.

9. The method according to claim 8, characterized in that, The contingency plan adaptation status has been obtained, including: Obtain the weight parameters corresponding to each dimension; Based on the weight parameters, a weighted geometric average is calculated for the single-dimensional fit of each dimension to obtain the comprehensive fit. The comprehensive fit is then determined as the pre-plan fit state, so that the comprehensive fit is zero when the single-dimensional fit of any dimension is zero.

10. A smart emergency response system for urban flood control, characterized in that: include: The data acquisition module is used to acquire real-time flood control monitoring data and resource configuration data of the current implementation plan for the target area; The feature construction module is used to construct scenario demand features that reflect the needs of disaster evolution based on real-time flood monitoring data, and to extract plan supply features that reflect the supply of emergency response capabilities based on resource allocation data. The adaptation assessment module is used to comprehensively calculate the matching relationship between the scenario demand characteristics and the contingency plan supply characteristics within a preset target time window to obtain the contingency plan adaptation status. The dynamic adjustment module is used to generate a plan adjustment instruction in response to the pre-set dynamic adjustment trigger conditions being met in the plan adaptation state.

Images