Intelligent Decision-Making Methods for Emergency Response to Sudden Environmental Pollution Incidents

Abstract

This invention discloses an intelligent decision-making method for emergency response to sudden environmental pollution incidents, specifically relating to the field of environmental emergency management. The method includes generating an event table from riverside edge computing nodes, constructing a river hypergraph and calculating a propagation table; performing source-based comparison propagation of successful and failed interceptions at candidate interception locations to generate a failure table; binding the first-round interception, remedial interception, resource scheduling paths, and deployment priority as disposal columns; and outputting the emergency response decision results after K-adaptive recovery robust optimization and write-back verification. By generating successful and failed interception propagation rows for the same candidate interception location, and calculating the consequences of interception failure based on the difference in their destinations and time compression, the interception deployment priority no longer depends solely on distance, concentration, or resource arrival time, but incorporates the pollution destination, remediation window, and the impact on sensitive targets after failure into the selection process.

Description

Technical Field

[0001] This invention relates to the field of environmental emergency management technology, and more specifically, to an intelligent decision-making method for emergency response to sudden environmental pollution incidents. Background Technology

[0002] Emergency response to sudden water pollution incidents usually focuses on predicting pollutant migration and on-site interception and dispatch. Existing systems mostly rely on river hydrological data, cross-sectional online monitoring data, pollutant concentration changes, and emergency resource locations to estimate the arrival time of the pollution front downstream of the river and select the deployment locations of oil booms, absorbent materials, emergency dosing points, or temporary interception facilities at downstream cross-sections. In scenarios where rainwater from a chemical industrial park is abnormally discharged into a river, the polluted water mass will pass through tributary inlets, sluice gates, water intakes, slow-flowing areas on beaches, and construction cofferdams under bridges in a short period of time. Although edge computing nodes deployed on the side of the park or river can access water level, flow velocity, flow rate, cross-sectional concentration, team location, and material warehouse status nearby, the deployment of oil booms, the placement of adsorbent materials, the setting of anchor points, the arrival of personnel, and the passage of vehicles all occupy a limited treatment window. If any near-end cross-section fails to be deployed or forms a bypass, the polluted water mass will continue to enter the downstream river section and change the subsequent interception conditions. Existing systems typically prioritize interception points that are close to the accident site, are expected to arrive quickly, have high pollution concentrations, or have short resource arrival times. However, they fail to consider the destination of the pollution after the interception at the current point fails, the remaining treatment time at the backup point, and the impact on sensitive targets in the priority determination. During on-site treatment, it is easy for polluted water masses to be introduced into tributary inlets, drinking water source intakes, or low-velocity stagnation areas after the near-end interception fails. Although subsequent teams can continue to track the polluted water masses, they have already missed the time to deploy downstream remedial points. The technical problem this application aims to solve is: how to use edge computing nodes to calculate the consequences of failed interception at candidate interception locations in emergency response to sudden water pollution entering a river, so that the order of interception deployment takes into account the destination of the pollution after failure, the remaining time of the remedial section, and the risk reduction relationship of sensitive targets. Summary of the Invention

[0003] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide an intelligent decision-making method for emergency response to sudden environmental pollution incidents. This method constructs a river hypergraph using riverside edge computing nodes, performs source-to-source propagation on the same candidate interception location under both successful and failed interception states, calculates the pollution destination, remedial interception location, remaining remediation time, and affected status of sensitive targets after interception failure, and binds the initial interception location, remedial interception location, resource scheduling path, and deployment order as a disposal column to perform K-adaptive recovery robust optimization, thereby addressing the problems mentioned in the background art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: an intelligent decision-making method for emergency response to sudden environmental pollution incidents, comprising: S1. The riverside edge computing node acquires pollution event data, river data, and resource data, performs time alignment and spatial mapping according to the same event number and the same simulation time, and generates an event table. S2. Based on the event table, construct a river hypergraph by including the propagation bifurcation points, flow control points, candidate interception points, and sensitive targets in the river channel, and calculate the propagation duration and flow rate of the polluted water mass in the river hypergraph according to the pollution event data to generate a propagation table. S3. Based on the river hypergraph and propagation table, by setting the interception success status and interception failure status of the same candidate interception location, the same source comparison propagation is performed to obtain the pollution destination after interception failure, the affected status of sensitive targets, the remedial interception location and the remaining remedial time, and generate the failure table. S4. Based on the failure table and resource data, K-adaptive recovery robust optimization is performed by binding the first-round interception position, the remedial interception position, the resource scheduling path and the deployment order to the same disposal column, so that the disposal column retains the remedial interception position and the remaining remedial duration after the interception fails, and a disposal table is generated. S5. Based on the disposal table, write the first round of interception locations back to the river supermap and update the flow distribution. Re-execute the same-source comparison propagation. When the updated sensitive target affected status and the remaining remediation time meet the disposal table, output the emergency disposal decision results.

[0005] In a preferred embodiment, S1 includes: S1-1. Read the river entry point coordinates, river entry time and pollutant identifier from the pollution event data. Read the river segment number, river segment start coordinates, river segment end coordinates and river segment connection order from the river channel data. Project the river entry point coordinates onto the nearest river segment centerline. Generate the river entry mileage position according to the centerline distance from the projection point to the river segment start point. Output the event baseline row. S1-2. Read the resource number, resource location and resource activation status from the resource data, project the resource location onto the river segment centerline in the river data, generate the resource mileage position according to the river segment number where the projection point is located and the distance from the centerline, and write the resource number, resource mileage position and resource activation status into the event base line to generate the resource attachment line. S1-3. Using the river entry time as the first simulation time, bind the event base row and resource connection row to the same event number in ascending order of adjacent simulation times, and generate an event table including event number, simulation time, river entry mileage position, river segment connection order, resource number, resource mileage position, and resource activation status.

[0006] In a preferred embodiment, S2 includes: S2-1. Read the river inlet mileage position from the event table, and read the river segment row, tributary confluence row, gate control row, candidate interception row, and sensitive target row from the river data; determine the confluence mileage position in the tributary confluence row as the propagation bifurcation position, the gate mileage position in the gate control row as the flow control position, the interception mileage position in the candidate interception row as the candidate interception position, and the water intake mileage position, residential water supply section mileage position, and aquaculture water intake section mileage position in the sensitive target row as the sensitive target position; and generate a river node table by assigning the river inlet mileage position, propagation bifurcation position, flow control position, candidate interception position, and sensitive target position to the river segment number and the distance from the center line. Among them, the river section rows, tributary confluence rows, gate control rows, candidate interception rows, and sensitive target rows in the river data refer to the records of the river section centerline, the confluence of tributaries into the main river, the records of gate or control dam flow control facilities, the records of interception sections with conditions for personnel and material deployment, and the protection target records of water intakes, residential water use sections, and aquaculture water intake sections extracted from basic river mapping data, water conservancy facility ledgers, emergency interception point plan databases, and water environment protection target lists, respectively.

[0007] In a preferred embodiment, S2 further includes: S2-2. Within the same river segment number, sort the river nodes in the river node table from upstream to downstream according to the distance from the center line, and connect adjacent river nodes to the main river edge; at the same propagation bifurcation position, connect the propagation bifurcation position to the first river node of each downstream tributary and generate a branching edge; at the same flow control position, connect the flow control position to the adjacent downstream river node and generate a flow control edge; generate a river hypergraph from the river node table, main river edge, branching edge, and flow control edge; S2-3. Read the inflow volume and inflow time from the pollution event data. Read the side length, flow velocity, and downstream flow of each riverbank in the river data. Divide the side length by the flow velocity to obtain the propagation time. Multiply the inflow volume by the proportion of the single downstream flow at the same propagation bifurcation point to the sum of all downstream flows to obtain the pollution distribution volume of the bifurcation side. Use the inflow volume as the pollution distribution volume of the non-bifurcation side. Write the riverbank number, upstream river node, downstream river node, propagation time, pollution distribution volume, and arrival time into the propagation table.

[0008] In a preferred embodiment, S3 includes: S3-1. Read the river hypergraph, propagation table and candidate interception locations. Record the river edge where the candidate interception location is located as the control edge. Use the arrival time of the upstream node of the control edge and the pollution distribution flow of the control edge as the same source input to generate the successful interception propagation line and the failed interception propagation line respectively. The successful interception propagation line and the failed interception propagation line share the same source input: the arrival time of the upstream node on the riverbank where the same candidate interception location is located and the pollution distribution flow. The successful interception propagation line corresponds to the remaining pollution after deducting the interception amount propagating downstream, while the failed interception propagation line corresponds to the undeducted pollution distribution flow propagating along the same downstream path. S3-2. In the successful propagation row of interception, the downstream output of the control edge is recorded as the remaining amount after subtracting the interception amount from the pollution distribution flow of the control edge, and is recursively calculated along the downstream connection sequence of the river supergraph. In the failed propagation row of interception, the downstream output of the control edge is recorded as the pollution distribution flow of the control edge, and is recursively calculated along the same downstream connection sequence to obtain the successful path table and the failed path table. The edge-by-edge recursion refers to starting from the initial riverbank and, according to the downstream connection order in the river supergraph, using the arrival time and pollution fraction of the previous riverbank as the input for the next riverbank, calculating the arrival time and pollution fraction of the next riverbank line by line. In addition, the interception quantity is obtained by multiplying the quantity of interception material, the adsorption capacity or retention capacity of a single material in the resource data corresponding to the candidate interception location, and the deployment effectiveness coefficient. The unit is consistent with the pollution flow rate. The pollution flow rate is measured in kilograms when measured by mass, in cubic meters when measured by volume, and in kilograms per second when measured by concentration flux.

[0009] In a preferred embodiment, S3 further includes: S3-3. For riverbanks with the same downstream node in the successful path table and the failed path table, calculate the destination difference by subtracting the successful pollution distribution from the failed pollution distribution based on the same downstream node, and calculate the time compression by subtracting the failed arrival time from the successful arrival time. Record the downstream node with a destination difference greater than zero as the destination of the pollution after the interception failure. S3-4. Read the candidate interception positions and sensitive target positions in the failure path table, delete the candidate interception positions whose arrival time is later than the arrival time of the sensitive target position, record the remaining candidate interception positions as the remedial interception positions, and obtain the remaining remedial time by subtracting the arrival time of the remedial interception position from the arrival time of the sensitive target position. Write the candidate interception position, contamination destination, destination difference, time compression amount, affected status of the sensitive target, remedial interception position and remaining remedial time into the failure table. The affected status of sensitive targets is generated by the riverbank where the endpoint node in the failure path table is the location of the sensitive target and the pollution diversion flow is greater than zero. Specifically, it includes the sensitive target number, arrival time, pollution diversion flow, and candidate interception location that triggers the arrival result.

[0010] In a preferred embodiment, S4 includes: S4-1. Read the failure table, and record the failure table rows with different contamination destinations or different sensitive target affected states under the same candidate interception position as recovery scenario rows, and record the number of recovery scenario rows under the same candidate interception position as the K value. S4-2. Read resource data, record the same candidate interception position as the first round interception position, and generate a remediation column for each recovery scenario row; each remediation column includes the remediation interception position, resource scheduling path, resource transition time from the first round interception position to the remediation interception position, remediation deployment time, and remediation completion time; Among them, the remedial interception location is the candidate interception location in the failure table that is located downstream of the pollution destination and arrives earlier than the arrival time of the sensitive target; the resource scheduling path is the path record in the resource data from the current location of the resource to the first-round interception location via the road or river patrol channel and then to the remedial interception location; the transfer time is the time required for the resource to travel from the first-round interception location to the remedial interception location along the resource scheduling path; the remedial deployment time is the time required for the resource to complete the deployment of the oil boom, the placement of absorbent materials or the deployment of temporary interception facilities at the remedial interception location; and the remedial completion time is the time obtained by adding the transfer time and the remedial deployment time.

[0011] In a preferred embodiment, S4 further includes: S4-3. For the K remedial columns under the same first-round interception position, subtract the remedial completion time from the remaining remedial time to obtain the scene remaining time, and subtract the interception amount corresponding to the remedial interception position from the pollution distribution flow in the recovery scene row to obtain the remaining pollution amount after remediation, and generate the recovery column. S4-4. Merge K recovery columns according to the same first-round interception position, delete the first-round interception positions with negative scene remaining time, and determine the target interception position according to the sum of the remaining pollution after remediation in the K recovery columns from low to high and the sum of scene remaining time from high to low. Write the target interception position, K recovery columns, resource scheduling path and deployment order into the disposal table. The resource scheduling path is obtained by accumulating the current location of the resource, the first interception location, and the remedial interception location in the resource data in the road or river patrol channel network according to the travel time. The deployment order is obtained by arranging the target interception location before the remedial interception locations in the K recovery columns, and the scene remaining time in the same recovery column from shortest to longest.

[0012] In a preferred embodiment, S5 includes: S5-1. Read the target interception location, interception amount and K recovery columns in the treatment table, locate the riverbank where the target interception location is located, and subtract the interception amount from the pollution distribution flow of the riverbank where the target interception location is located in the propagation table to obtain the remaining distribution flow. Write the remaining distribution flow back to the riverbank where the target interception location is located to generate the write-back propagation table. S5-2. Based on the write-back propagation table, starting from the riverbank where the target interception location is located, the arrival time and pollution distribution flow are recursively calculated along the downstream connection sequence of the river supergraph. Successful verification path tables and failed verification path tables are generated according to the interception success status and interception failure status, respectively.

[0013] In a preferred embodiment, S5 includes: S5-3. Read the K recovery columns in the failure review path table and the disposal table, and compare the remediation interception location, remediation remaining time, and remediation remaining pollution amount in each recovery column; Among them, the remedial interception position is obtained from the candidate interception position in the failure review path table that is downstream of the pollution destination and arrives earlier than the sensitive target. The remaining remedial time is the arrival time of the sensitive target minus the arrival time of the remedial interception position. The remaining pollution amount after remediation is the pollution diversion flow of the remedial interception position minus the remedial interception amount. S5-4. When the same remedial interception position exists in all K recovery columns, and the recalculated remaining remedial time is not less than the remedial completion time in the corresponding recovery column, and the recalculated remaining pollution after remediation is not greater than the remaining pollution after remediation in the corresponding recovery column, the target interception position, the remedial interception position in the K recovery columns, the resource scheduling path, and the deployment order are written into the emergency response decision results.

[0014] The technical effects and advantages of this invention are as follows: 1. By generating successful interception propagation lines and failed interception propagation lines for the same candidate interception location, and calculating the consequences of interception failure based on the difference in destination and time compression between the two, the order of interception deployment no longer depends solely on distance, concentration, or resource arrival time, but can incorporate the destination of pollution after failure, the remediation window, and the impact on sensitive targets into the selection process. 2. By binding the initial interception location, the backup interception location, the resource scheduling path, and the deployment priority into the same handling column, the corresponding backup action is formed when the target interception location is selected, which relatively reduces the processing delay of temporarily finding a backup section after the initial interception fails. 3. Record the main channel side, branch side and control side through the river super map, and calculate the propagation time and pollution distribution flow according to the side length, flow velocity and downstream flow, so that the polluted water mass forms a recursive propagation table when it passes through tributaries, gates and water intakes; 4. The remaining time of the scenario is obtained by subtracting the completion time of the remediation from the remaining remediation time, and the remaining pollution after remediation is obtained by subtracting the remediation interception amount from the pollution diversion amount, so that the selection of the remediation interception location has two calculation bases: time and pollution amount. 5. By jointly filtering target interception locations by K recovery columns, and deleting first-round interception locations with negative scene time, the selected target interception locations can cover multiple failure destinations under the same candidate interception location; 6. By writing the target interception location back to the river supermap and recalculating the successful verification path table and the failed verification path table, the emergency response decision results are verified by the propagation results before being output, which relatively reduces the risk of the response table becoming out of sync with the actual propagation changes. Attached Figure Description

[0015] Figure 1 This is a flowchart of the method steps of the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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 are within the scope of protection of the present invention.

[0017] Refer to the instruction manual appendix Figure 1 The intelligent decision-making method for emergency response to sudden environmental pollution incidents of the present invention includes: S1. The riverside edge computing node acquires pollution event data, river data, and resource data, performs time alignment and spatial mapping according to the same event number and the same simulation time, and generates an event table. This implementation method converts pollution event data, river data, and resource data into an event table, integrating river entry location, resource location, river segment priority, and projection time into the same river mileage system for subsequent river super-map construction, pollution propagation calculation, and disposal column generation. The processing starts with the river entry point, supplemented by resource integration, and uses the projection time as a time index to form an event table that can directly participate in subsequent calculations. This implementation method includes the following steps: In S1-1, the river channel edge calculation node reads the event number, inflow point coordinates, inflow time, pollutant identifier, inflow volume, and measurement method from the pollution event data. It also reads the river segment number, river segment centerline, river segment start coordinates, river segment end coordinates, river segment length, and river segment connection order from the river channel data. For each river segment centerline, it calculates the vertical distance from the inflow point coordinates to the river segment centerline, selects the river segment with the lowest vertical distance value as the river segment to which the inflow point belongs, and projects the inflow point coordinates onto the centerline of the river segment to which the inflow point belongs. When the projection point is located between the river segment start and end points, the projection point is used as the river channel projection point for the inflow point. The projection point is located outside the river segment start or at the river segment end. When the point is outside the river, the endpoint of the river segment closest to the river entry point coordinates is used as the river channel projection point of the river entry point; when there are two river segments with the same vertical distance value, the river segment located downstream in the river segment connection sequence is selected as the river segment to which the river entry point belongs; with the starting point of the river segment to which the river entry point belongs as the zero position, the distance of the center line is accumulated along the river segment centerline to the river segment end point to obtain the river entry mileage position, and the event number, river entry time, pollutant identification, river entry volume, measurement method, river segment number to which the river entry point belongs, river entry mileage position and river segment connection sequence are written into the event reference line; when the river entry point coordinates cannot be projected to the centerline of any river segment, no event reference line is generated, and a river entry point connection failure record is output; In S1-2, the riverside edge computing node reads the resource number, resource location, resource activation status, resource type, resource quantity, single-piece material adsorption capacity or interception capacity, deployment effectiveness coefficient, first-round deployment duration, and remedial deployment duration from the resource data. When the resource activation status is not activated, the resource data corresponding to the resource number does not participate in resource attachment. When the resource activation status is activated, the vertical distance from the resource location to the centerline of each river segment is calculated, the river segment with the lowest vertical distance value is selected as the river segment to which the resource belongs, and the resource location is projected onto the centerline of the river segment to which the resource belongs. When the projection point is located between the start and end points of the river segment, the projection point is used as the resource river channel projection point. When the projection point is located outside the start or end point of the river segment, the distance to the resource is used as the projection point. The endpoint of the river segment closest to the source location is used as the projection point of the resource channel. When two river segments have the same vertical distance, the river segment downstream of the inflow point and connected to the inflow point is selected as the river segment to which the resource belongs. Taking the starting point of the river segment to which the resource belongs as the zero point, the distance between the center lines along the river segment centerline and the ending point is accumulated to obtain the resource mileage position. The resource number, resource type, resource quantity, resource mileage position, resource activation status, single material adsorption capacity or interception capacity, deployment effectiveness coefficient, first round deployment time and remedial deployment time are written into the event baseline line to generate the resource attachment line. When the resource location cannot be projected to the centerline of any river segment, the resource number is written into the resource attachment failure record, but not into the resource attachment line. In S1-3, the riverside edge computing node uses the river entry time in the event baseline row as the first simulation time, reads the simulation step size configured by the environmental emergency platform, and generates an increasing sequence of simulation times according to the simulation step size. The simulation step size comes from the scheduling cycle or online monitoring data upload cycle of the environmental emergency platform and is fixed under the same event number. For the event baseline row and resource connection row under the same event number, the river entry mileage position, river segment connection order, resource number, resource mileage position, resource activation status, resource type, resource quantity, and deployment duration are bound together. At each simulation time, an event table is generated; when multiple resource connection rows with the same resource number exist at the same simulation time, the resource connection row with the later collection time is retained, and the number of the covered resource connection row is written into the overwrite record field of the event table; when the measurement method is mass, the inflow, pollution diversion, interception and residual pollution are all in kilograms; when the measurement method is volume, they are all in cubic meters; when the measurement method is concentration flux, they are all in kilograms per second; the simulation time, the first round of deployment duration and the remedial deployment duration are in minutes, and the inflow mileage and resource mileage are in meters; Through the above processing, the coordinates of the river entry point are converted into river entry mileage positions, and the resource location is converted into resource mileage positions. Both pollution event data and resource data are entered into the time series record under the river segment connection order. Subsequently, S2 can read the river entry mileage position to generate river node, and S4 can read the resource mileage position to calculate the resource scheduling path. When the river entry point or resource location falls outside the river segment endpoint, the endpoint value is used to avoid the projection result from deviating from the river segment centerline. When multiple river segments are the same distance, the upstream and downstream relationship is limited by the river segment connection order to avoid the same coordinate point being attached to unconnected river segments. In practical applications, for example, if an abnormal discharge occurs at the storm drain inlet of a chemical industrial park at 10:00, the riverside edge calculation node projects the storm drain coordinates onto the center line of the river segment numbered R03, obtaining the river inlet mileage of 320 meters. It also projects two oil boom storage points and one adsorbent material storage point onto the corresponding river segment center line, obtaining the resource mileage. The system generates simulation times at 10:00, 10:05, and 10:10, and writes the river inlet volume, metering method, resource quantity, resource mileage, and deployment duration into the event table.

[0018] S2. Based on the event table, construct a river hypergraph by including the propagation bifurcation points, flow control points, candidate interception points, and sensitive targets in the river channel, and calculate the propagation duration and flow rate of the polluted water mass in the river hypergraph according to the pollution event data to generate a propagation table. This implementation method converts the entry mileage in the event table and key locations in the river channel data into a river hypergraph, and calculates the propagation time, pollution fraction, and arrival time of the polluted water mass along the main channel, tributaries, and flow control points on the river hypergraph. The process first generates a river node table, then generates main channel edges, tributary edges, and flow control edges according to upstream and downstream connections, and finally calculates a propagation table based on edge length, flow velocity, and downstream flow, enabling subsequent source-based comparative propagation to read the arrival time and pollution fraction along each edge. This implementation method includes the following steps: In S2-1, this step is used to convert the river entry location, bifurcation location, flow control location, interception location, and sensitive target location into river nodes under the same river segment number, avoiding direct processing of discrete coordinates in subsequent propagation calculations; the river side edge calculation node reads the event number, simulation time, river entry mileage, and river segment number to which the river entry point belongs from the event table, and reads the river segment row, tributary confluence row, gate control row, candidate interception row, and sensitive target row from the river data; The river segment row includes the river segment number, river segment centerline, river segment start point, river segment end point, river segment length, river segment flow velocity, and downstream flow rate. The tributary confluence row includes the confluence inlet number, confluence inlet mileage, upstream tributary number, and downstream river segment number. The gate control row includes the gate number, gate mileage, gate downstream river segment number, and gate flow rate. The candidate interception row includes the interception location number, interception mileage, applicable resource type, and deployment duration. The sensitive target row includes the sensitive target number, sensitive target type, and sensitive target mileage. Write the mileage of the confluence point in the tributary confluence row as the propagation bifurcation position, write the mileage of the gate in the gate control row as the flow control position, write the mileage of the interception in the candidate interception row as the candidate interception position, and write the mileage of the water intake, the mileage of the residential water supply section and the mileage of the aquaculture water intake section in the sensitive target row as the sensitive target position. Then, the river entry mileage, propagation bifurcation location, flow control location, candidate interception location, and sensitive target location are written into the river node table according to the river segment number and the distance from the center line; when the candidate interception location or sensitive target location falls at the connection point of two river segments, the downstream river segment is selected according to the connection order and written into the river node table; when the candidate interception location or sensitive target location does not fall within the projection range of the center line of any river segment, the corresponding location number is written into the node attachment failure record and is not involved in the generation of the river node table; In S2-2, this step converts the river node table into a recursive river hypergraph, ensuring that the downstream main channel, tributary inflows, and gate-controlled flow each have independent propagation edges. The river side edge calculation nodes read the river node table, arranging the river nodes from upstream to downstream according to the center-to-center distance within the same river segment number, and connecting adjacent river nodes to the main channel edge. The upstream and downstream river nodes of the main channel edge are taken from the upstream and downstream nodes of their respective adjacent river nodes. At the same propagation bifurcation point, the bifurcation point is used as the starting point for the branching, and the first river node in each downstream tributary is... Nodes serve as the endpoints of flow branching, generating flow branching hyperedges. These hyperedges are then split into multiple flow branching edges based on the number of downstream tributaries and written into the river hypergraph. At the same flow control location, the flow control location is used as the flow control start point, and the adjacent river node downstream of the flow control location is used as the flow control endpoint, generating flow control edges. When candidate interception locations or sensitive target locations are located inside the main river edge, the candidate interception locations or sensitive target locations are inserted as new river nodes, and the original main river edge is split into upstream and downstream edges, so that both candidate interception locations and sensitive target locations can serve as nodes for subsequent same-source comparison propagation and affected state calculation. The river hypergraph includes a river node table, main river edge, branch edge, flow control edge, branch hyperedge number, and river segment connection order; if there is no flow control position with a downstream adjacent river node, no flow control edge is generated, and the flow control position number is written into the flow control edge missing record; In S2-3, this step is used to generate a propagation table on the river hypergraph, so that each river edge has a propagation duration, pollution fraction, and arrival time. The river edge computing nodes read the inflow amount, inflow time, and metering method from the pollution event data, and read the edge length, flow velocity, and downstream flow of each river edge in the river hypergraph. When the flow velocity is greater than zero, the propagation duration is obtained by dividing the edge length by the flow velocity. When the flow velocity is equal to zero, the propagation duration of the river edge is recorded as a null value, and the propagation path from that river edge to the downstream is not written into the propagation table. The arrival time of the upstream river node of the starting river edge where the inflow point is located is the inflow time, and the arrival time of the downstream river node of each river edge is the arrival time of the upstream river node plus the propagation duration. For the same propagation bifurcation point, when the sum of all downstream flows is greater than zero, the contamination flow of the bifurcation edge is the contamination flow entering the propagation bifurcation point multiplied by the single downstream flow divided by the sum of all downstream flows. When the sum of all downstream flows is equal to zero, the contamination flow of each bifurcation edge is recorded as a null value and writing to the propagation table of the corresponding downstream tributary stops. The inflow rate is not recalculated for non-diversion sides. The pollution distribution rate of non-diversion sides between the inflow point and the first diversion point is the inflow rate. The pollution distribution rate of non-diversion sides after the first diversion point inherits the pollution distribution rate of the upstream river side. The river side number, upstream river node, downstream river node, propagation time, pollution distribution rate, arrival time, and measurement method are written into the propagation table. When the measurement method is mass, both the inflow rate and the pollution distribution rate are in kilograms; when the measurement method is volume, both are in cubic meters; and when the measurement method is concentration flux, both are in kilograms per second. Through the above processing, bifurcation, flow control, candidate interception, and sensitive targets in the river channel are transformed into a river hypergraph that can be recursively derived edge by edge. Pollution propagation duration and pollution splitting volume are written into the propagation table. Subsequently, S3 can generate successful interception propagation rows and failed interception propagation rows on the same river hypergraph and propagation table. This processing also avoids two common errors: non-splitting edges reusing inflow after splitting leads to amplified pollution; candidate interception locations or sensitive target locations falling inside river edges are not split, making them unsuitable as reference starting points or affected nodes. In practical applications: for example, inflow... The mileage is located at 320 meters in the R03 river section. There is a tributary confluence 500 meters downstream of the R03 river section. The tributary confluence connects the R04 main channel and the R11 tributary. The flow rate of R04 is 6 cubic meters per minute, and the flow rate of R11 is 4 cubic meters per minute. When the inflow is 100 kg, after propagation to the tributary confluence, the pollution diversion flow rate at the corresponding diversion side of R04 is 60 kg, and the pollution diversion flow rate at the corresponding diversion side of R11 is 40 kg. If there are subsequent candidate interception locations and water intakes at R04, they will be inserted into the river node table and used as calculation nodes for subsequent same-source comparison propagation.

[0019] S3. Based on the river hypergraph and propagation table, by setting the interception success status and interception failure status of the same candidate interception location, the same source comparison propagation is performed to obtain the pollution destination after interception failure, the affected status of sensitive targets, the remedial interception location and the remaining remedial time, and generate the failure table. This implementation method generates propagation results of successful and failed interception states at the same candidate interception location. The difference between these two propagation results determines the contamination destination, remedial interception location, and remaining remediation time after interception failure. The processing uses the arrival time of the upstream node of the same reference edge and the contamination flow rate as common inputs. It recursively calculates the successful path after deducting the interception amount and the failed path without deducting the contamination flow rate. Then, it uses the failed path to deduce the remediation space for S4 to generate the recovery scenario row. This implementation method includes the following steps: In S3-1, this step is used to establish a common-source reference input for the same candidate interception location, so that the interception success status and the interception failure status differ only in the interception amount deduction action; the riverside edge calculation node reads the river hypergraph, propagation table, and candidate interception location; when the candidate interception location has already been written into the river hypergraph as a river node in S2, the downstream adjacent river edge of the candidate interception location is recorded as the reference edge; when the candidate interception location is still located inside the river edge, the candidate interception location is used as a new river node, and the river edge where the candidate interception location is located is split into an upstream edge and a downstream edge, and the downstream edge is recorded as the reference edge; the arrival time of the upstream node of the reference edge and the pollution distribution flow of the reference edge are read, and the two are used as common-source inputs and written into the interception success propagation line and the interception failure propagation line, respectively; when the candidate interception location cannot be located to the river edge in the river hypergraph, the candidate interception location number is written into the reference edge missing record, and no interception success propagation line and interception failure propagation line are generated; In S3-2, this step is used to recursively deduce successful propagation and failure propagation along the same downstream connection sequence, so that the failure consequences are generated by the difference in pollution diversion flow reduction; the riverside edge calculation node reads the interception successful propagation line, the interception failure propagation line, and the interception amount in the resource data; the interception amount is obtained by multiplying the interception material quantity, the adsorption capacity or retention capacity of a single material in the resource data, and the deployment effectiveness coefficient. The deployment effectiveness coefficient is obtained from the equipment calibration record or emergency drill record corresponding to the resource type. If the equipment calibration record and emergency drill record are missing, the deployment effectiveness coefficient is taken as one; In the successfully intercepted propagation row, the downstream output of the control edge is obtained by subtracting the interception amount from the pollution distribution flow of the control edge. When the difference is less than zero, the downstream output is recorded as zero. In the failed intercepted propagation row, the pollution distribution flow of the control edge is used as the downstream output of the control edge. Then, the downstream connection order of the river hypergraph is recursively applied. When applying the recursion, the arrival time and pollution distribution flow of the previous river edge are used as the input of the next river edge. The arrival time of the next river edge is the arrival time of the previous river edge plus the propagation time of the next river edge. The pollution distribution flow of the next river edge is obtained according to the rule that non-diverting edges inherit the upstream pollution distribution flow and diverting edges are obtained according to the diversion ratio in the propagation table. Successful path table and failed path table are obtained respectively. In addition, when the propagation time along the riverbank is null, the propagation records from the riverbank to the downstream are not written into the successful path table and the failed path table. In S3-3, this step is used to identify the newly added contaminated destination after the interception failure by using the node difference between the successful path table and the failed path table, instead of just judging the contamination direction by the downstream traversal results. The riverside edge calculation node reads the successful path table and the failed path table, and merges the riverside edges in the two path tables according to the same downstream node. When the failed path table has a downstream node and the successful path table has the same downstream node, the destination difference is obtained by subtracting the successful pollution distribution from the failed pollution distribution, and the time compression is obtained by subtracting the failed arrival time from the successful arrival time. When the failed path table has a downstream node but the successful path table does not have the same downstream node, the successful pollution distribution is recorded as zero, the destination difference is recorded as the failed pollution distribution, and the time compression is recorded as a null value. When the successful path table has a downstream node but the failed path table does not have the same downstream node, no pollution destination record is generated. The downstream node with a destination difference greater than zero is recorded as the pollution destination after the interception failure, and the pollution destination, destination difference, time compression, and candidate interception position number that triggered the pollution destination are written into the pollution destination record. The destination difference and the pollution distribution use the same measurement method, and the time compression is in minutes. In S3-4, this step is used to filter out candidate interception positions that are still within the remediation window from the failure path and generate a failure table for S4 to construct the recovery scenario row; The riverside edge computing node reads the failure path table, pollution destination records, candidate interception positions and sensitive target positions in the river hypergraph. When the sensitive target position is located inside the riverside, a new river node is added based on the sensitive target position, and the corresponding riverside is split so that the sensitive target position can be used as the endpoint node in the failure path table. For riversides in the failure path table where the endpoint node is the sensitive target position and the pollution diversion flow is greater than zero, a sensitive target affected state is generated. The sensitive target affected state includes the sensitive target number, arrival time, pollution diversion flow, and candidate interception positions that trigger the arrival result. For the same pollution destination and the same sensitive target position, candidate interception positions upstream of the pollution destination are deleted, candidate interception positions downstream of the sensitive target position are deleted, and candidate interception positions whose arrival time is later than the arrival time of the sensitive target position are deleted. The remaining candidate interception positions are recorded as remedial interception positions, and the remaining remedial time is obtained by subtracting the arrival time of the remedial interception position from the arrival time of the sensitive target position. The candidate interception positions, pollution destination, destination difference, time compression, sensitive target affected state, remedial interception positions, and remedial remaining time are written into the failure table. When there is no remedial interception location, the failure table retains the candidate interception location, the contamination destination, and the affected status of sensitive targets, and records the remedial interception location and the remaining remedial time as null values; Through the above processing, each candidate interception position obtains a set of comparison propagation results with successful interception and failed interception states. The contamination destination is generated by the difference in contamination flow between the two path results. The remedial interception position is jointly defined by the contamination destination, the sensitive target position, and the arrival time. Subsequently, S4 can merge the failure table according to the same candidate interception position and generate a recovery scenario row. This processing avoids the error of generating remedial positions only according to the downstream nodes of the candidate interception position, because although the downstream node is located within the propagation path, if it is already located downstream of the sensitive target position or the arrival time is later than the arrival time of the sensitive target, it cannot be used as a remedial interception position. In practical applications, for example, the pollution flow rate of the control edge where candidate interception position C1 is located is 60 kg, the interception amount is 45 kg, the downstream output of the successfully intercepted propagation line is 15 kg, and the downstream output of the failed interception propagation line is 60 kg; if the arrival time of the failed path at the water intake W1 is 10:40, and the arrival time of candidate interception position C2, which is located downstream of the pollution destination and upstream of the water intake W1, is 10:28, then C2 is written as the remedial interception position, and the remaining remedial time is 12 minutes.

[0020] S4. Based on the failure table and resource data, K-adaptive recovery robust optimization is performed by binding the first-round interception position, the remedial interception position, the resource scheduling path and the deployment order to the same disposal column, so that the disposal column retains the remedial interception position and the remaining remedial duration after the interception fails, and a disposal table is generated. This implementation method generates a handling table based on a failure table, grouping the initial interception location with the remedial interception locations under each failure scenario to avoid selecting the initial location first and then temporarily searching for a remedial location after failure. The processing uses the same candidate interception location as the merging object, converting the contamination destination and the affected status of sensitive targets after failure into recovery scenario rows. Then, it combines resource transfer, remedial deployment, and remaining remedial time to generate recovery columns. Finally, it determines the target interception location according to the rule that all recovery scenarios can be covered by remediation. This implementation process includes the following steps: In S4-1, this step is used to organize the failure consequences under the same candidate interception position into a recovery scenario optimized by K-adaptive recovery robustness, so that each failure destination participates in the subsequent remediation constraints; the riverside edge computing node reads the candidate interception position, pollution destination, destination difference, affected status of sensitive targets, remediation interception position, and remaining remediation time from the failure table; the failure table rows with the same pollution destination and the same sensitive target number under the same candidate interception position are merged. During merging, the pollution flow is added together, the arrival time is taken as the earlier arrival time, and the remaining remediation time is taken as the remediation time with the shorter time value. The merged result is recorded as a recovery scenario row; the number of recovery scenario rows under the same candidate interception position is recorded as the K value; when the K value is zero, no disposal column is generated for the corresponding candidate interception position, and a record of no recovery scenario is written; when the K value is greater than zero, the candidate interception position, K value, and recovery scenario row are written to the recovery scenario group for S4-2 to read; In S4-2, this step is used to pre-bind remedial actions for each recovery scenario row, so that the first-round interception position and the remedial interception position are generated synchronously in the same disposal column; the riverside edge computing node reads the recovery scenario group and resource data, and records the same candidate interception position as the first-round interception position; for each recovery scenario row, the remedial interception position in the failure table is read, and the current resource position, resource number, resource activation status, road or patrol channel network, travel speed, first-round deployment duration and remedial deployment duration in the resource data are read; when the resource activation status is not activated, the resource number does not participate in the generation of the remedial column; when the resource activation status is activated, the path from the current resource position to the first-round interception position and from the first-round interception position to the remedial interception position is calculated according to the road or patrol channel network. The travel time of the road segment is the road segment length divided by the travel speed. When multiple paths exist, the path with the lower sum of travel times is taken. If the sum of travel times is the same, the path with fewer road segments is taken. The travel time of a resource from the initial interception position to the remedial interception position is recorded as the transition time, and the remedial completion time is the sum of the transition time and the remedial deployment time. The recovery scene row, remedial interception position, resource number, resource scheduling path, transition time, remedial deployment time, and remedial completion time are written into the remedial column. When the remedial interception position is empty, no corresponding remedial column is generated, and the remedial missing record is retained in the recovery scene group. In S4-3, this step is used to calculate whether each remedial column is still within the remedial window and to calculate the amount of pollution that will still propagate downstream after remediation. The riverside edge computing node reads the K remedial columns and corresponding recovery scenario rows under the same first-round interception position. For each remedial column, the remaining time of the scenario is obtained by subtracting the remedial completion time from the remaining remedial time in the recovery scenario row. When the remaining time of the scenario is zero, it means that the remedial deployment completion time is the same as the arrival time of the sensitive target and the corresponding remedial column is retained. When the remaining time of the scenario is negative, the corresponding remedial column is marked as timeout. The interception amount corresponding to the remedial interception position in the resource data is read. The remaining pollution amount after remediation is obtained by subtracting the pollution diversion amount corresponding to the remedial interception position from the pollution diversion amount in the recovery scenario row. When the difference is less than zero, the remaining pollution amount after remediation is recorded as zero. The first-round interception position, recovery scenario row, remedial interception position, resource scheduling path, scenario remaining time, and remaining pollution amount after remediation are written into the recovery column. When multiple recovery columns correspond to the same remedial interception position, only the recovery column with the shorter resource scheduling path travel time is retained. If the travel time is the same, the recovery column with the shorter scenario remaining time is retained. In S4-4, this step is used to group and filter the K recovery columns at the first-round interception position level, so that the selected target interception positions have remedial actions in all recovery scenarios; the riverside edge computing nodes are merged into recovery columns according to the same first-round interception position. If any recovery scenario row does not have a recovery column, or the scenario remaining time of any recovery column is negative, the corresponding first-round interception position is deleted; among the retained first-round interception positions, the sum of the remaining pollution after remediation and the sum of the scenario remaining time of the K recovery columns are calculated respectively, and arranged in ascending order of the sum of the remaining pollution after remediation; when the sum of the remaining pollution after remediation is the same, they are arranged in descending order of the sum of the scenario remaining time; when the sum of the remaining pollution after remediation and the sum of the scenario remaining time are both the same, they are arranged in ascending order of the sum of the resource scheduling path travel time. The first-round interception position ranked first is recorded as the target interception position; in the deployment order, the target interception position is ranked before the remedial interception positions in the K recovery columns, and the remedial interception positions in the K recovery columns are arranged from shortest to longest remaining time in the scenario; the target interception position, K recovery columns, resource scheduling path, and deployment order are written into the disposal table; if there is no reserved first-round interception position, no disposal column record is output; Through the above processing, K-adaptive recovery robust optimization does not simply select the interception position that is close or fast to reach. Instead, it requires that the same target interception position has pre-set remedial actions in all K failure paths, and the disposal table is determined by the remaining pollution after remediation, the scene time remaining and the resource scheduling path. This processing avoids the situation where the first-round interception position performs well in a single scene but has no remedial space in other failure paths. In practical applications, the following example illustrates the situation: For instance, under candidate interception location C1, there are two recovery scenario rows, corresponding to tributary inlet B1 and water intake W1 respectively, with a K value of 2; it takes 8 minutes for resources to be transferred from C1 to the remedial interception location C2 and deployed, and the remaining remediation time for B1 is 10 minutes, so the scenario time remaining is 2 minutes; it takes 14 minutes for resources to be transferred from C1 to the remedial interception location C3 and deployed, and the remaining remediation time for W1 is 12 minutes, so C1 has a timeout recovery column and is deleted; if both recovery columns under candidate interception location C4 are retained and the sum of the remaining pollution after remediation is low, then C4 is written as the target interception location.

[0021] S5. Based on the disposal table, write back the first round of interception locations to the river supermap and update the flow distribution. Re-execute the same-source comparison propagation. When the updated sensitive target affected status and the remaining remediation time meet the disposal table, output the emergency disposal decision results. This implementation method is used to write back and verify the target interception locations in the treatment table, so that the target interception locations selected in S4 are not only kept in the sorting results, but are re-entered into the pollution propagation calculation of the river super-map to verify the actual impact of the target interception locations on pollution diversion, remedial interception locations, and remaining remedial time. The process first writes the interception volume of the target interception location back to the propagation table, then re-progresses the successful verification path and the failed verification path downstream from the target interception location, and then compares K recovery columns one by one. After the verification conditions are met, the emergency treatment decision result is output. This implementation process includes the following steps: In S5-1, this step is used to return the target interception location from the treatment table to the river hypergraph, so that the interception action at the target interception location changes the subsequent pollution distribution. The riverside edge calculation node reads the target interception location, interception amount, and K recovery columns from the treatment table, and reads the riverside number, upstream riverside node, downstream riverside node, and pollution distribution of the riverside where the target interception location is located from the propagation table. When the target interception location is located inside the riverside, the riverside where the target interception location is located is split into the upstream target edge and the downstream target edge, with the target interception location as a new riverside node. The downstream edge of the target is used as the write-back edge; when the target interception position already exists as a river node, the downstream adjacent river edge of the target interception position is used as the write-back edge; the remaining flow rate is obtained by subtracting the interception amount corresponding to the target interception position from the pollution flow rate of the write-back edge in the propagation table, and the remaining flow rate is recorded as zero when the difference is less than zero; the remaining flow rate is written back to the write-back edge, and the propagation time and downstream connection relationship of the write-back edge are retained, and a write-back propagation table is generated; when the target interception position cannot be located to the river edge in the river hypergraph, no write-back propagation table is generated, and a record of write-back failure of the target interception position is output. In S5-2, this step is used to perform the same-source comparison propagation again based on the write-back propagation table to determine whether the downstream propagation results after the target interception position is written back still support the recovery column in the disposal table; the river side edge calculation node reads the write-back propagation table and the river hypergraph, takes the write-back edge as the starting edge, and reads the arrival time of the upstream node of the write-back edge and the remaining flow rate. In the case of successful interception, the remaining flow rate is used as the downstream output of the write-back edge, and the arrival time and pollution flow rate are recursively calculated along the downstream connection sequence of the river hypergraph to generate a successful verification path table. In the state of interception failure, the contaminated flow of the write-back edge in the pre-write-back propagation table is used as the downstream output of the write-back edge, and the arrival time and contaminated flow are recursively deduced along the same downstream connection order to generate the failure verification path table. When iterating along each side, the arrival time of each subsequent riverside is the arrival time of the upstream river node plus the riverside propagation time. The pollution distribution of non-diversion sides inherits the pollution distribution of upstream riversides, and the pollution distribution of diversion sides is obtained according to the corresponding diversion ratio in the propagation table. When the riverside propagation time is null, the propagation record from the riverside to the downstream is not written into the successful verification path table and the failed verification path table. In S5-3, this step is used to compare the verification propagation results with the K recovery columns in the treatment table item by item, and to re-obtain the remedial interception location, the remaining remedial duration, and the remaining pollution amount after remediation. The riverside edge calculation node reads the failure verification path table, the K recovery columns in the treatment table, the candidate interception locations and sensitive target locations in the river supermap. For each recovery column, the pollution destination and sensitive target number in the recovery column are read first, and then the candidate interception location located downstream of the pollution destination and upstream or at the same position as the sensitive target location is found in the failure verification path table. Candidate interception locations whose arrival time is later than the arrival time of the sensitive target are deleted. When there are multiple remaining candidate interception locations, the remaining pollution amount after remediation for each candidate interception location is calculated first. The remaining pollution amount after remediation is the pollution diversion flow of the candidate interception location minus the remedial interception amount. If the difference is less than zero, it is recorded as zero. Then, one item is selected as the verification remedial interception location according to the remaining pollution amount after remediation from low to high, the remaining remedial duration from high to low, and the resource scheduling path travel time from low to high. The remaining time for review and remediation is the arrival time of the sensitive target minus the arrival time of the review and remediation interception location. The remaining amount of pollution after review and remediation is the pollution diversion flow of the review and remediation interception location minus the remediation interception amount. The review and remediation interception location, the remaining time for review and remediation, and the remaining amount of pollution after review and remediation are written into the review column. If the review and remediation interception location is not obtained, the corresponding recovery column is written into the review failure record. In S5-4, this step is used to determine whether the target interception location still meets the K recovery columns after the write-back, and to form the final emergency response decision result when the review is passed; the riverside edge computing node reads the K recovery columns in the response table and the review columns generated in S5-3, and compares the remedial interception location, remedial completion time, and remaining pollution after remediation in the recovery column with the remedial interception location, remedial remaining time, and remedial remaining pollution after remediation in the review column; when all K recovery columns have the same remedial interception location, and the remedial remaining time in each review column is not less than the remedial completion time in the corresponding recovery column, and the remedial remaining pollution after remediation is not greater than the remedial remaining pollution after remediation in the corresponding recovery column, the target interception location, the remedial interception location in the K recovery columns, the resource scheduling path, the deployment order, the remedial remaining time, and the remedial remaining pollution after remediation are written into the emergency response decision result; If any recovery column does not meet the above comparison relationship, delete the corresponding handling column of the current target interception position, return to S4-4 to read the next round interception position according to the handling table as the new target interception position, and re-execute S5; if there is no next round interception position in the handling table, output no available handling column record; Through the above processing, the disposal table generated by S4 is re-submitted into the river supergraph propagation process. The interception amount at the target interception location, the downstream pollution diversion flow, the remedial interception location, and the remaining remedial duration are all verified and calculated before being output. This processing avoids the situation where the target interception location satisfies K recovery columns in the S4 sorting, but the disposal table becomes invalid after writing back due to changes in diversion flow or remedial location. In practical applications: For example, if the disposal table identifies C4 as the target interception location, the pollution diversion flow along the riverbank where C4 is located is 80 kg, the interception amount of C4 is 50 kg, and the remaining diversion flow after writing back is 30 kg; the system re-infers the failure verification path from downstream of C4. If the remedial interception locations C6 and C8 in the K recovery columns are still upstream of the sensitive target, and the recalculated remedial remaining time is 9 minutes and 13 minutes respectively, both not less than the corresponding remedial completion time, and the remaining pollution amount after remediation is not higher than the value recorded in the disposal table, then the emergency disposal decision result containing C4, C6, C8, resource scheduling path, and deployment order will be output.

[0022] Working principle: This scheme first places the pollution event, river structure, and emergency resources on the same river mileage line, so that the entry point, tributary outlet, gate, interception location, sensitive target, and resource location can all participate in the calculation according to the upstream and downstream relationship; then, a river hypergraph is constructed based on the river connectivity relationship to calculate the arrival time and pollution fraction of the polluted water mass along the main channel, tributaries, and flow control locations; based on this, two propagation results are calculated for the same candidate interception location: successful interception and interception failure. The difference between the two results is used to determine where the pollution will flow after interception failure, whether it will affect sensitive targets, and how much pollution remains downstream. Minimize remedial time; then, filter the initial interception, remedial interception, resource scheduling path, and deployment priority in the same disposal column, so that the selected target interception location can not only intercept at present, but also has remedial arrangements after failure; finally, write the target interception location back to the river supermap to recalculate the propagation results, and after confirming that the remedial location, remedial time, and remaining pollution amount still meet the disposal requirements, output the emergency disposal decision results; overall, the previous step does not give results alone, but provides a basis for judgment for the next step, forming a continuous decision-making process from pollution propagation to failure consequences, and then to disposal selection and back-writing verification; When abnormal discharge of stormwater into a river occurs at a stormwater inlet in a chemical industrial park, the system first locates the stormwater inlet to a specific river section mileage and converts the locations of oil booms, absorbent materials, and emergency teams into resource mileage. As the polluted water mass travels downstream and passes through tributary outlets, sluice gates, and water intakes, the system calculates the arrival time and pollution distribution flow along each path. For a nearby interception location, the system does not simply select it because of its proximity; instead, it simultaneously calculates the remaining downstream pollution volume if the interception is successful, and whether the pollution will enter tributaries, water intakes, or residential water supply sections if the interception fails. If there is no time to remedy the situation downstream after the nearby location fails, the system selects an interception location where there is still a window for remedy after the failure, and simultaneously provides the remedy interception location, resource scheduling path, and deployment priority. This ensures that on-site personnel know the subsequent remedy points and resource directions when deploying the first round of interception, avoiding the need to temporarily chase the polluted water mass after the interception fails.

[0023] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An intelligent decision-making method for emergency disposal of sudden environmental pollution incidents, characterized in that, include: S1. The riverside edge computing node acquires pollution event data, river data, and resource data, performs time alignment and spatial mapping according to the same event number and the same simulation time, and generates an event table. S2. Based on the event table, construct a river hypergraph by including the propagation bifurcation points, flow control points, candidate interception points, and sensitive targets in the river channel, and calculate the propagation duration and flow rate of the polluted water mass in the river hypergraph according to the pollution event data to generate a propagation table. S3. Based on the river hypergraph and propagation table, by setting the interception success status and interception failure status of the same candidate interception location, the same source comparison propagation is performed to obtain the pollution destination after interception failure, the affected status of sensitive targets, the remedial interception location and the remaining remedial time, and generate the failure table. S4. Based on the failure table and resource data, K-adaptive recovery robust optimization is performed by binding the first-round interception position, the remedial interception position, the resource scheduling path and the deployment order to the same disposal column, so that the disposal column retains the remedial interception position and the remaining remedial duration after the interception fails, and a disposal table is generated. S5. Based on the disposal table, write the first round of interception locations back to the river supermap and update the flow distribution. Re-execute the same-source comparison propagation. When the updated sensitive target affected status and the remaining remediation time meet the disposal table, output the emergency disposal decision results.

2. The intelligent decision-making method for emergency response to sudden environmental pollution incidents according to claim 1, characterized in that: S1 includes: S1-1. Read the river entry point coordinates, river entry time and pollutant identifier from the pollution event data. Read the river segment number, river segment start coordinates, river segment end coordinates and river segment connection order from the river channel data. Project the river entry point coordinates onto the nearest river segment centerline. Generate the river entry mileage position according to the centerline distance from the projection point to the river segment start point. Output the event baseline row. S1-2. Read the resource number, resource location and resource activation status from the resource data, project the resource location onto the river segment centerline in the river data, generate the resource mileage position according to the river segment number where the projection point is located and the distance from the centerline, and write the resource number, resource mileage position and resource activation status into the event base line to generate the resource attachment line. S1-3. Using the river entry time as the first simulation time, bind the event base row and resource connection row to the same event number in ascending order of adjacent simulation times, and generate an event table including event number, simulation time, river entry mileage position, river segment connection order, resource number, resource mileage position, and resource activation status.

3. The intelligent decision-making method for emergency response to sudden environmental pollution incidents according to claim 2, characterized in that: S2 includes: S2-1. Read the river inlet mileage position in the event table, and read the river segment row, tributary confluence row, gate control row, candidate interception row, and sensitive target row in the river data. Determine the confluence mileage position in the tributary confluence row as the propagation bifurcation position, the gate mileage position in the gate control row as the flow control position, the interception mileage position in the candidate interception row as the candidate interception position, and the water intake mileage position, residential water supply section mileage position, and aquaculture water intake section mileage position in the sensitive target row as the sensitive target position. Generate a river node table by assigning the river inlet mileage position, propagation bifurcation position, flow control position, candidate interception position, and sensitive target position to the river segment number and center-of-line distance.

4. The intelligent decision-making method for emergency response to sudden environmental pollution incidents according to claim 3, characterized in that: S2 further includes: S2-2. Within the same river segment number, sort the river nodes in the river node table from upstream to downstream according to the distance from the center line, and connect adjacent river nodes to the main river edge; at the same propagation bifurcation position, connect the propagation bifurcation position to the first river node of each downstream tributary and generate a branching edge; at the same flow control position, connect the flow control position to the adjacent downstream river node and generate a flow control edge; generate a river hypergraph from the river node table, main river edge, branching edge, and flow control edge; S2-3. Read the inflow volume and inflow time from the pollution event data. Read the side length, flow velocity, and downstream flow of each riverbank in the river data. Divide the side length by the flow velocity to obtain the propagation time. Multiply the inflow volume by the proportion of the single downstream flow at the same propagation bifurcation point to the sum of all downstream flows to obtain the pollution distribution volume of the bifurcation side. Use the inflow volume as the pollution distribution volume of the non-bifurcation side. Write the riverbank number, upstream river node, downstream river node, propagation time, pollution distribution volume, and arrival time into the propagation table.

5. The intelligent decision-making method for emergency response to sudden environmental pollution incidents according to claim 4, characterized in that: S3 includes: S3-1. Read the river hypergraph, propagation table and candidate interception locations. Record the river edge where the candidate interception location is located as the control edge. Use the arrival time of the upstream node of the control edge and the pollution distribution flow of the control edge as the same source input to generate the successful interception propagation line and the failed interception propagation line respectively. S3-2. In the successful propagation row of interception, the downstream output of the control edge is recorded as the remaining amount after subtracting the interception amount from the pollution distribution flow of the control edge, and is recursively applied to each edge along the downstream connection sequence of the river supergraph; in the failed propagation row of interception, the downstream output of the control edge is recorded as the pollution distribution flow of the control edge, and is recursively applied to each edge along the same downstream connection sequence to obtain the successful path table and the failed path table.

6. The intelligent decision-making method for emergency response to sudden environmental pollution incidents according to claim 5, characterized in that: S3 further includes: S3-3. For riverbanks with the same downstream node in the successful path table and the failed path table, calculate the destination difference by subtracting the successful pollution distribution from the failed pollution distribution based on the same downstream node, and calculate the time compression by subtracting the failed arrival time from the successful arrival time. Record the downstream node with a destination difference greater than zero as the destination of the pollution after the interception failure. S3-4. Read the candidate interception positions and sensitive target positions in the failure path table, delete the candidate interception positions whose arrival time is later than the arrival time of the sensitive target position, record the remaining candidate interception positions as the remedial interception positions, and obtain the remaining remedial time by subtracting the arrival time of the remedial interception position from the arrival time of the sensitive target position. Write the candidate interception position, contamination destination, destination difference, time compression amount, affected status of the sensitive target, remedial interception position, and remaining remedial time into the failure table.

7. The intelligent decision-making method for emergency response to sudden environmental pollution incidents according to claim 6, characterized in that: S4 includes: S4-1. Read the failure table, and record the failure table rows with different contamination destinations or different sensitive target affected states under the same candidate interception position as recovery scenario rows, and record the number of recovery scenario rows under the same candidate interception position as the K value. S4-2. Read resource data, record the same candidate interception position as the first round interception position, and generate a remediation column for each recovery scenario row; each remediation column includes the remediation interception position, resource scheduling path, resource transition time from the first round interception position to the remediation interception position, remediation deployment time, and remediation completion time.

8. The intelligent decision-making method for emergency response to sudden environmental pollution incidents according to claim 7, characterized in that: S4 further includes: S4-3. For the K remedial columns under the same first-round interception position, subtract the remedial completion time from the remaining remedial time to obtain the scene remaining time, and subtract the interception amount corresponding to the remedial interception position from the pollution distribution flow in the recovery scene row to obtain the remaining pollution amount after remediation, and generate the recovery column. S4-4. Merge K recovery columns according to the same first-round interception position, delete the first-round interception positions with negative scene remaining time, and determine the target interception position according to the sum of the remaining pollution after remediation in the K recovery columns from low to high and the sum of scene remaining time from high to low. Write the target interception position, K recovery columns, resource scheduling path and deployment order into the disposal table.

9. The intelligent decision-making method for emergency response to sudden environmental pollution incidents according to claim 8, characterized in that: S5 includes: S5-1. Read the target interception location, interception amount and K recovery columns in the treatment table, locate the riverbank where the target interception location is located, and subtract the interception amount from the pollution distribution flow of the riverbank where the target interception location is located in the propagation table to obtain the remaining distribution flow. Write the remaining distribution flow back to the riverbank where the target interception location is located to generate the write-back propagation table. S5-2. Based on the write-back propagation table, starting from the riverbank where the target interception location is located, the arrival time and pollution distribution flow are recursively calculated along the downstream connection sequence of the river supergraph. Successful verification path tables and failed verification path tables are generated according to the interception success status and interception failure status, respectively.

10. The intelligent decision-making method for emergency response to sudden environmental pollution incidents according to claim 9, characterized in that: S5 includes: S5-3. Read the K recovery columns in the failure review path table and the disposal table, and compare the remediation interception location, remediation remaining time, and remediation remaining pollution amount in each recovery column; Among them, the remedial interception position is obtained from the candidate interception position in the failure review path table that is downstream of the pollution destination and arrives earlier than the sensitive target. The remaining remedial time is the arrival time of the sensitive target minus the arrival time of the remedial interception position. The remaining pollution amount after remediation is the pollution diversion flow of the remedial interception position minus the remedial interception amount. S5-4. When the same remedial interception position exists in all K recovery columns, and the recalculated remaining remedial time is not less than the remedial completion time in the corresponding recovery column, and the recalculated remaining pollution after remediation is not greater than the remaining pollution after remediation in the corresponding recovery column, the target interception position, the remedial interception position in the K recovery columns, the resource scheduling path, and the deployment order are written into the emergency response decision results.

Images