Road health code and pipe network hydraulic coupling municipal hidden danger linkage processing method

Abstract

The application discloses a kind of municipal hidden danger linkage processing methods based on road health code and pipe network hydraulic coupling, including obtaining road state data, drainage state data, environmental meteorological data and historical disposal data;Road health code evaluation module is constructed, and the comprehensive health index is generated by the AHP-entropy weight method combination weighting calculation road health grade and disease type vector;SWMM drainage pipe network hydraulic coupling module is constructed, and drainage risk index is calculated based on SWMM hydraulic model;Road-drainage hidden danger causal association diagram is constructed, and leading cause of complex hidden danger is inferred by Bayes network;Road maintenance, drainage maintenance and emergency rescue linkage disposal scheme are generated, and closed loop is updated according to work order acceptance result.The application can improve the correlation identification ability of road disease and drainage hidden danger, reduce repeated repair and flood season waterlogging risk, and be suitable for urban road, inspection well, gully, pipeline and pump station collaborative maintenance scene.

Description

Technical Field

[0001] This invention belongs to the technical field of municipal road maintenance, drainage network operation and maintenance and digital management of urban infrastructure, and specifically relates to a method for joint handling of municipal hidden dangers based on road health code and hydraulic coupling of pipeline network. Background Technology

[0002] As China's urbanization enters a high-quality renewal phase, the scale of existing urban roads and stormwater / sewage pipe networks continues to expand. A large number of aging roads and pipe networks are gradually entering a period of high incidence of defects, leading to frequent occurrences of complex municipal hazards such as potholes, subsidence, pipe network siltation and blockage, road flooding during heavy rains, and urban waterlogging. Road surface defects and drainage system deficiencies interact and amplify safety risks, placing higher demands on refined urban maintenance and flood control. Traditional manual inspections, paper-based records, and independent operation and maintenance by different specialties are no longer sufficient to meet the needs of predicting potential hazards, coordinating response, and efficiently allocating maintenance funds. Traditional maintenance suffers from significant shortcomings: emphasizing post-incident repairs over pre-incident warnings, fragmented road and pipe network data, and isolated handling of hazards. This results in high treatment costs, low efficiency, and a high likelihood of recurring waterlogging and road damage. Currently, municipal governments are actively promoting the construction of smart cities and resilient cities, and the industry is generally promoting the digital management of municipal facilities. The municipal pipeline network data maintenance system aims to optimize maintenance costs, flood risk, and road repair efficiency. It automatically generates balanced and optimal maintenance plans, balancing financial investment and treatment effects. It coordinates road repair and pipeline dredging during construction, reducing repeated excavation and inspection, significantly lowering the overall cost of municipal maintenance, and batch coordinating the priority of road maintenance across the entire region to achieve the optimal allocation of limited human and financial resources.

[0003] The existing technical solutions have the following problems: 1) Single road defect management system: The existing intelligent road inspection system only collects superficial defects such as road surface cracks and potholes, evaluates the degree of road damage by scoring, and lacks correlation analysis of pipeline drainage conditions; it has not established a standardized digital identity for road units (road health code), cannot quantify the long-term health evolution pattern, can only statically record defect locations, cannot predict future road condition deterioration trends, and relies solely on manual experience to formulate maintenance plans. 2) Shortcomings in health assessment and trend prediction technology: Existing road scoring systems mostly use simple weighted calculations and do not scientifically determine the weights of multiple disease indicators through the AHP (Analytic Hierarchy Process); they lack a short-term trend prediction mechanism based on a continuous 12-month health index time series, and cannot automatically output three trend labels: deterioration / stable / improvement, resulting in a lack of quantitative prediction support for maintenance decisions. Summary of the Invention

[0004] To address the aforementioned issues, the purpose of this invention is to provide a method for the coordinated handling of municipal hazards based on road health codes and hydraulic coupling of pipe networks. This method integrates closed-loop data from road health monitoring, drainage network operation, and maintenance work orders to dynamically assess the causal relationship between road defects and drainage hazards, and automatically generates cross-disciplinary coordinated response strategies.

[0005] This was achieved through the following technical solutions: A method for joint handling of municipal hidden dangers based on road health codes and pipeline hydraulic coupling includes the following steps: S1. Obtain the initial dataset of the target road unit, including road status data, drainage status data, environmental meteorological data, and historical treatment data; S2. Construct a road health code evaluation module based on the initial dataset of the target road unit to generate the health level, disease type vector, and disease evolution trend of the target road unit; S3. Construct the SWMM drainage network hydraulic coupling module and calculate the drainage risk index by associating the initial dataset of the target road unit; S4. Based on the health level and disease type vector of the target road unit in step S2 and the drainage risk index in step S3, construct a causal relationship diagram of road-drainage hazards and infer the dominant causes of hazards. S5. Based on the primary causes of hidden dangers, calculate and sort the comprehensive risk scores of the target roads in descending order, output the joint response plan, and generate work orders; S6. Read and collect work order completion information, and update the road health code and drainage risk index for model iteration and optimization; work order completion information includes: on-site photos, list of disposal materials, operation time, acceptance opinions and recurrence tracking results.

[0006] Optionally, the road condition data includes one or more of the following: pavement images, potholes, cracks, subsidence, manhole damage, water accumulation on the road surface, sidewalk damage, curbstone damage, and abnormal data on road ancillary facilities.

[0007] Optionally, the drainage status data may include one or more of the following: storm drain blockage status, manhole water level, pipeline defects, pipeline siltation rate, pump station operation status, dredging records, and drainage emergency response records.

[0008] Optionally, the drainage network hydraulic coupling module establishes a one-dimensional hydraulic model of the drainage network based on SWMM, inputs the initial dataset of the associated target road units, and calculates the drainage risk index under current or predicted rainfall conditions; the model input includes: pipe segment geometric parameters, manhole parameters, rainwater inlet parameters, pump station parameters, and boundary conditions.

[0009] Optionally, the formula for calculating the drainage risk index is: R_k = β1·f_full(k) + β2·f_depth(k) + β3·f_silt(k) + β4·f_pump(k), Where f_full, f_depth, f_silt, and f_pump are the risk functions for pipe section fullness, manhole water depth, pipeline siltation rate, and pump station operation, respectively; k is the drainage facility number; and β is the weighting coefficient.

[0010] Optionally, the associated drainage risk index R_i for each road unit i is the weighted average of the drainage risk indices of all drainage facilities within the catchment area of ​​each road unit, expressed as: Ri^road=Σ{k∈Φ(i)}γ_k·R_k, Where Φ(i) is the set of drainage facilities within the catchment area of ​​road unit i; γ_k is the spatial weight coefficient, which is inversely proportional to the distance from the facility to the road unit.

[0011] Optionally, the road-drainage hazard causal relationship graph is structured as follows: road units, storm drains, manholes, pipe sections, pumping stations, and work order events are used as nodes, and spatial proximity, confluence direction, upstream and downstream pipeline relationships, and historical work order co-occurrence relationships are used as edges. According to claim 1, the method for linking and handling municipal hazards based on road health codes and pipeline hydraulic coupling is characterized in that the road-drainage hazard causal relationship graph is structured as follows: road units, storm drains, manholes, pipe sections, pumping stations, and work order events are used as nodes, and spatial proximity, confluence direction, upstream and downstream pipeline relationships, and historical work order co-occurrence relationships are used as edges.

[0012] Optionally, the road health code and drainage risk index can be updated to adjust the weights of corresponding nodes and related edges of the same unit in the road-drainage hazard causal relationship graph within a preset observation period. This helps to uncover hidden and related hazards and effectively solve the blind spot problem of single monitoring. Optionally, the road health code evaluation module is used to calculate a comprehensive health index based on road grade, traffic volume, type of damage, area of ​​damage, depth of damage, frequency of recurrence, and historical maintenance results, by combining weights using the entropy weight method, to classify road health into levels and achieve standardized and quantifiable control over road health status.

[0013] Furthermore, a municipal hazard linkage handling system based on road health codes and pipeline hydraulic coupling is proposed, including a data acquisition module, a road health code evaluation module, an SWMM drainage pipeline hydraulic coupling module, a causal relationship graph inference module, a work order generation module, and a closed-loop update module. The data acquisition module acquires the initial dataset of the target road unit. The road health code evaluation module receives the initial dataset from the data acquisition module and generates health level and hazard type vectors for the target road unit. The SWMM drainage pipeline hydraulic coupling module calculates the drainage risk index by associating the initial dataset of the target road unit with the data. The causal relationship graph inference module constructs a causal relationship graph of road-drainage hazards based on the health level from the road health code evaluation module and the drainage risk index from the SWMM drainage pipeline hydraulic coupling module, and infers the dominant causes of the hazards. The work order generation module outputs a linkage handling plan and generates a work order based on the dominant causes of the hazards from the causal relationship graph inference module. The closed-loop update module collects work order information from the work order generation module, updates the road health code and drainage risk index, and performs model iterative optimization. The beneficial effects of this invention compared to the prior art are: 1) Hydraulic coupling to predict potential risks: By connecting to the SWMM hydraulic model of the pipeline network for real-time calculation, it is possible to simulate hydraulic changes such as pipeline overflow, pipe section siltation, and pressurized leakage under rainfall conditions in advance. Combined with the road health code, it can determine whether underground pipeline defects will induce road collapse, potholes, and settlement, thus achieving early warning of potential risks rather than post-event repairs.

[0014] 2) Digital and standardized management of road health codes: Unify the quantitative evaluation standards for road health, quantify indicators such as road surface damage, roadbed settlement, and weak soil cover areas for pipelines into health code scores, and digitally archive the status of the road throughout its entire life cycle; real-time correction of road health scores based on hydraulic data, automatic downgrading of road health levels when pipeline leakage continuously erodes the roadbed, and dynamic updating of risk labels. Attached Figure Description

[0015] Figure 1 This is a flowchart of the municipal hazard linkage response process of the present invention; Figure 2 This is a schematic diagram of the road health code evaluation module and the drainage network hydraulic coupling module of the present invention; Figure 3 This is a schematic diagram illustrating the construction of the causal relationship diagram of road-drainage hazards according to the present invention; Figure 4 This is a schematic diagram of the generation and closed-loop update process of the linkage handling scheme of the present invention; Figure 5 This is the Pareto optimal solution of the present invention. Detailed Implementation

[0016] The technical solutions in the embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0017] This invention discloses a method for the coordinated handling of municipal safety hazards based on the coupling of road health codes and pipeline hydraulics. For example... Figure 1 The diagram shown is a flowchart of the municipal hazard linkage handling process of this invention. A municipal hazard linkage handling method based on road health codes and drainage network hydraulic coupling utilizes a system comprising six modules: a data acquisition module, a road health code evaluation module, a drainage network hydraulic coupling module, a causal relationship graph reasoning module, a work order generation module, and a closed-loop update module. This system achieves integrated closed-loop management of road drainage coupling hazards, from data collection, intelligent diagnosis, risk assessment, work order issuance to model iterative optimization. The system comprises the following modules: a data acquisition module for acquiring the initial dataset of the target road unit; a road health code evaluation module for receiving the initial dataset from the data acquisition module and generating health level and defect type vectors for the target road unit; a SWMM drainage network hydraulic coupling module for calculating the drainage risk index by associating the initial dataset of the target road unit; a causal relationship graph inference module for constructing a causal relationship graph of road-drainage hazards based on the health level from the road health code evaluation module and the drainage risk index from the SWMM drainage network hydraulic coupling module, and inferring the dominant causes of hazards; a work order generation module for outputting a joint response plan and generating a work order based on the dominant causes of hazards from the causal relationship graph inference module; and a closed-loop update module for collecting work order information from the work order generation module, updating the road health code and drainage risk index, and performing model iterative optimization. A method for joint handling of municipal hazards based on road health code and network hydraulic coupling includes the following steps: S1. Obtain the initial dataset for the target road unit, including road status data, drainage status data, environmental meteorological data, and historical handling data. The initial dataset is acquired through road patrol vehicles, mobile inspection terminals, fixed video equipment, drainage facility sensors, pipeline detection equipment, meteorological data interfaces, and historical work order systems, and is associated according to road unit number, facility number, spatial coordinates, and collection time. Before entering the road health code evaluation module, each data item is converted into a calculable value, level, label, or image feature. Road condition data includes one or more of the following: pavement images, potholes, cracks, subsidence, manhole damage, road surface water accumulation, sidewalk damage, curbstone damage, and abnormal data of road ancillary facilities; drainage condition data includes one or more of the following: storm drain blockage status, manhole water level, pipeline inspection defects, pipeline siltation rate, pump station operation status, dredging records, and drainage emergency records; environmental meteorological data includes rainfall, rainfall duration, short-term forecasts, water accumulation alarms, and tide or river water level; historical handling data includes maintenance work orders, emergency work orders, handling duration, acceptance results, and number of recurrences at the same location. The quantification methods for road condition data are as follows: road surface images are pictures or video frames with timestamps, road unit numbers, and geographic coordinates; potholes are represented by quantity, area, maximum length, maximum width, and estimated depth; cracks are represented by crack quantity, total length, average width, maximum width, crack type, and crack density; subsidence is represented by subsidence area, maximum subsidence depth, subsidence center location, and trend; manhole perimeter damage is represented by manhole cover height difference, manhole perimeter damage area, manhole perimeter crack length, subsidence depth, and damage type; road surface water accumulation is represented by water accumulation area, water accumulation depth, water accumulation duration, and water accumulation location; sidewalk damage is represented by damage quantity, damage area, maximum height difference, and damage type; curbstone damage is represented by damage length, damage quantity, misalignment distance, and height difference; and abnormalities in road ancillary facilities are represented by facility number, abnormality type, abnormality quantity, abnormality location, and severity.

[0018] Furthermore, the quantification of drainage status data is as follows: The clogging status of storm drains is obtained through inspection images, manual inspections, or cleaning records, and the clogging rate is represented by the percentage of the effective water passage area that is blocked. A clogging rate of 0-20% indicates minor clogging, 20%-60% indicates moderate clogging, and above 60% indicates severe clogging. The water level in inspection wells is obtained through level sensors, manual measurement, or video recognition, and is represented by real-time water level, water level rise rate, whether it exceeds the warning level, and duration. Pipeline defects are detected through CCTV, QV, sonar detection, or manual inspection. Inspection data is obtained and expressed according to defect type, defect level, defect location, defect length, or area; pipeline siltation rate is expressed as the proportion of the pipeline cross-section occupied by silt, sand, or debris; pump station operating status is expressed as the number of start-stop cycles, cumulative operating time, instantaneous flow rate, current and voltage, number of alarms, and fault status; dredging records are expressed as dredging time, dredging location, dredging length, dredging volume, operation method, and acceptance results; drainage emergency records are expressed as the event time, water accumulation location, cause of emergency response, response measures, response duration, vehicles and personnel dispatched, and response results. Environmental meteorological data is quantified as follows: rainfall is expressed as precipitation depth per unit time; rainfall duration is expressed as the duration of a single rainfall event; short-term forecasts are expressed as the predicted rainfall start time, predicted rainfall intensity, predicted duration, and affected area; water accumulation alarms are expressed as alarm time, water accumulation location, water accumulation depth, alarm level, and duration; tide levels or river levels are expressed as real-time water levels, the difference from warning levels, and the rate of water level rise. The quantification of historical handling data is as follows: maintenance work orders and emergency work orders are represented by work order number, location, type of disease or event, dispatch time, arrival time, handling measures, and completion status; handling time includes response time, operation time, and total handling time; acceptance results are represented as qualified, unqualified, requiring re-inspection, or acceptance score; the number of recurrences at the same location indicates the number of times the same type of disease or drainage anomaly occurs again at the same road unit or the same facility location within the preset statistical period.

[0019] S2. Construct a road health code evaluation module based on the initial dataset of the target road unit, generating the health level, disease type vector, and disease evolution trend of the target road unit. For example... Figure 2 As shown, the steps for generating a road health code, including the road health code evaluation module of the present invention, are as follows: S2.1 Multimodal Defect Identification and Multidimensional Index Quantization. Specifically, firstly, a target detection model (such as YOLOv8 or Faster R-CNN) is used to identify defects in the pavement images of road units; the training dataset for the target detection model comes from historical road inspection images and manual annotation results. Then, the indicators of the training dataset of the target detection model are quantified to construct a defect evaluation index vector; the defect evaluation index vector for each road unit i is represented as: V i=[v1,v2,v3,v4,v5,v6,v7], where v1 is the percentage of damaged area (the sum of the areas of all types of damaged areas as a percentage of the total area of ​​the road unit); v2 is the maximum depth of damaged areas (mm); v3 is the road grade coefficient (1.0 for expressways, 0.8 for arterial roads, 0.6 for secondary arterial roads, and 0.4 for local roads); v4 is the standardized value of average daily traffic volume (actual traffic volume / designed traffic volume); v5 is the recurrence frequency (the number of times the same type of damaged area has recurred at this location in the past 12 months); v6 is the historical maintenance effect score (0~1, 1 indicates no recurrence after maintenance); and v7 is the drainage impact factor (the normalized value of the associated drainage risk index calculated in step S3).

[0020] S2.2, Assign weights to the index combination, and calculate H using the disease evaluation index vector from step S2.1. i The road health code evaluation module calculates a comprehensive health index based on road grade, traffic volume, type of road damage, area of ​​damage, depth of damage, frequency of recurrence, and historical maintenance effectiveness. This is achieved through a combination of the analytic hierarchy process (AHP) and entropy weighting. The calculation process is as follows: First, subjective weights w_j^AHP are determined based on the road health score using the AHP method. To avoid subjective bias from a single expert, a group decision-making approach is adopted in the AHP stage. At least five experts in road maintenance, structural inspection, and drainage engineering are invited to independently construct a judgment matrix. Only valid matrices with a consistency ratio less than 0.1 are retained, and the geometric mean is taken as the group judgment matrix to obtain the subjective weights. Then, based on historical damage data, objective weights (w_j^entropy weight) are calculated using the entropy weighting method to determine the weights of each indicator. The final combined weights are expressed as: w j =α·w_j^AHP+(1-α)·w_j^Entropy weight, where α is the preference coefficient (0.5≤α≤0.7), used to balance subjective experience and objective data; the comprehensive health index H of road unit i. i Based on w_j, the formula is: H i =100×[1-(Σ j w j ·v ij ) / Σ j w j ].

[0021] S2.4. Based on the road health score, output a road health evaluation, including: road health code level classification, disease type vector generation, and disease evolution trend prediction. The road health code level classification relies on the comprehensive health index H. iRoad units are divided into four health levels. The comprehensive health index thresholds (85, 70, 55) for road unit i are determined based on the following: referring to the percentage-based excellent, average, and poor grading convention in the "Technical Specification for Urban Road Maintenance" (CJJ 36-2016); based on the cumulative frequency distribution of the comprehensive health index from the local historical disease database over the past three years (85, 70, and 55 correspond to approximately the top 15%, top 35%, and top 65 percentiles, respectively), ensuring that the grading results conform to the actual statistical patterns of disease; and achieving cost-benefit matching with the response times (4 hours, 24 hours, and 7 days) of the red, orange, and yellow work orders in step S5. Accordingly, when H... i When the health level of a road unit is ≥85, it is green (healthy); when 70≤H i When the health level is <85, the health level of the road unit is yellow (Note); when 55≤H i When the health level is <70, the road unit's health level is orange (warning); when H i When the health level is less than 55, the health level of the road unit is red (dangerous).

[0022] The representation of the disease type vector is: D i =[d1,d2,d3,d4], where D i The vector represents the disease type, where four components represent the probability that the disease is caused by structural damage (d1), surface wear (d2), construction disturbance (d3), and drainage impact (d4), respectively, and the sum of each component is 1. The output disease type vector is based on the spatial distribution characteristics of the disease (such as whether it is concentrated around the well or distributed along the fracture direction), historical maintenance records, and drainage impact factors. These features are input into a Softmax multi-classifier (using logistic regression or a shallow neural network structure), which outputs the probability values ​​of the four categories after being mapped by the softmax activation function. This transforms the cause determination problem into a standard probability classification problem, ensuring the objectivity and reproducibility of the output.

[0023] Disease evolution trend prediction is based on the health index sequence of a road unit over the past 12 months, using the grey prediction model GM(1,1) to predict the health index for the next 3 months. The health index sequence of a road unit over the past 12 months is represented as: H (0) ={H (0) (1),H (0) (2),……,H (0) (12)}. The specific steps are: using the grey prediction model GM(1,1) to infer short-term trends. Its core principle is to transform the original discrete sequence into a more monotonic cumulative sequence H through first-order cumulative generation (AGO). (1) (k)=∑km=1H (0) (m) is used to weaken random fluctuations and highlight potential exponential patterns; then, a sequence Z adjacent to the mean is constructed.(1) (k) = 0.5 × [H (1) (k)+H (1) (k-1)] (k≥2), establish the grey differential equation H (0) (k)+a·Z (1) (k) = b. Here, a is the development coefficient, reflecting the inherent growth or decline rate of the health index; b is the grey action quantity, used to capture the overall impact of external disturbances on health status. Using the least squares method to estimate the parameters of the above linear equation, the optimal estimates of a and b can be obtained. Then, the whitening differential equation dH... (1) (t) / dt+a*H (1) The time response function of (t)=b: H^ (1) (t)=[H (0) (1)-ba]e -a(t-1) +b / a. Calculate the predicted values ​​of the accumulated sequence at future times t=13, 14, 15 (i.e., the 1st, 2nd, and 3rd months in the future), and then restore them through the cumulative subtraction operation H^. (0) (t)=H^ (1) (t)-H^ (1) (t-1) yields the final predicted health index value H^ (0) (13) H^ (0) (14) H^ (0) (15). If the three predicted values ​​show an increasing trend month by month, the output is "improving trend"; if they show a decreasing trend month by month, the output is "deteriorating trend"; otherwise, the output is "stable trend". Considering that the health index sequence of some road units may be closer to linear change than strict exponential law, this invention also introduces linear regression as a supplementary discrimination method. Let H(t)=α+β⋅t+ε, where t is the time sequence (1~12), β is the regression slope, representing the monthly average rate of change of the health index (positive value is an increase, negative value is a decrease), α is the intercept term (only used for fitting, not involved in trend determination), and ε is the random error term. When β is greater than the preset positive threshold (e.g., +0.5 points / month, which is determined based on the standard deviation of historical data fluctuations), it is judged as an improving trend; when β is less than the preset negative threshold (e.g., -0.5 points / month), it is judged as a deteriorating trend; and when it is between the two, it is considered as a stable trend. The GM(1,1) and linear regression methods are calculated in parallel. The final trend label is mainly based on the GM(1,1) output. If the two conclusions conflict, the more conservative (deterioration direction) result is taken to ensure that no warning is missed.

[0024] S3. Construct the SWMM drainage network hydraulic coupling module and calculate the drainage risk index by associating it with the initial dataset of the target road unit. For example... Figure 2 The diagram illustrates the process of calculating hydraulic coupling risks based on SWMM pipe networks. The specific steps are as follows: S3.1 The drainage network hydraulic coupling module establishes a one-dimensional hydraulic model of the drainage network based on SWMM. It inputs the initial dataset of the associated target road units and calculates the drainage risk index under the current or predicted rainfall conditions. The model inputs are: pipe segment geometric parameters, including pipe diameter, pipe length, slope, and roughness coefficient; manhole parameters, including manhole bottom elevation and manhole depth; rainwater inlet parameters, including type, catchment area, and interception capacity; pump station parameters, including pump flow-head curve and start-stop water level; and boundary conditions, including downstream water level and tide level. The process of constructing the SWMM model is as follows: First, modeling data is extracted from the drainage network GIS database and field inspection records, including pipe segment geometric parameters (pipe diameter, pipe length, slope, roughness coefficient), manhole parameters (manhole bottom elevation, manhole depth), storm drain parameters (type, catchment area, interception capacity), pump station parameters (flow-head curve, start and stop water levels), and boundary conditions (downstream water level, tide level). Second, manholes, storm drains, and pump stations are generalized as nodes, and pipelines are generalized as pipe segments. Upstream and downstream topological connections are established based on the drainage flow direction. Then, the study area is divided into sub-catchment areas based on the DEM, and the data for each sub-catchment area is extracted using GIS tools. The area, width, slope, and impermeability of the catchment area are determined. These parameters are then assigned to the SWMM model. The roughness coefficient of the pipe sections is determined according to the Manning formula based on the pipe material (0.011~0.017 for concrete pipes, 0.009~0.015 for plastic pipes). Design rainfall or short-term forecast rainfall is set as the input boundary based on environmental meteorological data. Finally, the simulation time step is set (a smaller step size is required for dynamic wave solutions to maintain numerical stability). The model is run to calculate the hydraulic response under current or predicted rainfall conditions, outputting key hydraulic parameters such as the maximum filling degree of each pipe section, the maximum water depth of the manhole, the runoff ratio of the storm drain, and the operating water level of the pumping station. After the model is built, historical rainfall and measured water level data can be used to calibrate and verify sensitive parameters such as the Manning coefficient and impermeability to ensure simulation accuracy.

[0025] S3.2. Based on the environmental meteorological data obtained in step S1, set up the design rainfall scenario, including rainfall duration (e.g., 1h, 2h, 6h, 24h) and rainfall return period (e.g., 1 year, 2 years, 5 years, 10 years, 20 years), as well as the real-time input of short-term forecast rainfall.

[0026] S3.3 Run the SWMM network model to simulate the hydraulic response of each drainage facility under the set rainfall scenario, and output key hydraulic parameters, including: the maximum filling degree of each pipe section (h / D), the maximum water depth of each inspection well (m), the ratio of the actual runoff to the design runoff of each rainwater inlet, and the relationship between the operating water level and the start / stop water level of each pumping station. Specific relationships include: start and stop relationship: when the actual water level rises to the start water level, the pump starts; when the water level drops to the stop water level, the pump stops. The pump start-up water level is higher than the pump stop-down water level, and the difference between the two forms a hysteresis range to prevent frequent start-ups and shutdowns due to water level fluctuations. A multi-stage progressive relationship is established: pump stations equipped with multiple pumps have progressively increasing start-up water levels (e.g., first start-up water level, second start-up water level), with pumps starting up at each level as the water level rises. An alarm water level (maximum start-up water level) is set, triggering full pump operation. The pump stop-down water level is the minimum control water level; pumps are shut down at each level as the water level falls back to this level. Safety constraints are maintained: the operating water level must be maintained between the minimum operating water level (to ensure equipment cooling and cavitation margin) and the maximum warning water level (to prevent upstream pipeline overload and overflow). Optimized scheduling is achieved: appropriately increasing the start-up water level within a safe range can reduce head and energy consumption; increasing the difference between start-up and stop-down water levels can extend pump life; increasing the operating water level can utilize the storage capacity of the collection tank and upstream pipeline network to reduce peak flows.

[0027] S3.4. For each drainage facility k, calculate the drainage risk index using the following formula: R_k = β1·f_full(k) + β2·f_depth(k) + β3·f_silt(k) + β4·f_pump(k), Where f_full, f_depth, f_silt, and f_pump are the risk functions for pipe section fullness, manhole water depth, pipe siltation rate, and pump operation, respectively; k is the drainage facility number; and β is the weighting coefficient. f_full(k) = max(0, (h / D)_k - (h / D)_threshold) / (1 - (h / D)_threshold), is the pipe section fullness risk function (taken as 0 when h / D ≤ threshold); referring to the maximum design fullness provisions for different pipe diameters in the "Outdoor Drainage Design Code" (GB 50014), the maximum design fullness provisions are 0.55 for pipe diameters of 200~300mm, 0.65 for 350~450mm, 0.70 for 500~900mm, and 0.75 for ≥1000mm. f_depth(k) = H_well(k) / H_well design(k) is the inspection well water depth risk function; f_silt(k) = siltation rate_k, which is the pipeline siltation rate (0~1); f_pump(k) = max(0,(H_operation-H_pump start) / (H_pump stop-H_pump start)), which is the pump station operation risk function; β1, β2, β3, β4 are weighting coefficients (β1+β2+β3+β4=1), which are dynamically adjusted according to the facility type. For example, for pipe sections, β1 and β3 are mainly considered, and for pump stations, β4 is mainly considered. The weight allocation follows the principle of varying according to facility and focusing on the dominant risk. It is dynamically adjusted based on the core failure modes of different drainage facilities, mainly including: Pipeline facilities (dominant risks are insufficient flow capacity and siltation / blockage): Emphasis is placed on fullness and siltation rate, with typical weights of β1=0.40, β2=0.10, β3=0.40, β4=0.10; Pumping station facilities (dominant risk is abnormal operation): Emphasis is placed on pumping station operation risks, with typical weights of β1=0.15, β2=0.10, β3=0.15, β4=0.60; Manhole facilities (dominant risk is overflow due to excessive water depth): Emphasis is placed on manhole water depth risks, with typical weights of β1=0.20, β2=0.50, β3=0.20, β4=0.10. The initial values ​​of each weight are determined by expert experience combined with the Analytic Hierarchy Process (AHP), and are dynamically adjusted during model iteration based on work order acceptance data and recurrence tracking results, ensuring that the risk assessment continuously reflects actual operating conditions.

[0028] S3.5. Based on the drainage risk index of each drainage facility k calculated in S3.4, calculate the associated drainage risk index of road unit i. The associated drainage risk index R_i for each road unit i is the weighted average of the drainage risk indices of all drainage facilities within the catchment area of ​​each road unit, expressed as: Ri^road=Σ{k∈Φ(i)}γ_k·R_k, Where Φ(i) represents the set of drainage facilities within the catchment area of ​​road unit i; γ_k is the spatial weight coefficient, which is inversely proportional to the distance from the facility to the road unit. When the upstream storm drain of road unit P is blocked, the hydraulic gradient of the downstream pipe section is insufficient, or the pumping station is operating abnormally, the R_i^road value of road unit P increases.

[0029] S4. Based on the health level and disease type vector of the target road unit in step S2 and the drainage risk index in step S3, construct a causal relationship diagram of road-drainage hazards and deduce the dominant causes of hazards. For example... Figure 3 The diagram shown is a schematic representation of the construction of the road-drainage hazard causal relationship diagram of the present invention. The construction process of the road-drainage hazard causal relationship diagram is as follows: S4.1 Graph Structure Definition: Road units, storm drains, manholes, pipe sections, pumping stations, and work order events are defined as nodes (V), and spatial proximity, flow direction, upstream and downstream relationships within the pipeline network, and historical work order co-occurrence relationships are defined as edges (E). The structure of the road-drainage hazard causal relationship graph is represented as G = (V, E). Specifically, spatial proximity edges represent the connections between a road unit and drainage facilities within a 50m radius; flow direction edges represent storm drain paths determined by the Digital Elevation Model (DEM); pipeline network topology edges represent upstream and downstream connections determined by drainage network GIS data; and historical co-occurrence edges represent connections between nodes simultaneously involved in the same work order or maintenance operation.

[0030] S4.2 Edge Weight Initialization: The initial weight ω_e of each edge e is calculated using the formula: ω_e = ω_e^space · ω_e^topology · ω_e^history; where ω_e^space is the spatial proximity (the closer the distance, the higher the weight), using the actual distance parameter (within 50m) corresponding to the spatially adjacent edges in S4.1, the closer the distance, the higher the weight; ω_e^topology is the topological correlation, using the topological interval level n of the pipeline topology edge / convergence direction edge in S4.1 (the weight of direct upstream and downstream is 1, and the weight of interval n levels is 0.5^n); ω_e^history is the historical co-occurrence frequency, using the number of times the same work order co-occurred as counted by the historical co-occurrence edges in S4.1, and then normalized.

[0031] S4.3, Graph-based reasoning as the dominant cause: For the target road unit i and its associated drainage facility set, a message-passing algorithm based on Bayesian networks is used for causal reasoning to obtain the posterior probability of each candidate cause. Specifically, the health state H_i, the disease type vector D_i, and the drainage risk index R_i^road of road unit i are used as observation evidence. Belief propagation iterative calculation is performed along the graph structure, expressed as: P(cause_c|observation evidence) ∝ P(observation evidence|cause_c) · P(cause_c), where cause_c ∈ {drainage-dominant road disease, road structure-dominant drainage anomaly, construction disturbance-type composite hazard, and weather-triggered temporary hazard}. Typically, the cause with the highest posterior probability is selected as the dominant cause of the composite hazard. The sources of causes include the following three aspects: (1) Historical work order statistics: Based on the historical maintenance work orders, emergency repair work orders and acceptance records of the target area in the past 3 to 5 years, the proportion of each type of cause in all hidden danger events is statistically analyzed. For example, historical statistics of a certain area show that: drainage-dominated road defects account for 45%, road structure-dominated drainage anomalies account for 25%, construction disturbance-related composite hidden dangers account for 15%, and weather-triggered temporary hidden dangers account for 15%. The initial value of the prior probability is set accordingly.

[0032] (2) Quantification of expert experience: When historical data is insufficient or regional characteristics change significantly (such as after the construction of new areas or the renovation of pipelines), no less than 5 experts in the field of road maintenance and drainage engineering are invited to independently evaluate the prior probability of each cause using the Delphi method, and the geometric mean is taken as the prior value.

[0033] (3) Uniform prior (default): In extreme cases where historical data is lacking and expert evaluation is not feasible, no-information prior is adopted, that is, the prior probabilities of each cause are equal, and P(cause_c) = 0.25 is taken.

[0034] S4.4, Determination of Dominant Causes: The rules for determining dominant causes are used as an auxiliary judgment to estimate the dominant causes of hidden dangers. The specific rules are as follows: Drainage-dominant road defects: when R_i^road>0.6 and d4 (probability of drainage impact) in D_i>0.5, and the siltation rate of the downstream pipe section exceeds 30%; Drainage anomalies dominated by road structure: when R_i^road≤0.4 and d1 (probability of structural damage) in D_i>0.6; Construction disturbance-type composite hidden dangers: when there are recent (within 3 months) construction records within 200m of the road unit, and the rate of H_i decrease exceeds 5 minutes / month; Weather-triggered temporary hidden dangers: when the short-term forecast rainfall intensity exceeds 10mm / h, and the current H_i≥70 (yellow and above). As a specific implementation example: when a section of road experiences continuous subsidence around manholes and water accumulation after rain, and the siltation rate of its downstream pipe section exceeds 30%, the system will identify the hidden danger as a drainage-dominated road defect and prioritize the generation of a joint work order for dredging and manhole reinforcement.

[0035] S5. Based on the dominant causes of potential hazards, calculate a comprehensive risk score for the target road. Ranking the dominant risks (in descending order), output a coordinated response plan and generate a coordinated response work order. The comprehensive risk score calculation formula is: C_i = λ1·(100-H_i) + λ2·R_i^road; where λ1 + λ2 = 1 (λ1 = 0.5, λ2 = 0.5). Figure 4 The diagram shows the generation and closed-loop update process of the linkage response plan of the present invention. Specifically, the work order generation rules are as follows: For road units with a red health code (H_i<55) and a drainage risk index R_i^road>0.6, four types of work orders are generated: emergency drainage, traffic protection, storm drain cleaning, and temporary road surface repair, requiring a response within 4 hours; for road units with an orange health code (55≤H_i<70) and a recurrence frequency of ≥2 at the same location, a combination of pipeline inspection, manhole perimeter structure repair, and road surface milling and paving work orders are generated, requiring a response within 24 hours; for road units with a yellow health code (70≤H_i<85), a review inspection or planned maintenance suggestion is generated, requiring completion within 7 days; for road units with a green health code (H_i≥85), they are only included in the regular inspection plan, and no proactive work orders are generated. Furthermore, when generating work orders, the availability of operational resources (maintenance teams, machinery and equipment, and material inventory) and traffic impact (avoiding peak traffic hours and reducing lane closures) need to be considered. Therefore, a multi-objective optimization algorithm (NSGA-II) is used to optimize and constrain the execution order and resource allocation of work orders. The optimization objectives for the execution order and resource allocation of work orders include: minimizing the total processing time, minimizing the cost of operational resources, and minimizing the comprehensive traffic impact index. The quantification method for the optimization objectives of the execution order and resource allocation of work orders is as follows: calculate the minimum total processing time (T... total The formula is: T total = max(t completes m,n) - min(t start m,n), where m is the team number, n is the work order number, and t completes m,n and t start m,n are the times when the m-th team completes and starts the n-th work order, respectively; calculate the minimum operation resource consumption cost (C). total The formula is: C total =∑N n=1 (c_manpower_n ⋅T_operation_n + c_machinery_n ⋅T_operation_n + c_materials_n), where c_manpower_n and c_machinery_n are the unit time labor and machinery usage costs of work order n (yuan / h), c_materials_n is the one-time material consumption cost (yuan), and T_operation_n is the expected operation time (h); minimize the comprehensive traffic impact index (I traffic The formula is: I traffic =∑N n=1 (L n ⋅W n⋅ρ n ⋅T homework n / T 基准 ), where L n W represents the length of the road occupancy (m). n T represents the number of lane closures. 基准 As the baseline operation time (taken as 8 hours), ρ n The sensitivity coefficients for road classification are: expressway 1.5, arterial road 1.2, secondary arterial road 1.0, and local road 0.7.

[0036] Furthermore, the constraints on the execution order and resource allocation of work orders include: only one work order can be executed by the same team at any given time; the work order type must match the team's qualifications; the work order must be executed within the available work window; and material consumption must not exceed the current inventory. Therefore, the optimization method is obtained, and the process is as follows: First, each solution is encoded into two chromosomes: a process chromosome (work order execution order) and a resource chromosome (team allocation); then, through selection, crossover, and mutation iterations for 200 generations, retaining elite solutions in each generation; finally, a set of Pareto optimal solutions is obtained (a set of solutions where completion time, cost, and traffic impact are mutually advantageous). The scheduler selects the final solution from this set based on scenario preferences (priority time during flood season, priority cost during normal days, priority traffic impact on main roads). For example, the process chromosome encoding example: A→B→C indicates the work order of executing A first, then B, and finally C; the resource chromosome encoding example: Team 1→Team 2→Team 3 indicates that A is allocated to Team 1, B to Team 2, and C to Team 3. After 200 iterations using NSGA-II, as shown... Figure 5As shown, three Pareto optimal solutions are output: Solution 1: The execution order and resource allocation are carried out in parallel with three teams, A (team 1), B (team 2), and C (team 3). The total completion time is 4.0 hours, the resource cost is 6975 yuan, and the traffic impact index is 40.9. The corresponding optimal scenario is priority during the flood season (fastest completion). Solution 2: The execution order and resource allocation are carried out in the order of A→C→B, with each team assigned to team 1. The total completion time is 7.5 hours, the resource cost is 4200 yuan, and the traffic impact index is 28.5. The corresponding optimal scenario is priority during normal days (lowest cost). Solution 3: The execution order and resource allocation are carried out in parallel with teams A (team 1) and C (team 2), with team B following (team 1). The total completion time is 6.0 hours, the resource cost is 5100 yuan, and the traffic impact index is 18.2. The corresponding optimal scenario is priority for main roads (expressway C has the least traffic impact). In addition, during the flood season, dispatchers choose Option 1 to ensure that all three potential hazards are addressed as quickly as possible; when daily maintenance funds are tight, Option 2 is chosen, requiring only one team and consuming minimal resources; if work order C is located on an urban expressway, where traffic impact has the highest weight, then Option 3 is chosen, although the total completion time is 6 hours, the expressway occupancy time is reduced to 2 hours, minimizing traffic disruption. In a specific embodiment, assuming at 9:00 AM on a certain day, the system simultaneously generates three work orders to be handled: Work order A is for subsidence around a manhole on a main road (requiring a road repair team, estimated 4 hours), Work order B is for cleaning a storm drain inlet on an adjacent branch road (requiring a drainage and dredging team, estimated 1.5 hours), and Work order C is for pipeline inspection on an expressway (requiring a CCTV inspection team, estimated 2 hours), with a total of three dispatchable teams (each corresponding to a Class I qualification).

[0037] S6. Read and collect work order completion information, and update the road health code and drainage risk index for model iteration and optimization. Work order completion information includes: on-site photos, disposal material list, operation time, acceptance opinions, and recurrence tracking results. Updating the road health code and drainage risk index involves three parameters: health index, drainage risk index, and causal graph weights. The health index is updated based on the acceptance results, updating the road unit's health index H_i^new. If the acceptance is successful, H_i^new = min(100, H_i^old + ΔH_repair), where ΔH_repair is determined based on the repair area and repair quality (5~20 points). If the acceptance is unsuccessful, H_i^new = H_i^old (unchanged), and a second work order is triggered. The drainage risk index is updated based on the pipeline inspection results after dredging, clearing, or repair operations, updating parameters such as the siltation rate and fullness of drainage facilities, and recalculating R_k^new. The weights in the cause-effect graph are updated based on whether the same unit recurs within the observation period after the work order ends. If the same unit does not recur within the preset observation period (6 months) after treatment, the weight of the associated hazard is reduced (multiplied by 0.8); if it recurs, the contribution weight of the corresponding drainage facility or road structure node in the cause-effect graph is increased (multiplied by 1.2). Finally, each quarter, newly added work order data, acceptance data, and recurrence data are included in the training set to retrain the target detection model, optimize the AHP judgment matrix and entropy weight method weights, and achieve continuous iterative optimization of the model.

[0038] In summary, this invention provides a method for the coordinated handling of municipal hazards based on road health codes and hydraulic coupling of pipe networks. This method integrates closed-loop data from road health monitoring, drainage network operation, and maintenance work orders to dynamically assess the causal relationship between road defects and drainage hazards, and automatically generates cross-professional coordinated handling strategies. This improves the ability to identify the correlation between road defects and drainage hazards, reduces repeated repairs and the risk of flooding during the rainy season, and is applicable to collaborative maintenance scenarios for urban roads, manholes, storm drains, pipelines, and pumping stations, demonstrating significant advancements.

[0039] The above embodiments are merely illustrative of the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made to the technical solutions based on the technical concept proposed in this invention shall fall within the scope of protection of this invention.

Claims

1. A method for joint handling of municipal hidden dangers based on road health codes and pipeline hydraulic coupling, characterized in that, Includes the following steps: S1. Obtain the initial dataset of the target road unit, including road status data, drainage status data, environmental meteorological data, and historical treatment data; S2. Construct a road health code evaluation module based on the initial dataset of the target road unit, and generate the health level and disease type vector of the target road unit; S3. Construct the SWMM drainage network hydraulic coupling module and calculate the drainage risk index by associating the initial dataset of the target road unit; S4. Based on the health level and disease type vector of the target road unit in step S2 and the drainage risk index in step S3, construct a causal relationship diagram of road-drainage hazards and infer the dominant causes of hazards. S5. Based on the primary causes of hidden dangers, calculate and sort the comprehensive risk scores of the target roads in descending order, output the joint response plan, and generate work orders; S6. Collect work order information and update the road health code and drainage risk index for model iteration and optimization; work order information includes: on-site photos, list of disposal materials, operation time, acceptance opinions and recurrence tracking results.

2. The method for joint handling of municipal hidden dangers based on road health code and pipeline hydraulic coupling as described in claim 1, characterized in that, Road condition data includes one or more of the following: pavement images, potholes, cracks, subsidence, manhole damage, water accumulation on the road, sidewalk damage, curbstone damage, and abnormal data on road ancillary facilities.

3. The method for joint handling of municipal hidden dangers based on road health code and pipeline hydraulic coupling as described in claim 1, characterized in that, Drainage status data includes one or more of the following: storm drain blockage status, manhole water level, pipeline defects, pipeline siltation rate, pump station operation status, dredging records, and drainage emergency response records.

4. The method for joint handling of municipal hidden dangers based on road health code and pipeline hydraulic coupling as described in claim 1, characterized in that, The upper force model takes an initial dataset of associated target road units as input and calculates the drainage risk index under current or predicted rainfall conditions. The model inputs include: pipe segment geometric parameters, inspection well parameters, rainwater inlet parameters, pump station parameters, and boundary conditions.

5. The method for joint handling of municipal hidden dangers based on road health code and pipeline hydraulic coupling according to claim 1, characterized in that, The formula for calculating the drainage risk index is: R_k = β1·f_full(k) + β2·f_depth(k) + β3·f_silt(k) + β4·f_pump(k), Where f_full, f_depth, f_silt, and f_pump are the risk functions for pipe section fullness, manhole water depth, pipeline siltation rate, and pump station operation, respectively; k is the drainage facility number; and β is the weighting coefficient.

6. The method for joint handling of municipal hidden dangers based on road health code and pipeline hydraulic coupling according to claim 7, characterized in that, The associated drainage risk index R_i^ for each road unit i is the weighted average of the drainage risk indices of all drainage facilities within the catchment area of ​​each road unit, expressed as: Ri^road=Σ{k∈Φ(i)}γ_k·R_k, Where Φ(i) is the set of drainage facilities within the catchment area of ​​road unit i; γ_k is the spatial weight coefficient, which is inversely proportional to the distance from the facility to the road unit.

7. The method for joint handling of municipal hidden dangers based on road health code and pipeline hydraulic coupling according to claim 1, characterized in that, The road-drainage hazard causal relationship diagram is structured as follows: road units, storm drains, manholes, pipe sections, pumping stations, and work order events are used as nodes, and spatial proximity, flow direction, pipeline upstream and downstream, and historical work order co-occurrence relationships are used as edges.

8. The method for joint handling of municipal hidden dangers based on road health code and pipeline hydraulic coupling as described in claim 1, characterized in that, Update the road health code and drainage risk index to adjust the weights of corresponding nodes and associated edges of the same unit in the road-drainage hazard causal relationship graph within a preset observation period.

9. A method for joint handling of municipal hidden dangers based on road health code and pipeline hydraulic coupling as described in claim 1, characterized in that, The road health code evaluation module is used to calculate a comprehensive health index based on road grade, traffic volume, type of damage, area of ​​damage, depth of damage, frequency of recurrence, and historical maintenance results, and to classify road health into different levels.

10. A municipal hazard linkage system based on road health code and pipeline hydraulic coupling, characterized in that, It includes a data acquisition module, a road health code evaluation module, an SWMM drainage network hydraulic coupling module, a causal relationship graph reasoning module, a work order generation module, and a closed-loop update module. The data acquisition module is used to acquire the initial dataset of the target road unit. The road health code evaluation module is used to receive the initial dataset from the data acquisition module and generate the health level and disease type vector of the target road unit. The SWMM drainage network hydraulic coupling module is used to calculate the drainage risk index by associating the initial dataset of the target road unit. The causal relationship graph reasoning module constructs a causal relationship graph between roads and drainage hazards based on the health level of the road health code evaluation module and the drainage risk index of the SWMM drainage network hydraulic coupling module, and infers the dominant causes of hazards. The work order generation module outputs a joint response plan and generates a work order based on the dominant causes of hazards from the causal relationship graph reasoning module. The closed-loop update module is used to collect work order information from the work order generation module, update the road health code and drainage risk index, and perform model iterative optimization.

Images