A method and system for protecting vehicles from urban flooding based on precipitation forecasting

Abstract

This invention relates to the field of urban flood control technology, specifically to a method and system for urban flood vehicle risk protection based on precipitation forecasting. The method includes acquiring meteorological forecast data and real-time water level data; calculating the flood risk level based on the meteorological forecast data, and performing fusion analysis and dual verification with the real-time water level data; generating a flood risk warning when triggering conditions are met; responding to the warning by simultaneously triggering associated insurance protection services and generating insurance compensation commitment information; identifying affected vehicle users and sending them notifications containing the warning and protection information. This invention deeply integrates very early precipitation forecasting with insurance protection services. While generating a flood risk warning, it simultaneously triggers insurance protection services and generates clear insurance compensation commitment information, providing predictive warnings with accompanying insurance compensation commitments to vehicle users before flooding occurs, enabling them to make the safest decisions based on clear economic protection.

Description

Technical Field

[0001] This invention relates to the field of urban flood prevention technology, specifically to a method and system for protecting vehicles from urban flooding risks based on precipitation forecasting. Background Technology

[0002] Current urban flooding vehicle damage prevention and control faces two main challenges: first, delayed warnings, often leaving car owners with no time to move their vehicles by the time they receive a warning; second, a dilemma in risk decision-making, where car owners face the choice between "protecting the car" and "protecting themselves," lacking authoritative decision-making support and financial security; and third, even if vehicles are damaged, the insurance claims process is cumbersome and fails to alleviate the anxiety of car owners during risk decision-making. Existing technologies either only provide weather warnings or only offer post-accident insurance services, failing to effectively link "advanced prediction" and "risk protection" at critical decision-making moments. Car owners face the dilemma of "whether to risk moving their vehicles," lacking an authoritative and financially secure decision-making basis.

[0003] Existing technical solutions have the following significant drawbacks: Disconnect between early warning and protection: Weather warning systems and insurance claims services operate independently, failing to provide car owners with an integrated solution at critical decision-making moments. Delayed intervention: Real-time water accumulation-based warning models do not allow sufficient response time for car owners. Lack of decision support: Existing solutions only provide risk warnings, failing to address car owners' concerns about financial losses, leading some to risk moving their vehicles, increasing safety risks. Summary of the Invention

[0004] The purpose of this invention is to overcome the above-mentioned shortcomings and provide a method and system for protecting vehicles from urban flooding based on precipitation prediction.

[0005] To achieve the above objectives, the specific solution of the present invention is as follows: This invention provides a method for protecting vehicles from urban flooding based on precipitation forecasting, comprising the following steps: S1. Obtain meteorological forecast data for the target area within a preset time period, as well as real-time water level data for each monitoring point within the target area; S2. Calculate the waterlogging risk level based on the meteorological forecast data; perform fusion analysis and double verification based on the waterlogging risk level and the real-time water level data to determine whether the preset warning triggering conditions are met. If the preset warning triggering conditions are met, generate a waterlogging risk warning. S3. In response to the waterlogging risk warning, simultaneously trigger the insurance protection service associated with the waterlogging risk warning and generate insurance compensation commitment information for the waterlogging risk warning; S4. Identify vehicle users within the area affected by the flood risk warning and push a warning notification containing the flood risk warning information and the insurance compensation commitment information to the vehicle users.

[0006] Preferably, in step S2 of this invention, calculating the urban flooding risk level based on the meteorological forecast data includes: Based on the meteorological forecast data and geographical environment data, the risk index R of each location in the target area is calculated using an urban flooding risk prediction model. The risk index R is mapped to the corresponding urban flooding risk level.

[0007] Preferably, the risk index R is calculated using the following multi-factor weighted model: Where P is the predicted precipitation factor, D is the real-time water level and depth factor, T is the topographic index factor, and V is the regional vulnerability index factor. to The dynamic weighting coefficients for each factor are: the topography index factor is calculated based on digital elevation model data; the regional vulnerability index factor is calculated based on historical flooding event data and vehicle density data; and the dynamic weighting coefficients are adjusted according to seasonal and / or urban regional characteristics.

[0008] Preferably, in step S2, the method of determining whether a preset warning triggering condition is met through dual verification is used. The condition is considered met if one of the following conditions is met: First triggering condition: If the waterlogging risk level reaches or exceeds a preset risk threshold, and any of the real-time water level data reaches or exceeds the first preset water level threshold, then the posterior probability of the waterlogging event is calculated based on a Bayesian inference model; when the posterior probability exceeds a preset probability threshold, it is determined that the warning triggering condition is met. Second triggering condition: If any real-time water level data reaches or exceeds the second preset water level threshold, it is directly determined that the warning triggering condition is met; wherein, the second preset water level threshold is greater than the first preset water level threshold.

[0009] Preferably, in this invention, generating an urban flooding risk warning includes defining a dynamic warning range, wherein the dynamic warning range includes: Starting from the monitoring point where the water level data triggers the warning, based on the hydrodynamic model, combined with the digital elevation model, land cover type and drainage network data, the process of water accumulation diffusion is simulated, and the geographical influence range where the water depth exceeds the preset depth threshold is dynamically delineated.

[0010] Preferably, in this invention, the process of acquiring the real-time water level data in step S1 includes an intelligent power consumption management step: Based on the aforementioned waterlogging risk level and / or early warning trigger status, dynamically and remotely configure the heartbeat reporting interval of the water level monitoring network; When in a low-risk or no-warning state, configure to long-interval mode to reduce power consumption; When the risk level rises or enters an early warning state, configure it to short interval mode to increase the monitoring frequency; The water level monitoring network supports a dual-mode communication mechanism. In the long interval mode, if the real-time water level data of any monitoring point exceeds the locally stored emergency threshold, it will be immediately reported through the emergency alarm channel.

[0011] Preferably, in step S3 of this invention, triggering the insurance coverage service includes: The event information of the flood risk warning and the information of the affected insurance policies are stored on the blockchain for evidence; based on the machine learning model, the compensation amount is estimated according to the water level depth, regional vehicle density and average vehicle value, and a time-sensitive digital pre-authorization token is generated as a fast claims settlement certificate.

[0012] Preferably, in step S4 of this invention, pushing a warning notification to the vehicle user includes: The system acquires the geographical location information of insured vehicle users in real time and performs semantic recognition on the geographical location to distinguish whether the user is located in an underground garage, a low-lying road section, or a high-lying parking lot. Users are classified into individual risk levels based on their semantic location type and their distance from the flood risk center; Based on the individual risk level of the user, different combinations of notification channels, notification templates, and content urgency are automatically matched for differentiated push notifications.

[0013] Preferably, step S4 of this invention further includes a step to ensure that the upgrade notification is delivered: For high-risk users who do not confirm receipt of the push notification within the preset time, the notification will be upgraded in a tiered manner according to the priority of the notification channel, until human customer service intervention is required.

[0014] In another aspect, the present invention provides an urban flooding vehicle risk protection system based on precipitation prediction, characterized in that it includes: The multi-source data acquisition module is used to acquire meteorological forecast data for the target area within a preset time period, as well as real-time water level data of each monitoring point within the target area. The data fusion and decision-making module is used to calculate the waterlogging risk level based on the meteorological forecast data; perform fusion analysis and dual verification based on the waterlogging risk level and the real-time water level data to determine whether the preset warning triggering conditions are met; if the preset warning triggering conditions are met, a waterlogging risk warning is generated. The insurance linkage service module is used to respond to the waterlogging risk warning, simultaneously trigger the insurance protection service associated with the waterlogging risk warning, and generate insurance compensation commitment information for the waterlogging risk warning; The multi-channel notification module is used to identify vehicle users in areas affected by the flood risk warning and push warning notifications containing the flood risk warning information and the insurance compensation commitment information to the vehicle users.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention deeply integrates early rainfall forecasting with insurance protection services. While generating flood risk warnings, it simultaneously triggers insurance protection services and generates clear insurance compensation commitment information. It provides vehicle users with predictive warnings and insurance compensation commitments before flooding occurs, enabling vehicle users to make the safest personnel decisions based on clear economic protection.

[0016] Furthermore, by defining a dynamic early warning range, this invention accurately targets vehicle users affected by the risk of urban flooding, avoiding the waste of resources and user annoyance caused by the traditional "broad-based" early warning approach. This increases users' attention to and responsiveness to the early warning, thereby improving the overall efficiency of urban flooding risk prevention and control. Attached Figure Description

[0017] Figure 1 This is a flowchart of a method for protecting vehicles from urban flooding based on precipitation prediction, provided in an embodiment of the present invention. Figure 2 This is a judgment logic diagram of a dual verification mechanism provided in an embodiment of the present invention. Detailed Implementation

[0018] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments, but this is not to limit the scope of the invention to this.

[0019] This invention discloses a method and system for urban flooding vehicle risk protection based on precipitation forecasting. The core innovation lies in the deep integration of very early precipitation forecasting with insurance coverage services. Through multi-source data collaborative collection, intelligent risk decision-making, insurance linkage activation, and differentiated early warning push, it provides vehicle users with a basis for decision-making regarding urban flooding risks with economic protection, addressing the pain points of existing technologies such as delayed early warnings, disconnected protection, and lack of decision support. The following detailed description of the technical solution of this invention, combined with specific implementation scenarios, ensures that those skilled in the art can fully reproduce the technical effects of this invention.

[0020] Example 1:

[0021] Taking the flood season (June-September) urban flood control in a rainy provincial capital city (hereinafter referred to as "target city") as an application scenario, this embodiment details the entire process of implementation. Figure 1 and Figure 2 As shown in this embodiment, a method for protecting vehicles from urban flooding based on precipitation forecasting may specifically include the following steps: Step S1: Obtain meteorological forecast data for the target area within a preset future time period, as well as real-time water level data for each monitoring point within the target area. Specifically, this step includes the following steps: Step S11: Employing a multi-source authoritative meteorological data access strategy, a stable data link is established with the China Meteorological Administration's Meteorological Data Center and the European Centre for Medium-Range Weather Forecasts (ECMWF) via a RESTful API interface. The "future preset time period" is set to 6 hours, which can be adjusted to 3-12 hours according to city needs. Quantitative Precipitation Forecast (QPF) data with a spatial resolution of 1km×1km for the target city is obtained. The data update frequency is once every 5 minutes to ensure the timeliness of the forecast data.

[0022] Quality control (QC) operations are performed on the accessed meteorological data. Abnormal precipitation data (such as single predicted precipitation exceeding 1.5 times the historical extreme value for the same period in the region) are removed by the 3σ principle. A small number of missing data are filled by linear interpolation. Then, spatial interpolation is performed by Kriging interpolation to eliminate the spatial resolution differences between different meteorological sources, forming standardized and fully covered gridded prediction data, which provides a unified input for subsequent risk assessment.

[0023] Step S12: Real-time water level data acquisition: Deploy 256 water level monitoring points in key areas of the target city (including low-lying road sections, underground parking garage entrances and exits, historical flooding sites, etc.) to form a heterogeneous sensor network. The heterogeneous sensor network includes point-triggered sensors and ultrasonic / pressure-type precision ranging sensors. For example, the heterogeneous sensor network may include 192 point-triggered sensors (installed at heights of 5cm, 15cm, and 30cm respectively) and 64 ultrasonic precision ranging sensors (measurement accuracy ±1cm). Point-triggered sensors are used for threshold alarms; ultrasonic precision ranging sensors are used for continuous monitoring. The sensor communication protocol adopts dual-mode adaptation of NB-IoT (60% coverage, suitable for low-power, wide-coverage scenarios) and 4G (40% coverage, suitable for high-frequency reporting scenarios); the power supply method is "solar panel + lithium battery backup", with a lithium battery capacity of 12V / 20Ah, ensuring continuous operation for more than 72 hours under no sunlight conditions.

[0024] The sensor can employ an adaptive sampling mechanism: during non-warning periods, it uses a low-power mode and samples once every 10 minutes; when rainfall is predicted, it automatically switches to a high-frequency mode and samples once every 1 minute.

[0025] Step S2: Calculate the urban flooding risk level based on the meteorological forecast data; perform fusion analysis and dual verification based on the urban flooding risk level and the real-time water level data to determine whether the preset warning triggering conditions are met. If the preset warning triggering conditions are met, an urban flooding risk warning is generated. Specifically, this step includes the following steps: Step S21, Calculation of Urban Flooding Risk Level: First, collect basic geographic environmental data for the target city, including: 1) 1:5000 high-precision digital elevation model (DEM) data, used to extract topographic relief features; 2) Land cover type data (divided into 5 categories: asphalt pavement, cement pavement, grassland, green space, and building areas); 3) Drainage network distribution and design flow data (obtained from the city's water affairs department); 4) Database of historical urban flooding events in the past 5 years (including occurrence time, location, water depth, and affected area); 5) Urban area vehicle density heat map data (generated based on statistics from the transportation department).

[0026] The risk index R is calculated using a multi-factor weighted model: The factors and their weights are set as follows: P represents the predicted precipitation factor: the predicted precipitation for the next 6 hours is mapped to a score of 0-10, where ≤10mm is 2 points, 10-30mm is 4 points, 30-50mm is 7 points, and >50mm is 10 points; D is the real-time water level depth factor: based on real-time monitoring data from sensors, ≤3cm is 1 point, 3-8cm is 4 points, 8-15cm is 7 points, and >15cm is 10 points; T represents the topographic index factor: slope and elevation are calculated based on DEM data. Low-lying areas (elevation more than 2m below the surrounding average elevation) are scored 10 points, flat areas (slope < 3°) are scored 6 points, and high-lying areas (elevation more than 3m above the surrounding average elevation) are scored 3 points. V represents the regional vulnerability index factor: calculated using a weighted summation method, with a weight of 0.6 for the frequency of historical flooding events (10 points for 3 or more occurrences in the past 5 years, 6 points for 1-2 occurrences, and 2 points for 0 occurrences) and a weight of 0.4 for vehicle density (10 points for ≥300 vehicles / km², 6 points for 150-300 vehicles / km², and 2 points for <150 vehicles / km²). - Dynamic weighting coefficients: adjusted according to season and / or urban area characteristics, such as emphasizing precipitation factors during the flood season and emphasizing topographic factors during the non-flood season. For example, the flood season (June-September) is set to... =0.35、 =0.3、 =0.2、 =0.15 (focusing on precipitation and real-time water level); set to [value] during the non-flood season (October-May). =0.2、 =0.25、 =0.35、 =0.2 (focusing on topography and regional vulnerability).

[0027] Risk level mapping rule: R < 4 is low risk, 4 ≤ R < 7 is medium risk, 7 ≤ R < 9 is high risk, and R ≥ 9 is extremely high risk. In other words, risk levels are divided into low risk, medium risk, high risk, and extremely high risk.

[0028] Step S22, Dual Authentication and Early Warning Trigger: Preset trigger condition parameters: The preset risk threshold is "medium risk" (R≥4 points). The preset risk threshold can be adjusted to low risk or high risk according to the city's flood control level; the first preset water level threshold is 5cm, the preset probability threshold is 0.75, and the second preset water level threshold is 15cm (higher than the first preset water level threshold, used for direct triggering in emergency situations). These can be adjusted according to factors such as the city's flood control level and geographical environment.

[0029] Bayesian inference model training and application: Based on the meteorological forecast data, real-time water level data and actual flooding records of the target city over the past 5 years, a joint probability model of meteorological forecast risk and real-time water level is trained. The model input is "meteorological risk level" and "whether the real-time water level exceeds the threshold", and the output is the posterior probability of the occurrence of flooding events.

[0030] For example, scenario 1, also known as the first triggering condition: At 14:00 on June 15, meteorological forecast data shows that the precipitation in a certain area of ​​the target city in the next 6 hours is 45mm, and the risk index R is calculated to be 7.2 (medium risk); at 14:05, three water level monitoring points in the area reported water levels of 6.3cm, 5.8cm, and 4.9cm respectively (two of which exceeded the first preset water level threshold of 5cm); the system calculates the posterior probability of waterlogging using a Bayesian inference model to be 0.83 (exceeding the preset probability threshold of 0.75), and determines that the warning triggering condition is met.

[0031] Scenario 2, also known as the second triggering condition: At 16:30 on June 18, the water level monitoring point at the entrance of an underground parking garage directly reported a water level of 16.2cm (exceeding the second preset water level threshold of 15cm). The system does not need to calculate the posterior probability and directly determines that the warning triggering condition is met.

[0032] It should be noted that the purpose of setting the first triggering condition in this invention is to initiate a high-precision Bayesian model for secondary verification when the risk reaches a medium level (R≥4 points) and preliminary on-site evidence is obtained (water level exceeds the first preset water level threshold by 5cm), thus balancing the timeliness and accuracy of the warning. A high-risk level (R≥7 points) is itself a strong risk signal, usually accompanied by water level exceeding the threshold, thus naturally satisfying the first triggering condition. Setting the second triggering condition (water level exceeding the second preset water level threshold by 15cm) serves as an independent safety baseline, ensuring an unconditional and immediate response regardless of the predicted risk level when the sensor clearly detects an extremely high water level.

[0033] like Figure 2 As shown, when performing dual verification in practical applications, the second trigger condition can be checked first. If the condition is met, the warning trigger condition is directly determined to be met. If the condition is not met, the first trigger condition is checked. If the condition is met, the posterior probability of waterlogging is calculated using a Bayesian inference model. If the posterior probability is greater than the preset probability threshold, the warning trigger condition is determined to be met. Otherwise, the monitoring status is returned.

[0034] Step S23: Delineating the dynamic early warning range: Starting from the monitoring point where water level data triggers an early warning, a two-dimensional surface runoff model (based on shallow water equations) is activated, and the following data is input: 1) 1:5000 high-precision DEM data; 2) roughness coefficients corresponding to different surface cover types (asphalt pavement 0.012, cement pavement 0.015, grassland 0.03, green space 0.04, building area 0.02); 3) design flow rate, pipe diameter, and distribution data of the drainage network in the area; 4) real-time water level data and predicted precipitation intensity.

[0035] Numerical calculations using the finite difference method simulate the diffusion path, velocity, and depth of accumulated water under gravity. A preset depth threshold of 10cm is set, and areas with water depths exceeding 10cm in the simulation results are designated as the warning impact area. The final output is polygonal geofence data (in WKT format), achieving meter-level accuracy in area delineation. For example, for the area triggering the warning in Scenario 1, the final warning area is a polygonal region with an actual radius of 850 meters centered on the monitoring point, accurately covering three surrounding residential communities, two low-lying road sections, and the underground parking garage of a commercial complex.

[0036] Step S3: In response to the flood risk warning, simultaneously trigger the insurance protection service associated with the flood risk warning, and generate insurance compensation commitment information for the flood risk warning. The core of this step is to achieve synchronous linkage between "flood risk warning generation" and "insurance protection activation and compensation commitment generation," which is implemented as follows: Step S31, Synchronous Trigger Mechanism: After the data fusion and decision-making module generates an urban flood risk warning, it sends a trigger signal to the insurance linkage service module through the internal event bus (response delay ≤200ms). The signal contains core information such as the warning event ID, warning level, dynamic warning range WKT data, and trigger time, ensuring that there is no time difference between "warning generation" and "insurance activation".

[0037] Step S32, On-chain Evidence Storage Implementation: For example, deploying an insurance smart contract on the Hyperledger Fabric consortium blockchain. The contract pre-defines core terms such as the conditions for coverage activation, the scope of compensation, and evidence storage rules. After receiving the warning trigger signal, the insurance linkage service module automatically calls the smart contract interface to store the following information on the blockchain: 1) The hash value of the warning event (generated by encrypting the warning ID, trigger time, and scope of impact); 2) The screening criteria for affected policies (within the warning scope + the policy status is valid + the insurance type includes coverage for flooding losses); 3) The coverage activation time (consistent with the warning trigger time). The on-chain data is confirmed through the consortium blockchain's consensus mechanism (PBFT algorithm) to ensure that the data is tamper-proof and traceable, providing legal and data support for the insurance compensation commitment.

[0038] Step S33, Compensation Amount Estimation and Pre-authorization Token Generation: A Gradient Boosting Regression (GBR) machine learning model is used. This model is trained based on the target city's flooding claims data (including water depth, vehicle damage level, and repair costs) from the past three years, regional vehicle density data, and local average vehicle value data (provided by the Insurance Association; the current average vehicle value in the target city is 198,000 yuan). The model's prediction error is controlled within ±7%. Specific operation: Input the average water depth of the warning area (6.05cm in scenario 1), regional vehicle density (286 vehicles / square kilometer), and average vehicle value (198,000 yuan). The model quickly estimates the compensation amount per vehicle to be between 6,000 and 18,000 yuan. The total compensation amount for this warning area is estimated at 1.56 million yuan. The system generates a unique digital pre-authorization token for each affected user. The token is generated using the UUIDv4 algorithm and contains the following information: user policy number, warning event ID, compensation range, token validity period (72 hours), and fast claims entry identifier. The token is stored in both the consortium blockchain and the core insurance business system to ensure that it is verifiable.

[0039] Step S34: Generation of Insurance Compensation Commitment Information: Based on the on-chain stored guarantee rules and pre-authorized token information, standardized insurance compensation commitment information is automatically generated. Key elements include: 1) Guarantee activation status (clearly marked "activated"); 2) Compensation range (e.g., "6000-18000 yuan"); 3) Activation conditions (direct losses to the vehicle body, engine, electrical system, etc., caused by flooding within the affected area of ​​this warning); 4) Claim voucher (pre-authorized token ID); 5) Fast claims process (applicable scenarios with one-click application via APP and no on-site inspection required); 6) Exclusion clauses (excluding losses caused by intentionally driving through floodwaters or failing to take reasonable precautions according to the warning). The compensation commitment information is generated synchronously with the flood risk warning information, with a time difference of ≤300ms, ensuring the core logic of "warning equals commitment" is implemented.

[0040] Step S4: Identify vehicle users within the area affected by the flood risk warning and send a warning notification containing the flood risk warning information and the insurance compensation commitment information to the vehicle users. This step specifically includes the following: Step S41: User Geographic Location Acquisition and Semantic Recognition: The geographical location data of the insured user is collected through the location SDK built into the insurance app, integrating GPS, cell tower positioning, and Wi-Fi positioning. GPS positioning is used as the primary positioning source (positioning accuracy ≤ 5 meters in outdoor scenarios), supplemented by cell tower positioning (accuracy ≤ 20 meters in indoor scenarios) and Wi-Fi positioning (accuracy ≤ 10 meters in indoor scenarios). A Kalman filter algorithm is used to smooth the positioning trajectory, eliminating positioning jump errors, ultimately stabilizing the positioning accuracy within 10 meters.

[0041] Location semantic processing: The smoothed user location coordinates are matched with the POI database of a third-party map service provider (such as Gaode Maps). The spatial inclusion relationship is used to determine the type of scene the user is in, specifically divided into four semantic labels: "underground parking garage", "low-lying road section", "high-lying parking lot", and "ordinary road", providing scene support for individual risk level classification. For example, if user A's location coordinates fall within the POI range of "XX Community Underground Parking Garage", the semantic recognition result is "underground parking garage"; if user B's coordinates fall in an area where the elevation is 1.8 meters lower than the surrounding average elevation and the POI label is "road", the semantic recognition result is "low-lying road section".

[0042] Step S42, Individual User Risk Level Classification: Based on the semantic location type of the user and its distance from the flood risk center, three levels of individual risk are classified: Level 1 risk corresponds to high risk: the semantic location is "underground parking garage" or "low-lying road section", and the distance from the risk center is less than 500 meters; Level 2 risk corresponds to medium risk: the semantic location is "underground parking garage" or "low-lying road section", 500-1000 meters away from the risk center; or the semantic location is "ordinary road", less than 500 meters away from the risk center; Level 3 risk corresponds to low risk: the semantic location is "high ground parking lot", or 1000-2000 meters away from the risk center (regardless of the semantic location type).

[0043] Step S43, Differentiated Early Warning Push Implementation: The system has a built-in multi-channel push gateway, integrating APP push (JPush SDK), SMS (Alibaba Cloud SMS Service), voice calls (Tencent Cloud TTS), and human customer service access channels. It automatically matches the notification channel combination, notification template, and content urgency according to the individual risk level. For Level 1 risk users (e.g., User A, underground parking garage + 320 meters away): A triple-channel notification system will be used, consisting of "APP high-priority push notification (vibration + continuous ringing for 30 seconds) + SMS + voice call." The notification template is: "[Emergency Warning] Your vehicle is currently located in the underground parking garage of XX Community (high-risk area for flooding). The surrounding water level has reached 6cm and is expected to continue rising within 1 hour! Your flooding insurance has been activated, with a compensation commitment of 6,000-18,000 yuan. Pre-authorization token ID: XXX. You can apply for quick claims with one click through the APP. Please evacuate immediately. Do not risk moving your vehicle. Prioritize personal safety! Claims hotline: 400-XXX-XXXX." Push notification execution time requirements: From obtaining user information to successful push notification via the first channel (APP) ≤ 1 minute. The voice call will be initiated within 3 minutes after the APP push notification.

[0044] For users at Level 2 risk (e.g., User B, located in a low-lying area within 650 meters): A dual-channel notification system (APP push notification + SMS) will be used. The notification template will be: "[Flood Warning] There is a risk of flooding in the XX section of road you are in. Floodwater is expected to spread to this area within 2 hours. The current risk level is medium. Your vehicle flood insurance has been activated, with a compensation commitment of 6,000-18,000 yuan. A pre-authorization token has been generated and can be viewed in the 'My Insurance' section of the APP. It is recommended to prioritize personal safety and arrange vehicle parking accordingly." Push notification execution timeframe: All channels must push notifications within 5 minutes.

[0045] Level 3 risk users (e.g., User C, in the high-altitude parking lot + 1200 meters away): The "APP Standard Notification" single-channel notification will be used. The notification template is: "[Flooding Warning] Flooding warning has been issued for your area. Your vehicle's flooding insurance has been activated, with a compensation commitment of 6000-18000 yuan. If your vehicle is damaged due to this flooding, you can quickly claim compensation using a pre-authorized token. Please pay attention to changes in weather and road flooding." Push notification execution time requirement: Complete the push within 10 minutes.

[0046] Step S44, Implementation of Tiered Upgrade Notifications: The system establishes a notification status monitoring mechanism to track the user's "delivered → read → confirmed" status in real time. Level 1 risk users: If an app push notification is not read within 15 minutes, an SMS message is not replied to within 20 minutes, or a voice call is not answered, an automatic escalation process will be triggered: First, a voice call will be re-initiated (with a 5-minute interval). If the call is not answered twice, the user's information will be transferred to a live customer service representative, who will then call the user manually to notify them until the user confirms receipt or completes three attempts to call the representative manually.

[0047] Level 2 risk users: If the user does not confirm the notification within 30 minutes of receiving the app push notification and SMS message, a new SMS reminder will be sent to ensure the user is aware of the notification.

[0048] Example 2:

[0049] Based on Example 1, this embodiment further optimizes the power consumption management of the water level monitoring network composed of sensors, specifically including the following: Dynamic configuration of heartbeat reporting interval: The system has a built-in historical flooding data seasonal distribution model. Based on the flooding occurrence time and frequency data of the target city over the past 5 years, it divides the season into "low-risk season" (November to March of the following year) and "high-risk season" (April to October). Combined with the real-time flooding risk level, the sensor heartbeat reporting interval is dynamically and remotely configured. Low-risk season + low-risk level (R < 4 points) + no precipitation forecast: Configured in long interval mode, the heartbeat reporting interval is set to 72 hours. The sensor enters an ultra-low power sleep state. At this time, the sensor operating current is ≤1.5mA. Compared with the normal mode (10-minute interval), the power consumption is reduced by more than 95%, and the battery life is increased to 18 months. Low-risk season + medium-risk level (4≤R<7 points) or weak precipitation forecast: Configure as medium interval mode, heartbeat reporting interval set to 300 minutes; High-risk season or risk level ≥ high risk (R ≥ 7 points) or enter early warning state: Configure to short interval mode, heartbeat reporting interval is set to 60 minutes to ensure high-frequency risk monitoring.

[0050] Preferably, the sensor hardware integrates dual communication modules: a "heartbeat channel" and an "emergency alarm channel". Heartbeat channel: used for routine heartbeat data reporting, such as low-power transmission using the NB-IoT communication protocol; Emergency Alarm Channel: If using a 4G communication protocol, the sensor is normally in sleep mode. It locally stores an emergency threshold (preset at 8cm). When the water level suddenly exceeds this threshold, a local hardware interrupt triggers wake-up, immediately switching to the emergency alarm channel to report data. The reporting delay is ≤3 seconds, ensuring rapid response to the second trigger condition. Upon receiving such an emergency report, the system cloud determines whether the water level data has reached the second preset threshold (e.g., 15cm) at the business level. If it has, the second trigger condition is directly triggered; if not, it is recorded as a high-priority alarm event and may trigger more frequent monitoring. This design achieves decoupling and coordination between device-level rapid safety response and system-level business decision-making. For example, at a low-risk monitoring point, a sudden pipe rupture causes the water level to rise from 2cm to 9cm within 10 minutes. The sensor immediately reports via the emergency alarm channel, and the system directly triggers an early warning, avoiding missed reports due to long intervals between heartbeats.

[0051] Based on the trial operation data from 100 monitoring points in the target city, after adopting this intelligent power consumption management solution, the average battery life of the sensors increased from 6 months in the traditional mode to 15.8 months, and the maintenance frequency decreased by 62%. The equipment failure identification rate increased to 98% during high-risk seasons, and the monitoring blind spots caused by equipment power failure were reduced by 90%, effectively solving the problems of "short battery life, difficult maintenance, and high risk of missed reports" of traditional sensors.

[0052] Example 3:

[0053] like Figure 1 and Figure 2 As shown, this embodiment also provides an urban flooding vehicle risk protection system based on precipitation prediction. This system adopts a three-layer "cloud-edge-device" architecture: the terminal layer consists of water level sensors and user mobile terminals; the edge layer consists of regional data processing nodes (for real-time data preprocessing); and the cloud layer consists of core service modules (such as those deployed on an Alibaba Cloud ECS cluster), ensuring the system's scalability, real-time performance, and reliability. Specifically, it includes the following: A multi-source data acquisition module is used to acquire meteorological forecast data for the target area within a preset future time period, as well as real-time water level data from various monitoring points within the target area; including: Meteorological forecasting unit: Used to acquire high-resolution gridded precipitation forecast data for the target area in real time for a preset period from authoritative meteorological agencies. Its core function is to perform data access, quality control and spatial standardization processing, transforming raw meteorological data from different sources into a standardized forecast field that is spatiotemporally continuous and can be directly used by the urban flooding risk model, providing the most critical meteorological input for the system's early warning.

[0054] The water level sensing unit consists of a network of intelligent sensors deployed in key areas such as low-lying urban areas and underground facilities. Its core function is to monitor the depth of water accumulation on the surface or underground in real time and accurately, transmitting the monitoring data back via a low-power wide-area network. This unit features adaptive sampling and dual-mode communication capabilities, enabling ultra-low power operation under normal conditions to extend battery life, and immediate switching to high-speed reporting in case of sudden water level changes. This provides the system with accurate and reliable surface water condition verification data and emergency direct trigger signals.

[0055] Location Acquisition Unit: Integrated into the user terminal application, its core function is to acquire the geographical location information of insured vehicle users in real time and with high accuracy. By integrating multiple positioning technologies and filtering and optimizing trajectory data, this unit ensures the accuracy and stability of user location data. More importantly, this unit transforms raw coordinate data into semantic location tags with risk identification significance (such as "underground parking garage" or "low-lying road section"), providing a direct data foundation for subsequent accurate user risk classification and differentiated early warning notifications.

[0056] The data fusion and decision-making module is used to calculate the urban flooding risk level based on the meteorological forecast data; perform fusion analysis and dual verification based on the urban flooding risk level and the real-time water level data to determine whether the preset warning triggering conditions are met; if the preset warning triggering conditions are met, an urban flooding risk warning is generated; it includes a risk index calculation submodule, a Bayesian inference verification submodule, and a dynamic range delineation submodule.

[0057] Risk index calculation submodule: Based on real-time acquired meteorological forecast data, geographical environment data, and water level monitoring data, it uses a built-in multi-factor weighted model ( This submodule dynamically calculates the urban flooding risk index R for each grid unit within the target area. It dynamically configures weight coefficients based on seasonal and regional characteristics, mapping the calculated risk index to qualitative risk levels such as "low, medium, high, and extremely high," providing a quantitative and spatially distributed core basis for subsequent risk decision-making. It can pre-store geographic environmental data and a dynamic weight configuration table for the target city, supporting flexible adjustment of weight coefficients by season and region. It receives multi-source data every 5 minutes and calculates the risk index R for each grid unit in real time.

[0058] The Bayesian inference verification submodule performs a high-confidence probabilistic verification for scenarios initially assessed as medium-risk or higher and with water levels exceeding thresholds. This submodule uses a joint probabilistic model trained on historical data, taking the meteorological risk level and real-time water level exceedance as input, to quickly calculate the posterior probability of the actual occurrence of the flooding event. Its core function is to effectively balance the timeliness and accuracy of early warnings by introducing probabilistic judgments, reducing false alarm rates and providing scientific and rigorous decision support for triggering early warnings.

[0059] Once an alert is triggered, the dynamic range delineation submodule immediately simulates the water spread process based on a hydrodynamic model, using the trigger point as the core. This submodule integrates high-precision digital elevation models, surface roughness coefficients, and drainage network data to predict the depth and extent of water accumulation through numerical calculations. Its core function is to accurately delineate the geographical impact range where the water depth exceeds a preset threshold (e.g., 10cm) and output it in a standard geofencing format, thereby precisely extending the alert from a "point" to a "surface," enabling accurate location of affected vehicles and users.

[0060] The insurance linkage service module is used to respond to the flood risk warning, simultaneously trigger insurance protection services associated with the flood risk warning, and generate insurance compensation commitment information for the flood risk warning. This includes the following: Smart contract deployment provides underlying blockchain support for the automated and reliable execution of insurance protection services. Smart contracts deployed on the consortium blockchain pre-code the protection rules, activation conditions, and claims logic. Their core function is to automatically and immutably record the association information between the warning event and the affected policies when an alert is triggered, completing the "on-chain notarization" and "instant activation" of the protection service. This ensures the legal validity and transparency of subsequent compensation commitments, achieving the codified and automated operation of business logic.

[0061] The compensation amount estimation submodule utilizes a machine learning model to quickly and intelligently pre-assess the potential vehicle damage caused by this flooding event. This submodule takes the average water depth, vehicle density, and average vehicle value in the warning area as input, and outputs the estimated compensation amount range for individual vehicles and the overall situation based on a model trained on historical claims data. Its core function is to provide specific and reasonable quantitative basis for insurance compensation commitments, making the "commitment" clear and credible, and setting the monetary basis for generating pre-authorization tokens.

[0062] The pre-authorization token generation submodule generates a unique, time-sensitive digital credential containing key claims information for each affected insured user. This submodule binds information such as the user's policy, the warning event, and the estimated payout amount to generate a verifiable digital token. Its core function is to serve as an "electronic key" for subsequent rapid claims processing. Users can directly initiate claims with this token, simplifying the process, shortening the time frame, and technically fulfilling the user experience of rapid protection promised in the "warning as a commitment" principle.

[0063] A multi-channel notification module is used to identify vehicle users within the area affected by the flood risk warning and push warning notifications containing the flood risk warning information and the insurance compensation commitment information to these vehicle users. Specifically, it includes four sub-units: The location acquisition and semantic recognition unit is used to acquire the geographical location information of the insured vehicle user in real time and perform semantic recognition on the geographical location to distinguish whether the user is located in an underground garage, a low-lying road section or a high-lying parking lot.

[0064] The user risk classification unit is used to classify individual user risk levels based on the semantic location type of the user and its distance from the flood risk center.

[0065] The tiered push unit is used to automatically match different combinations of notification channels, notification templates, and content urgency based on the individual user's risk level, and to conduct differentiated push notifications.

[0066] The upgrade notification unit is used to send tiered upgrade notifications to high-risk users who do not confirm receipt within a preset time after the push notification is sent, according to the priority of the notification channel, until human customer service is involved.

[0067] It should be noted that this invention does not rely on any specific third-party service, and all technical solutions that can achieve the same data acquisition / deployment / exhibition functions are within the scope of protection.

[0068] The above description is only a preferred embodiment of the present invention. Therefore, any equivalent changes or modifications made to the structure, features and principles described in the claims of this patent application are included within the protection scope of this patent application.

Claims

1. A method for protecting vehicles from urban flooding based on precipitation forecasting, characterized in that, Includes the following steps: S1. Obtain meteorological forecast data for the target area within a preset time period, as well as real-time water level data for each monitoring point within the target area; S2. Calculate the urban flooding risk level based on the meteorological forecast data; Based on the fusion analysis and dual verification of the waterlogging risk level and the real-time water level data, it is determined whether the preset warning triggering conditions are met. If the preset warning triggering conditions are met, a waterlogging risk warning is generated. S3. In response to the waterlogging risk warning, simultaneously trigger the insurance protection service associated with the waterlogging risk warning and generate insurance compensation commitment information for the waterlogging risk warning; S4. Identify vehicle users within the area affected by the flood risk warning and push a warning notification containing the flood risk warning information and the insurance compensation commitment information to the vehicle users.

2. The method according to claim 1, characterized in that, Step S2, which involves calculating the urban flooding risk level based on the meteorological forecast data, includes: Based on the meteorological forecast data and geographical environment data, the risk index R of each location in the target area is calculated using an urban flooding risk prediction model. The risk index R is mapped to the corresponding urban flooding risk level.

3. The method according to claim 2, characterized in that, The risk index R is calculated using the following multi-factor weighted model: Where P is the predicted precipitation factor, D is the real-time water level and depth factor, T is the topographic index factor, and V is the regional vulnerability index factor. to The dynamic weighting coefficients for each factor are: the topography index factor is calculated based on digital elevation model data; the regional vulnerability index factor is calculated based on historical flooding event data and vehicle density data; and the dynamic weighting coefficients are adjusted according to seasonal and / or urban regional characteristics.

4. The method according to any one of claims 1 to 3, characterized in that, In step S2, the preset warning triggering conditions are determined through dual verification. A condition is considered met if one of the following conditions is satisfied: First triggering condition: If the waterlogging risk level reaches or exceeds a preset risk threshold, and any of the real-time water level data reaches or exceeds the first preset water level threshold, then the posterior probability of the waterlogging event is calculated based on a Bayesian inference model; when the posterior probability exceeds a preset probability threshold, it is determined that the warning triggering condition is met. Second triggering condition: If any real-time water level data reaches or exceeds the second preset water level threshold, it is directly determined that the warning triggering condition is met; wherein, the second preset water level threshold is greater than the first preset water level threshold.

5. The method according to claim 4, characterized in that, The generation of urban flooding risk warnings includes defining a dynamic warning range, which includes: Starting from the monitoring point where the water level data triggers the warning, based on the hydrodynamic model, combined with the digital elevation model, land cover type and drainage network data, the process of water accumulation diffusion is simulated, and the geographical influence range where the water depth exceeds the preset depth threshold is dynamically delineated.

6. The method according to claim 4, characterized in that, The process of acquiring the real-time water level data in step S1 includes intelligent power consumption management steps: Based on the aforementioned waterlogging risk level and / or early warning trigger status, dynamically and remotely configure the heartbeat reporting interval of the water level monitoring network; When in a low-risk or no-warning state, configure to long-interval mode to reduce power consumption; When the risk level rises or enters an early warning state, configure it to short interval mode to increase the monitoring frequency; The water level monitoring network supports a dual-mode communication mechanism. In the long interval mode, if the real-time water level data of any monitoring point exceeds the locally stored emergency threshold, it will be immediately reported through the emergency alarm channel.

7. The method according to claim 1, characterized in that, The triggering of insurance coverage services mentioned in step S3 includes: The event information of the flood risk warning and the information of the affected insurance policies are stored on the blockchain for evidence; based on the machine learning model, the compensation amount is estimated according to the water level depth, regional vehicle density and average vehicle value, and a time-sensitive digital pre-authorization token is generated as a fast claims settlement certificate.

8. The method according to claim 1, characterized in that, The step S4 of pushing a warning notification to the vehicle user includes: The system acquires the geographical location information of insured vehicle users in real time and performs semantic recognition on the geographical location to distinguish whether the user is located in an underground garage, a low-lying road section, or a high-lying parking lot. Users are classified into individual risk levels based on their semantic location type and their distance from the flood risk center; Based on the individual risk level of the user, different combinations of notification channels, notification templates, and content urgency are automatically matched for differentiated push notifications.

9. The method according to claim 8, characterized in that, Step S4 also includes steps to ensure that upgrade notifications are delivered: For high-risk users who do not confirm receipt of the push notification within the preset time, the notification will be upgraded in a tiered manner according to the priority of the notification channel, until human customer service intervention is required.

10. A vehicle risk protection system for urban flooding based on precipitation forecasting, characterized in that, include: The multi-source data acquisition module is used to acquire meteorological forecast data for the target area within a preset time period, as well as real-time water level data of each monitoring point within the target area. The data fusion and decision-making module is used to calculate the urban flooding risk level based on the meteorological forecast data; Based on the fusion analysis and dual verification of the waterlogging risk level and the real-time water level data, it is determined whether the preset warning triggering conditions are met. If the preset warning triggering conditions are met, a waterlogging risk warning is generated. The insurance linkage service module is used to respond to the waterlogging risk warning, simultaneously trigger the insurance protection service associated with the waterlogging risk warning, and generate insurance compensation commitment information for the waterlogging risk warning; The multi-channel notification module is used to identify vehicle users in areas affected by the flood risk warning and push warning notifications containing the flood risk warning information and the insurance compensation commitment information to the vehicle users.

Images