Municipal construction safety dispatching management system based on cloud platform

Abstract

The application relates to the technical field of construction safety management, and discloses a municipal construction safety dispatching management system based on a cloud platform, which realizes online and advanced accurate prediction of construction disturbance risks of underground concealed pipelines by constructing a pipeline attribute-detection feature-risk level benchmark database and a modified pipeline disturbance risk advanced prediction physical model; the two action mechanisms of disturbance conduction and pipeline failure are distinguished through gradient disturbance tests, a disturbance conduction-pipeline failure classification model is established, and a risk correction coefficient is introduced, so that the false judgment defect caused by the fact that traditional monitoring can only give an alarm after the fact and cannot distinguish the failure stage is eliminated in principle; meanwhile, the physical model is constructed by fusing pipeline attribute parameters, defect features and the modified disturbance magnitude, the risk evolution trend can be output in advance for several hours, sufficient response time is reserved for construction dispatching, and the accuracy and advancement of risk prediction are significantly improved.

Description

Technical Field

[0001] This invention relates to the technical field of construction safety management, specifically a municipal construction safety dispatch and management system based on a cloud platform. Background Technology

[0002] In the field of urban municipal engineering construction, underground concealed pipelines (gas, electricity, water supply and drainage, etc.) are the core lifeline of urban operation. Excavation and disturbance during municipal construction can easily lead to safety accidents such as pipeline damage, media leakage, power outages and water outages, which are the core risk points for safety management in municipal construction. Currently, municipal construction is generally characterized by narrow and dispersed work areas, cross-operation of multiple sections and trades, lack of underground pipeline archives in old urban areas, and frequent dynamic changes in construction conditions. The accuracy, foresight, and coordination of construction safety scheduling directly determine the level of project safety management. The existing management model in the industry mainly relies on manual on-site inspections, single-point pipeline detection, and post-event video surveillance for traceability. Some projects have built basic cloud monitoring platforms to achieve centralized display of monitoring data, but a complete closed loop from risk prediction to scheduling execution has not yet been formed. Existing technologies have three prominent shortcomings in practical applications. The problems are threefold: First, underground concealed pipeline detection equipment is diverse in type and data heterogeneity, lacking a unified risk quantification model. Traditional monitoring can only provide post-event alarms after pipeline damage, failing to distinguish between disturbance propagation and pipeline failure mechanisms, resulting in high false alarm and missed alarm rates and insufficient response time for construction scheduling. Second, online monitoring data lacks a quantitative mapping relationship with construction scheduling strategies. Even if anomalies are detected, accurate scheduling plans cannot be automatically generated, relying solely on management experience for decision-making. This leads to significant scheduling delays and large deviations, making it impossible to achieve a balanced control of safety and progress. Third, frequent interventions during construction, such as cross-operations, capacity acceleration, and process switching, easily trigger superimposed disturbances and sudden risk changes. The existing control system lacks a feedforward compensation mechanism, failing to mitigate the safety impact of changes in operating conditions in advance, making it highly susceptible to exceeding safety thresholds and causing accidents. Existing publicly available technologies are mostly focused on on-site data collection and anomaly alarms, without establishing a linkage mechanism between risk prediction and construction scheduling, thus failing to fundamentally improve the precision of municipal construction safety scheduling.

[0003] Chinese invention patent CN119963387B, published on July 18, 2025, discloses an internet-based comprehensive safety management platform and method for railway tunnel construction personnel. Through strategies such as refined area management, real-time monitoring and early warning, rapid response and emergency dispatch, and efficient collaboration and resource optimization, it significantly improves the safety level of construction personnel. This platform and method, driven by data, scientifically assesses and classifies the safety levels of management areas, ensuring the rational allocation of resources. Simultaneously, by setting multi-level safety thresholds, it enables the timely detection and effective response to potential safety hazards. Whether it's rapid response in emergencies or efficient collaboration under normal circumstances, it greatly enhances the safety of railway tunnel construction and provides a good guarantee for the safety management of construction personnel. However, the above technical solution cannot intelligently detect underground concealed pipelines during railway construction, increasing the risks of railway construction. Summary of the Invention

[0004] (a) Technical problems to be solved To address the shortcomings of existing technologies in practical applications, such as the diverse types of underground concealed pipeline detection equipment, highly heterogeneous data, and the lack of a unified risk quantification model, traditional monitoring can only provide post-incident alarms after pipeline damage, failing to distinguish between disturbance propagation and pipeline failure mechanisms. This results in high false alarm and false negative rates, making it difficult to allow sufficient response time for construction scheduling. Furthermore, the lack of a quantitative mapping relationship between online monitoring data and construction scheduling strategies means that even when anomalies are detected, accurate scheduling plans cannot be automatically generated, relying solely on management experience for decision-making. This leads to significant scheduling delays and large deviations, hindering the balanced management of safety and progress. During construction, frequent interventions such as cross-operations, capacity acceleration, and process switching can easily trigger superimposed disturbances and sudden risk changes, and the existing control system lacks feedforward mechanisms. The compensation mechanism cannot mitigate the safety impact of changes in operating conditions in advance, and is prone to exceeding safety thresholds and causing accidents. Existing publicly available technologies are mostly focused on on-site data collection and anomaly alarms, without establishing a linkage mechanism between risk prediction and construction scheduling. This fails to fundamentally improve the precision of municipal construction safety scheduling. The solution is to construct a modified physical model for the advanced prediction of pipeline disturbance risks, enabling accurate and advanced prediction of disturbance risks from underground concealed pipeline construction, thereby eliminating the misjudgment defects of traditional monitoring in principle. The solution is to establish a risk-scheduling multi-dimensional mapping matrix to connect monitoring and scheduling in a closed-loop management system, replacing manual experience-based decision-making. The solution is to introduce a working condition intervention feedforward compensation mechanism to proactively mitigate the impact of changes in operating conditions, balancing construction safety and project progress, and significantly improving the precision of municipal construction safety scheduling.

[0005] (II) Technical Solution This invention is achieved through the following technical solution: a municipal construction safety dispatch and management system based on a cloud platform, the system comprising: The municipal pipeline safety benchmark database and prediction model construction module collects standard detection data and disturbance accident sample data of underground concealed pipelines under different geological and construction conditions, and establishes a benchmark database of pipeline attributes, detection characteristics and risk levels; it also collects disturbance transmission data and pipeline failure morphology data during multiple sets of gradient construction disturbance tests, constructs a disturbance transmission-pipeline failure classification model, obtains construction disturbance risk correction coefficients, and forms a physical model for advanced prediction of pipeline disturbance risks. The municipal construction risk-scheduling mapping matrix construction module collects pipeline detection data, construction process parameters, and work surface layout data of the current section during the construction preparation and trial operation phases. It calculates the benchmark risk value by combining the pipeline disturbance risk advance prediction physical model and binds the construction process parameters, equipment resource parameters, pipeline risk parameters, and scheduling strategies to construct a risk-scheduling multidimensional mapping matrix. The real-time monitoring and prediction module for construction disturbance risk collects pipeline disturbance monitoring data, construction condition data, and personnel and equipment location data of the current work face in real time during continuous construction, extracts disturbance characteristic parameters, and calls the pipeline disturbance risk advanced prediction physical model to calculate the real-time risk level and risk evolution trend. The construction safety scheduling closed-loop management module calculates the risk deviation of the real-time risk level relative to a preset safety threshold. When the risk deviation exceeds the dynamic adaptive early warning threshold, it calls the risk-scheduling multidimensional mapping matrix to calculate the adjustment amount of the construction scheduling strategy, sends it to the on-site construction execution terminal for closed-loop management, and synchronously updates the risk-scheduling multidimensional mapping matrix. When the risk deviation does not exceed the dynamic adaptive early warning threshold, it smoothly updates the current monitoring data to the risk-scheduling multidimensional mapping matrix. When a construction condition intervention feedforward scheduling module detects a construction condition intervention action, it calls the risk-scheduling multidimensional mapping matrix to calculate the expected risk fluctuation caused by the intervention action, and converts the expected risk fluctuation into a feedforward scheduling compensation signal and superimposes it onto the construction scheduling system.

[0006] Preferably, the specific steps in establishing the pipeline attribute-detection feature-risk level benchmark database include: collecting standard detection data and disturbance accident sample data of underground concealed pipelines under different geological and construction conditions; collecting disturbance transmission data and pipeline failure morphology data during multiple sets of gradient construction disturbance tests; constructing a disturbance transmission-pipeline failure classification model; obtaining construction disturbance risk correction coefficients; and forming a physical model for advanced prediction of pipeline disturbance risk. S101. Prepare gradient burial depth and gradient pipe diameter samples for three typical types of pipelines: gas, electricity, and water supply and drainage, as well as gradient defect degree samples including three typical defects: loose joints, pipe wall corrosion, and burial depth deviation. S102. Perform standard operating condition simulation disturbance tests on all samples, record the disturbance magnitude and risk level corresponding to the pipeline failure critical state, and label the characteristic response data of the corresponding detection equipment. S103. The pipeline attribute parameters, defect type and defect severity parameters, corresponding detection feature data, and standard risk level of the sample are stored in a structured manner to form the pipeline attribute-detection feature-risk level benchmark database.

[0007] Preferably, the specific steps in collecting standard detection data and disturbance accident sample data of underground concealed pipelines under different geological and construction conditions to establish a pipeline attribute-detection characteristic-risk level benchmark database; collecting disturbance transmission data and pipeline failure morphology data during multiple sets of gradient construction disturbance tests to construct a disturbance transmission-pipeline failure classification model, obtain construction disturbance risk correction coefficients, and form a pipeline disturbance risk advance prediction physical model include: S11. Set up multiple gradient construction disturbance test groups, each group corresponding to the first level of disturbance intensity, and conduct simulated excavation disturbance tests on standard pipeline samples of different types and burial depths, and simultaneously collect soil stress, pipeline deformation and damage morphology data at different time nodes. S12. Perform segmented fitting on the disturbance propagation curve, identify the curve abrupt inflection point, and combine the pipeline damage morphology data of the corresponding nodes to mark the stable deformation stage before the inflection point as the disturbance propagation-dominant stage, and mark the sudden damage stage after the inflection point as the pipeline failure-dominant stage, thus forming the disturbance propagation-pipeline failure classification model. S13. Using the standard risk threshold in the pipeline attribute-detection feature-risk level benchmark database as a benchmark, fit the quantitative relationship between the deformation rate and the standard risk level in the disturbance propagation dominant stage under different disturbance intensities, and combine the inflection point time correction amount in the pipeline failure dominant stage to calculate the risk correction coefficient under each disturbance gradient. S14. By integrating pipeline attribute parameters, detection feature parameters, and the corrected disturbance magnitude, a physical model for advanced prediction of pipeline disturbance risk is constructed.

[0008] Preferably, during the construction preparation and trial operation phase, pipeline detection data, construction process parameters, and work surface layout data for the current section are collected. Combined with the pipeline disturbance risk advance prediction physical model, a baseline risk value is calculated. Construction process parameters, equipment resource parameters, pipeline risk parameters, and scheduling strategies are bound together to construct a risk-scheduling multidimensional mapping matrix. When constructing the risk-scheduling multidimensional mapping matrix, a geological environment characteristic compensation step is also incorporated. Based on the soil type, groundwater level and surrounding building load parameters of the construction area, the soil disturbance transmission coefficient is calculated and used as a risk transmission correction factor to scale and correct the benchmark risk value in the risk-scheduling multidimensional mapping matrix, and then a corresponding matching scheduling strategy is applied.

[0009] Preferably, in the step of collecting pipeline disturbance monitoring data, construction condition data, and personnel and equipment location data of the current work face in real time during continuous construction, extracting disturbance characteristic parameters, and calling the pipeline disturbance risk advanced prediction physical model to calculate the real-time risk level and risk evolution trend, the specific steps for calculating the real-time risk level and risk evolution trend include: S31. During the current detection cycle, spatial location data of underground pipelines are collected using ground-penetrating radar and electromagnetic pipeline instrumentation, real-time disturbance data is collected using distributed optical fiber and soil stress gauge, and location data of construction machinery and personnel are collected using positioning equipment. S32. Perform spatiotemporal registration and feature extraction on multi-source detection data, identify pipeline types, and calculate disturbance intensity, disturbance distance, and deformation rate parameters to obtain disturbance characteristic parameters of the current working face; S33. Input the pipeline attribute parameters and disturbance characteristic parameters into the pipeline disturbance risk advance prediction physical model, and output the real-time risk level of the current working face and the risk evolution trend in the future period.

[0010] Preferably, the steps for calculating the risk deviation of the real-time risk level relative to a preset safety threshold, and when the risk deviation exceeds the dynamic adaptive early warning threshold, include: calling the risk-scheduling multidimensional mapping matrix to calculate the adjustment amount of the construction scheduling strategy, issuing it to the on-site construction execution terminal for closed-loop management, and synchronously updating the risk-scheduling multidimensional mapping matrix; and when the risk deviation does not exceed the dynamic adaptive early warning threshold, smoothly updating the current monitoring data to the risk-scheduling multidimensional mapping matrix. S41. Using a preset safety risk threshold as a benchmark value, calculate the ratio of the difference between the real-time risk level and the benchmark value to the benchmark value to obtain the risk deviation. S42. Determine whether the risk deviation is greater than the dynamic adaptive early warning threshold; if yes, proceed to S43; if no, proceed to S44. S43. Based on the risk impact weight of each scheduling element in the risk-scheduling multidimensional mapping matrix, prioritize the adjustment of high-weight core scheduling parameters, calculate the adjustment amount of the scheduling strategy for each element based on the risk deviation value and weight ratio, issue it to the corresponding field execution terminal for closed-loop management, and update the risk-scheduling multidimensional mapping matrix simultaneously. S44. Keep the current construction scheduling strategy unchanged, and use the preset exponential smoothing formula to smoothly update the current disturbance data, risk parameters and scheduling data to the risk-scheduling multidimensional mapping matrix.

[0011] Preferably, before the data is sent to the corresponding field execution terminal for closed-loop control in step S43, a scheduling boundary verification step is also included: S431. Read the target scheduling parameter value after the adjustment amount of the superimposed scheduling strategy; S432. Determine whether the target scheduling parameter value falls within the preset construction allowable range; if it exceeds the upper limit of the range, then forcibly truncate the scheduling strategy adjustment amount to the parameter value corresponding to the upper limit of the range; if it is lower than the lower limit of the range, then forcibly truncate the scheduling strategy adjustment amount to the parameter value corresponding to the lower limit of the range.

[0012] Preferably, when a construction condition intervention action is detected, the risk-scheduling multidimensional mapping matrix is ​​invoked to calculate the expected risk fluctuation caused by the intervention action, and the expected risk fluctuation is converted into a feedforward scheduling compensation signal and superimposed on the construction scheduling system. The intervention action includes cross-operation entry actions, and the specific steps include: S51. Real-time monitoring of cross-operation entry signals and parameters of newly added work surfaces through the construction scheduling system; S52. Extract the disturbance superposition coefficient of the corresponding work process from the risk-scheduling multidimensional mapping matrix, calculate the expected risk increase caused by cross-operations, and use it as the expected risk fluctuation amount. S53. Generate a corresponding feedforward scheduling compensation signal based on the expected risk fluctuation amount, and adjust the excavation rate, equipment path and personnel on-duty configuration in advance to offset the risk fluctuation caused by cross-operations.

[0013] Preferably, when a construction condition intervention action is detected, the risk-scheduling multidimensional mapping matrix is ​​invoked to calculate the expected risk fluctuation caused by the intervention action, and the expected risk fluctuation is converted into a feedforward scheduling compensation signal and superimposed on the construction scheduling system. The intervention action also includes a construction acceleration action. When a construction acceleration command is detected, the change in the magnitude of disturbance per unit time is calculated based on the excavation rate increase ratio. The change in the magnitude of disturbance is substituted into the risk-scheduling multidimensional mapping matrix to obtain the expected risk increase as the expected risk fluctuation, and corresponding speed reduction, current limiting, and encrypted monitoring feedforward compensation signals are generated and sent to the construction scheduling system.

[0014] (III) Beneficial Effects This invention provides a cloud-based municipal construction safety dispatch and management system. It has the following beneficial effects: I. By constructing a benchmark database of pipeline attributes, detection characteristics, and risk levels, and a modified physical model for advanced prediction of pipeline disturbance risks, online and accurate prediction of construction disturbance risks for underground concealed pipelines has been achieved. Through gradient disturbance tests, two mechanisms of action—disturbance propagation and pipeline failure—are distinguished. A disturbance propagation-pipeline failure classification model is established, and a risk correction coefficient is introduced. This fundamentally eliminates the shortcomings of traditional monitoring, which can only provide post-event alarms and cannot distinguish the failure stage, leading to misjudgments. Simultaneously, by integrating pipeline attribute parameters, defect characteristics, and the corrected disturbance magnitude to construct a physical model, the risk evolution trend can be output several hours in advance, allowing sufficient response time for construction scheduling and significantly improving the accuracy and anticipation of risk prediction.

[0015] Second, by establishing a risk-scheduling multidimensional mapping matrix, a closed-loop management system from risk monitoring to construction scheduling is achieved, breaking the traditional management model of "disconnection between monitoring and scheduling" in municipal construction. This matrix deeply binds construction process parameters, equipment resource parameters, pipeline risk parameters, and scheduling strategies, while introducing geological environmental characteristic compensation factors to correct the benchmark risk value, constructing a dynamic and self-learning multidimensional mapping relationship. When the risk deviation detected online exceeds the dynamic adaptive early warning threshold, the system can quickly calculate the adjustment amount of the scheduling strategy based on the mapping relationship and parameter weights, and directly issue it to the field execution end after boundary verification, simultaneously completing the matrix iteration update. This solves the pain point of traditional management "only measuring and not adjusting" and relying on manual experience for decision-making, realizing the automation and refinement of safety management.

[0016] Third, by introducing a feedforward compensation mechanism for work condition interventions, proactive prevention and stable management of risks during construction are achieved. For common work condition interventions in municipal construction, such as overlapping operations and accelerated construction, the system can capture intervention signals in real time, call the risk-scheduling multi-dimensional mapping matrix to calculate the expected risk fluctuations caused by the interventions in advance, and convert the expected fluctuations into feedforward scheduling compensation signals within the same cycle of the work condition action execution. This proactively adjusts construction parameters and resource allocation, mitigating the safety impacts of work condition changes in advance, avoiding the lag of traditional feedback control, effectively ensuring risk stability during continuous construction, and balancing construction safety and project progress. Attached Figure Description

[0017] Figure 1 A flowchart illustrating the operation method of the cloud-based municipal construction safety dispatch and management system provided by this invention; Figure 2 This is a schematic diagram of the modules of the cloud-based municipal construction safety dispatch and management system provided by the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] An example of the cloud-based municipal construction safety dispatch and management system is as follows: Example: Please see Figures 1-2 A cloud-based municipal construction safety dispatch and management system, comprising: The municipal pipeline safety benchmark library and prediction model construction module 01 collects standard detection data and disturbance accident sample data of underground concealed pipelines under different geological and construction conditions, and establishes a benchmark database of pipeline attributes, detection characteristics and risk levels; it also collects disturbance transmission data and pipeline failure morphology data during multiple sets of gradient construction disturbance tests, constructs a disturbance transmission-pipeline failure classification model, obtains construction disturbance risk correction coefficients, and forms a physical model for advanced prediction of pipeline disturbance risks. Module 02 of the Municipal Construction Risk-Dispatch Mapping Matrix Construction: During the construction preparation and trial operation phase, collect pipeline detection data, construction process parameters and work surface layout data of the current section, calculate the benchmark risk value by combining the pipeline disturbance risk advance prediction physical model, and bind the construction process parameters, equipment resource parameters, pipeline risk parameters and dispatch strategies to construct a risk-dispatch multidimensional mapping matrix; Module 03: During continuous construction, the module collects pipeline disturbance monitoring data, construction condition data and personnel and equipment location data of the current working face in real time online, extracts disturbance characteristic parameters, and calls the pipeline disturbance risk advanced prediction physical model to calculate the real-time risk level and risk evolution trend. The construction safety scheduling closed-loop management module 04 calculates the risk deviation of the real-time risk level relative to the preset safety threshold; when the risk deviation exceeds the dynamic adaptive early warning threshold, it calls the risk-scheduling multidimensional mapping matrix to calculate the adjustment amount of the construction scheduling strategy, sends it to the on-site construction execution terminal for closed-loop management, and updates the risk-scheduling multidimensional mapping matrix synchronously; when the risk deviation does not exceed the dynamic adaptive early warning threshold, it smoothly updates the current monitoring data to the risk-scheduling multidimensional mapping matrix. Construction Condition Intervention Feedforward Scheduling Module 05: When a construction condition intervention action is detected, the risk-scheduling multidimensional mapping matrix is ​​called to calculate the expected risk fluctuation caused by the intervention action, and the expected risk fluctuation is converted into a feedforward scheduling compensation signal and superimposed on the construction scheduling system.

[0020] Implementation Example Basic Configuration Description This example is based on a municipal engineering project for the reconstruction and expansion of a main road in the main urban area of ​​a city. The total construction length is 2.3km. Along the route are distributed DN300 high-pressure gas pipelines (design pressure 0.4MPa), 10kV power lines, and DN500 municipal water supply and drainage pipelines. The project adopts the open-cut construction method and involves three sections of work that are carried out simultaneously.

[0021] Hardware configuration Pipeline detection equipment: SIR-4000 ground-penetrating radar (center frequency 100MHz, detection depth 0-5m, depth resolution ±2cm), RD8000 electromagnetic pipeline instrument (positioning accuracy ±5cm, burial depth measurement accuracy ±3%). Disturbance monitoring equipment: distributed fiber optic vibration monitoring system (sampling frequency 100Hz, spatial positioning accuracy ±1m), vibrating wire soil stress gauge (range 0-2MPa, accuracy ±0.5%FS), resistance strain gauge pipeline deformation sensor (range ±50mm, accuracy ±0.1mm). Working conditions and positioning equipment: The excavator is equipped with a Beidou GPS positioning terminal (positioning accuracy ±10cm, data upload frequency 1Hz), a UWB personnel positioning base station (positioning accuracy ±30cm), and a laser ranging sensor for excavation depth (accuracy ±1cm). Cloud-edge collaborative equipment: On-site deployment of edge computing gateway (ARM architecture, 8-core CPU, 16GB memory, supports 4G / 5G dual-link transmission), and cloud platform deployed on the city government cloud (configured with 32-core CPU, 128GB memory, 2TB SSD storage, and adopts microservice architecture).

[0022] Software and data platform configuration Data storage: PostgreSQL relational database is used to store structured data such as pipeline attributes, risk levels, and scheduling strategies, while InfluxDB time series database is used to store time series data such as disturbance monitoring and operating condition acquisition. Algorithm Engine: The numerical computation and curve fitting engine is built based on Python, SciPy, and Matplotlib, and the risk prediction model is deployed on the cloud platform algorithm service layer; System integration: The system integrates with the on-site construction control system via the OPCUA protocol and with the urban pipeline ownership unit's supervision platform and traffic control platform via the API interface.

[0023] The specific steps for establishing the pipeline attribute-detection characteristic-risk level benchmark database include: collecting standard detection data and disturbance accident sample data of underground concealed pipelines under different geological and construction conditions; collecting disturbance transmission data and pipeline failure morphology data during multiple sets of gradient construction disturbance tests; constructing a disturbance transmission-pipeline failure classification model; obtaining construction disturbance risk correction coefficients; and forming a physical model for advanced prediction of pipeline disturbance risk. S101. Prepare gradient burial depth and gradient pipe diameter samples for three typical types of pipelines: gas, electricity, and water supply and drainage, as well as gradient defect degree samples including three typical defects: loose joints, pipe wall corrosion, and burial depth deviation. Specific implementation: Four sets of gradient burial depth samples were prepared at the test site, with burial depths of 1.0m, 1.5m, 2.0m, and 2.5m, corresponding to five pipe diameters: DN200, DN300, DN400, DN500, and DN600. For each type of pipeline, three types of defect samples were prepared simultaneously: pinhole defects were divided into four gradients based on area percentage (0.5%, 1%, 2%, and 5%), pipe wall corrosion was divided into three gradients based on wall thickness loss rate (10%, 20%, and 30%), and joint loosening was divided into three gradients based on joint displacement (2mm, 5mm, and 8mm). A total of 126 sets of valid test samples were formed.

[0024] S102. Perform standard operating condition simulation disturbance tests on all samples, record the disturbance magnitude and risk level corresponding to the pipeline failure critical state, and label the characteristic response data of the corresponding detection equipment. Specific implementation: Simulated excavation disturbance was applied using a tiered loading method, with each load level stabilized for 30 minutes, while simultaneously collecting soil stress and pipeline deformation data. Failure thresholds were defined according to the "Code for Construction and Acceptance of Urban Gas Transmission and Distribution Engineering" and the "Code for Comprehensive Planning of Urban Engineering Pipelines": gas pipeline interface leakage, electrical line insulation damage, and water supply and drainage pipe cracking and leakage constituted failure. Risks were divided into four levels: low risk (quantitative value 10 points), moderate risk (quantitative value 30 points), significant risk (quantitative value 60 points), and major risk (quantitative value 100 points). Peak soil vibration velocity, soil stress, pipeline deformation, and detection characteristic data such as ground-penetrating radar reflection amplitude, electromagnetic signal intensity, and fiber optic vibration energy were recorded for each risk level.

[0025] S103. The pipeline attribute parameters, defect type and defect severity parameters, corresponding detection feature data, and standard risk level of the samples are stored in a structured manner to form a pipeline attribute-detection feature-risk level benchmark database.

[0026] Specific implementation: The database fields include 11 categories, such as pipeline type, pipe diameter, burial depth, defect type, defect degree, soil vibration velocity, soil stress, pipeline deformation, radar reflection amplitude, electromagnetic signal strength, and risk level. A total of 126 sets of standard sample data are stored and deployed in the benchmark library module of the cloud platform, supporting retrieval and access by pipeline type and risk level.

[0027] The specific steps involved in collecting standard detection data and disturbance accident sample data of underground concealed pipelines under different geological and construction conditions to establish a pipeline attribute-detection characteristic-risk level benchmark database; collecting disturbance transmission data and pipeline failure morphology data from multiple sets of gradient construction disturbance tests to construct a disturbance transmission-pipeline failure classification model, obtaining construction disturbance risk correction coefficients, and forming a physical model for advanced prediction of pipeline disturbance risk. S11. Set up multiple gradient construction disturbance test groups, each group corresponding to the first level of disturbance intensity, and conduct simulated excavation disturbance tests on standard pipeline samples of different types and burial depths, and simultaneously collect soil stress, pipeline deformation and damage morphology data at different time nodes. Specific implementation: Four gradient disturbance test groups were set up for mild, moderate, severe, and extremely severe vibration, with corresponding peak vibration velocities of 0.5 cm / s, 1.5 cm / s, 2.5 cm / s, and 3.5 cm / s, respectively. Three parallel samples of DN300 defect-free gas pipelines were taken from each group and continuously loaded for 24 hours. Soil stress and pipeline deformation data were collected every hour, and the morphology of pipeline inner wall damage was observed every 6 hours using an industrial endoscope. Simultaneously, microscopic morphology photos of the pipe wall were taken.

[0028] S12. Perform segmented fitting on the disturbance propagation curve, identify the curve abrupt inflection point, and combine the pipeline damage morphology data of the corresponding nodes to mark the stable deformation stage before the inflection point as the disturbance propagation-dominant stage, and mark the sudden loss stage after the inflection point as the pipeline failure-dominant stage, thus forming a disturbance propagation-pipeline failure classification model. Specific implementation: Matlab was used to perform a cubic polynomial piecewise fitting on the pipeline deformation-time curve, and the inflection point of the curve abrupt change was identified by the zero-crossing point method of the second derivative. Taking the severe disturbance test group as an example, the fitting showed that the second derivative changed from positive to negative at the 12th hour, which was determined to be the abrupt inflection point. Combined with the endoscopic observation results, microcracks appeared at the pipeline interface and the deformation rate increased sharply at the 12th hour. Therefore, it was determined that the 0-12 hour period was the disturbance propagation-dominated stage (deformation increased linearly and slowly with the disturbance), and the 12-hour period was the pipeline failure-dominated stage (deformation increased exponentially). Based on this rule, a classification model was constructed, which can automatically identify the stage of any disturbance curve.

[0029] S13. Using the standard risk threshold in the pipeline attribute-detection feature-risk level benchmark database as a benchmark, fit the quantitative relationship between the deformation rate and the standard risk level in the disturbance propagation dominant stage under different disturbance intensities, and combine the inflection point time correction amount in the pipeline failure dominant stage to calculate the risk correction coefficient under each disturbance gradient. Specific implementation: Based on the Arrhenius perturbation propagation model, the deformation rate during the dominant phase of perturbation propagation is fitted. Risk level quantification value Relationship: ,in , The fitting coefficients are used; the pipeline failure inflection point time is also introduced. As a correction term, the risk correction factor is calculated using the following formula: Correction factor = Risk value in the transmission stage / (Risk value in the transmission stage + Risk increment in the failure stage) Based on the fitting calculation of 126 sets of samples, the risk correction coefficients for moderate, severe, and extremely severe disturbances were 0.95, 0.82, and 0.71, respectively, which were used to eliminate the conservative error in risk prediction caused by pipeline failure.

[0030] S14. By integrating pipeline attribute parameters, detection characteristic parameters, and the corrected disturbance magnitude, a physical model for advanced prediction of pipeline disturbance risk is constructed.

[0031] Specific implementation: The model input parameters are five features: pipeline diameter, burial depth, defect degree, soil vibration velocity, and disturbance distance. The output is a real-time risk level quantification value and a risk evolution curve for the next 4 hours. The model is deployed on the cloud platform algorithm layer, and the single-sample inference time is <100ms.

[0032] Verification results: 30 test samples that did not participate in the training were selected for verification. The accuracy of risk level prediction was ≥92%, and the relative error of 4-hour trend prediction was ≤8%, which meets the accuracy requirements of online scheduling.

[0033] By constructing a benchmark database of pipeline attributes, detection characteristics, and risk levels, and a modified physical model for advanced prediction of pipeline disturbance risks, online and accurate prediction of construction disturbance risks for underground concealed pipelines has been achieved. Through gradient disturbance tests, two mechanisms of action—disturbance propagation and pipeline failure—are distinguished, and a disturbance propagation-pipeline failure classification model is established, incorporating a risk correction coefficient. This fundamentally eliminates the shortcomings of traditional monitoring, which can only provide post-event alarms and cannot distinguish failure stages, leading to misjudgments. Simultaneously, by integrating pipeline attribute parameters, defect characteristics, and the corrected disturbance magnitude to construct a physical model, the risk evolution trend can be output several hours in advance, allowing sufficient response time for construction scheduling and significantly improving the accuracy and anticipation of risk prediction.

[0034] During the construction preparation and trial operation phase, pipeline detection data, construction process parameters, and work surface layout data for the current section are collected. A baseline risk value is calculated using a physical model for advanced prediction of pipeline disturbance risks. Construction procedure parameters, equipment resource parameters, pipeline risk parameters, and scheduling strategies are then bound together to construct a risk-scheduling multidimensional mapping matrix. Furthermore, geological environmental characteristic compensation steps are integrated into the construction of this risk-scheduling multidimensional mapping matrix. Based on the soil type, groundwater level and surrounding building load parameters of the construction area, the soil disturbance transmission coefficient is calculated and used as a risk transmission correction factor to scale and correct the benchmark risk value in the risk-scheduling multidimensional mapping matrix, and then the corresponding matching scheduling strategy is applied.

[0035] Specific implementation: During the construction preparation phase, ground-penetrating radar and electromagnetic pipeline surveying were used to conduct a detailed exploration of the entire pipeline, confirming that the DN300 gas pipeline was buried at a depth of 1.8m and at a minimum distance of 1.2m from the excavation edge, with no obvious defects. Steady-state trial operation conditions were set: excavation rate of 0.8m / h, single excavator operation, and 8 hours of construction per day. Disturbance monitoring data was collected continuously for 72 hours under these conditions, and the data were input into the predictive model to calculate a baseline risk level of 28 points.

[0036] Steps for incorporating geological environment characteristics during matrix construction: Based on the silty clay soil type in the construction area, the groundwater level depth of 3.2m, and the surrounding building load of 15kPa, calculate the soil disturbance transmission coefficient: Soil disturbance transmission coefficient = Soil shear modulus correction coefficient × Groundwater level influence coefficient × Building load reduction coefficient = 1.02 × 1.01 × 1.02 ≈ 1.05; Using this coefficient as a risk transmission correction factor, the baseline risk value is scaled and corrected. The corrected baseline risk value is 28 × 1.05 = 29.4 points. After rounding, the baseline risk quantification value for engineering applications is 30 points, corresponding to a general risk level.

[0037] Twelve construction procedure parameters across four categories—excavation rate, number of equipment, working distance, and working period—are structurally bound to revised risk parameters and corresponding scheduling strategies to construct a risk-scheduling multidimensional mapping matrix. The risk impact weights of each scheduling element are determined using orthogonal experimental design: excavation rate weight 0.5, working distance weight 0.3, and number of equipment weight 0.2. The matrix is ​​stored on a cloud platform, supporting rapid retrieval of corresponding scheduling schemes based on risk level.

[0038] During continuous construction, real-time online collection of pipeline disturbance monitoring data, construction condition data, and personnel and equipment location data at the current work site is performed. Disturbance characteristic parameters are extracted, and the real-time risk level and risk evolution trend are calculated using a pipeline disturbance risk advance prediction physical model. The specific steps for calculating the real-time risk level and risk evolution trend include: S31. During the current detection cycle, spatial location data of underground pipelines are collected using ground-penetrating radar and electromagnetic pipeline instrumentation, real-time disturbance data is collected using distributed optical fiber and soil stress gauge, and location data of construction machinery and personnel are collected using positioning equipment. Specific implementation: The system is set to a detection cycle of 5 minutes; the distributed fiber optic vibration sensor collects vibration data once per second, and calculates the average vibration velocity and peak value per cycle; the soil stress gauge collects stress data once every 30 seconds; the Beidou positioning data of the construction machinery is uploaded in real time, and the positioning data is uniformly converted into the CGCS2000 coordinate system; before each day's construction, the ground-penetrating radar is used to re-measure the work surface for the day and update the pipeline location data.

[0039] S32. Perform spatiotemporal registration and feature extraction on multi-source detection data, identify pipeline types, and calculate disturbance intensity, disturbance distance, and deformation rate parameters to obtain disturbance characteristic parameters of the current working face; Specific implementation: The timestamps of all equipment are unified through BeiDou time synchronization, with a time synchronization accuracy of ≤10ms; wavelet transform is used to denoise the vibration data and extract the peak vibration velocity and equivalent disturbance energy within a single cycle; the minimum horizontal distance (disturbance distance) between the excavation machinery and the pipeline is calculated through spatial coordinate matching; the deformation rate is calculated by differential calculation of deformation data from two consecutive cycles, and finally, a disturbance feature vector containing disturbance intensity, disturbance distance, and deformation rate is output.

[0040] S33. Input the pipeline attribute parameters and disturbance characteristic parameters into the pipeline disturbance risk advance prediction physical model, and output the real-time risk level of the current working face and the risk evolution trend in the future period.

[0041] Specific implementation: Data measured during a certain inspection cycle: the distance from the excavation face to the pipeline was 0.8m, the peak vibration velocity was 2.5cm / s, and the deformation rate was 0.15mm / h; after inputting into the model, the real-time risk level was calculated to be 62 points, corresponding to a relatively high risk level. It was predicted that the risk value would rise to 95 points after 2 hours of continuous operation under the current conditions, approaching the threshold of major risk.

[0042] By establishing a risk-scheduling multidimensional mapping matrix, a closed-loop management system from risk monitoring to construction scheduling is achieved, breaking the traditional management model of "disconnection between monitoring and scheduling" in municipal construction. This matrix deeply binds construction process parameters, equipment resource parameters, pipeline risk parameters, and scheduling strategies, while introducing geological environmental characteristic compensation factors to correct the baseline risk value, constructing a dynamic and self-learning multidimensional mapping relationship. When the risk deviation detected online exceeds the dynamic adaptive early warning threshold, the system can quickly calculate the adjustment amount of the scheduling strategy based on the mapping relationship and parameter weights, and directly issue it to the field execution end after boundary verification, simultaneously completing the matrix iteration update. This solves the pain points of traditional management that "only measures and does not adjust" and relies on manual experience for decision-making, realizing the automation and refinement of safety management.

[0043] Calculate the risk deviation of the real-time risk level relative to the preset safety threshold; when the risk deviation exceeds the dynamic adaptive early warning threshold, invoke the risk-scheduling multidimensional mapping matrix, calculate the adjustment amount of the construction scheduling strategy, and send it to the on-site construction execution terminal for closed-loop management, and synchronously update the risk-scheduling multidimensional mapping matrix; when the risk deviation does not exceed the dynamic adaptive early warning threshold, smoothly update the current monitoring data to the risk-scheduling multidimensional mapping matrix. The specific steps include: S41. Using a preset safety risk threshold as a benchmark, calculate the ratio of the difference between the real-time risk level and the benchmark value to the benchmark value to obtain the risk deviation. Calculation formula: Risk deviation = |Real-time risk quantification value - Benchmark risk quantification value| / Benchmark risk quantification value × 100%; In this embodiment, the baseline risk quantification value is 30 points, and the current real-time risk quantification value is 62 points. The calculated risk deviation is approximately 106.7% = |62-30| / 30×100%.

[0044] S42. Determine whether the risk deviation is greater than the dynamic adaptive early warning threshold; if yes, proceed to S43; if no, proceed to S44. Specific implementation: The dynamic adaptive early warning threshold is calculated using the 3σ criterion. A sample of risk deviation during normal construction periods over the past 30 days is taken, and the sample standard deviation σ = 10% is calculated. The early warning threshold is set to 3σ = 30%. If 106.7% > 30%, then S43 is executed, triggering the scheduling and control program; if 106.7% ≤ 30%, then S44 is executed. S43. Based on the risk impact weight of each scheduling element in the risk-scheduling multidimensional mapping matrix, prioritize the adjustment of high-weight core scheduling parameters, calculate the adjustment amount of the scheduling strategy for each element based on the risk deviation value and weight ratio, issue it to the corresponding field execution terminal for closed-loop management, and update the risk-scheduling multidimensional mapping matrix simultaneously. Specific implementation: Based on weight priority, the excavation rate is adjusted first (weight 0.5), followed by the working distance (weight 0.3); The formula for calculating the scheduling adjustment amount is as follows: Single parameter adjustment amount = risk deviation × parameter weight × unit adjustment step size The excavation rate is adjusted in increments of 0.1 m / h per 10% deviation, and the working distance is adjusted in increments of 0.1 m per 10% deviation. Calculations show that the excavation rate needs to be reduced by 0.4 m / h (from 0.8 m / h to 0.4 m / h), and the working distance needs to be increased by 0.3 m (from 0.8 m to 1.1 m).

[0045] The dispatch instructions are sent to the excavator's on-board terminal and the on-site dispatch room through the edge gateway, with a dispatch latency of <500ms. After the control is completed, the new working condition-risk corresponding data is written into the mapping matrix to complete the synchronous update.

[0046] Before S43 is sent to the corresponding field execution terminal for closed-loop control, a scheduling boundary verification step is also included: S431. Read the target scheduling parameter value after the adjustment amount of the superimposed scheduling strategy; S432. Determine whether the target scheduling parameter value falls within the preset construction allowable range; if it exceeds the upper limit of the range, then forcibly truncate the scheduling strategy adjustment amount to the parameter value corresponding to the upper limit of the range; if it is lower than the lower limit of the range, then forcibly truncate the scheduling strategy adjustment amount to the parameter value corresponding to the lower limit of the range.

[0047] Specific implementation: The preset allowable range for excavation rate is [0.2m / h, 1.2m / h], and the allowable range for working distance is [0.5m, 3.0m]. The calculated target excavation rate of 0.4m / h and target working distance of 1.1m both fall within the range. After verification, the operation is launched.

[0048] S44. Keep the current construction scheduling strategy unchanged, and use the preset exponential smoothing formula to smoothly update the current disturbance data, risk parameters and scheduling data to the risk-scheduling multidimensional mapping matrix.

[0049] Specific implementation: If the risk deviation for a certain period is 15%, which does not exceed the 30% threshold, it is judged as normal construction fluctuation; an exponential smoothing formula is used for updating. Updated baseline value = α × current measured value + (1-α) × historical baseline value The smoothing coefficient α = 0.05 enables the model to undergo slow adaptive iteration; after running continuously for 30 days, the fitting error of the baseline risk value is reduced by 12% compared with the initial value.

[0050] When a construction intervention action is detected, the risk-scheduling multidimensional mapping matrix is ​​invoked to calculate the expected risk fluctuation caused by the intervention action. This expected risk fluctuation is then converted into a feedforward scheduling compensation signal and superimposed onto the construction scheduling system. Intervention actions include cross-operation entry actions. Specific steps include: S51. Real-time monitoring of cross-operation entry signals and parameters of newly added work surfaces through the construction scheduling system; Specific implementation: The system connects to the project construction scheduling system through the API interface to obtain the work plan and entry instructions in real time; when the instruction of "pipeline relocation team enters the site" is detected, the feedforward calculation process is triggered to obtain the location of the new work surface, the type of work, and the number of equipment parameters.

[0051] S52. Extract the disturbance superposition coefficient of the corresponding work process from the risk-scheduling multidimensional mapping matrix, calculate the expected risk increase caused by cross-operations, and use it as the expected risk fluctuation. Specific implementation: Extract the double-workface disturbance superposition coefficient from the matrix as 1.3 (calibrated through multiple sets of parallel tests); the current benchmark risk value is 30 points, and the expected risk value is calculated as 30 × 1.3 = 39 points, with an expected risk increase of 9 points, corresponding to a 30% increase in risk deviation.

[0052] S53. Generate corresponding feedforward scheduling compensation signals based on the expected risk fluctuations, and adjust the excavation rate, equipment path and personnel on-duty configuration in advance to offset the risk fluctuations caused by cross-operations.

[0053] Specific implementation: A negative feedforward compensation signal is generated, and the excavation rate of the main working face is reduced by 30% in advance (from 0.8m / h to 0.56m / h). At the same time, one full-time safety officer is assigned to the cross-operation area. The compensation signal and the cross-operation entry command are issued simultaneously to achieve zero delay between changes in working conditions and compensation actions.

[0054] Implementation results: After the cross-operation began, the actual risk value stabilized at 32 points, with fluctuations controlled within 7%, and no warning was triggered.

[0055] When a construction condition intervention action is detected, the risk-scheduling multidimensional mapping matrix is ​​called to calculate the expected risk fluctuation caused by the intervention action. The expected risk fluctuation is then converted into a feedforward scheduling compensation signal and superimposed on the construction scheduling system. The intervention action also includes construction acceleration actions. When a construction acceleration command is detected, the change in the magnitude of disturbance per unit time is calculated based on the excavation rate increase ratio. The change in the magnitude of disturbance is then substituted into the risk-scheduling multidimensional mapping matrix to obtain the expected risk increase as the expected risk fluctuation, and corresponding speed reduction, current limiting, and encrypted monitoring feedforward compensation signals are generated and sent to the construction scheduling system.

[0056] Specific implementation: When the system receives an instruction to accelerate the construction schedule by 25%, the calculated disturbance level per unit time increases by 22%; substituting this into the mapping matrix, the expected risk value will rise to 46 points, with a deviation increase of 53%. A feedforward compensation combination signal is then generated: the actual excavation rate only increases by 10%, the monitoring frequency is increased to once per minute, and temporary soil reinforcement measures are added; the compensation signal is sent to the site one minute before the acceleration instruction is executed.

[0057] Implementation results: After the construction speed was accelerated, the actual risk value was stabilized within 35 points, and the risk fluctuation was controlled within 17%, which took into account both the construction period requirements and the safety management objectives.

[0058] By introducing a feedforward compensation mechanism for work condition interventions, proactive prevention and stable management of risks during construction are achieved. For common work condition interventions in municipal construction, such as overlapping operations and accelerated construction, the system can capture intervention signals in real time, call the risk-scheduling multi-dimensional mapping matrix to calculate the expected risk fluctuations caused by the interventions in advance, and convert these expected fluctuations into feedforward scheduling compensation signals within the same cycle of the work condition action execution. This proactively adjusts construction parameters and resource allocation, mitigating the safety impacts of work condition changes in advance, avoiding the lag of traditional feedback control, effectively ensuring risk stability during continuous construction, and balancing construction safety and project progress.

[0059] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A cloud-based municipal construction safety dispatch and management system, characterized in that, The system includes: The municipal pipeline safety benchmark database and prediction model construction module collects standard detection data and disturbance accident sample data of underground concealed pipelines under different geological and construction conditions, and establishes a benchmark database of pipeline attributes, detection characteristics and risk levels; it also collects disturbance transmission data and pipeline failure morphology data during multiple sets of gradient construction disturbance tests, constructs a disturbance transmission-pipeline failure classification model, obtains construction disturbance risk correction coefficients, and forms a physical model for advanced prediction of pipeline disturbance risks. The municipal construction risk-scheduling mapping matrix construction module collects pipeline detection data, construction process parameters, and work surface layout data of the current section during the construction preparation and trial operation phases. It calculates the benchmark risk value by combining the pipeline disturbance risk advance prediction physical model and binds the construction process parameters, equipment resource parameters, pipeline risk parameters, and scheduling strategies to construct a risk-scheduling multidimensional mapping matrix. The real-time monitoring and prediction module for construction disturbance risk collects pipeline disturbance monitoring data, construction condition data, and personnel and equipment location data of the current work face in real time during continuous construction, extracts disturbance characteristic parameters, and calls the pipeline disturbance risk advanced prediction physical model to calculate the real-time risk level and risk evolution trend. The construction safety scheduling closed-loop management module calculates the risk deviation of the real-time risk level relative to a preset safety threshold. When the risk deviation exceeds the dynamic adaptive early warning threshold, it calls the risk-scheduling multidimensional mapping matrix to calculate the adjustment amount of the construction scheduling strategy, sends it to the on-site construction execution terminal for closed-loop management, and synchronously updates the risk-scheduling multidimensional mapping matrix. When the risk deviation does not exceed the dynamic adaptive early warning threshold, it smoothly updates the current monitoring data to the risk-scheduling multidimensional mapping matrix. When a construction condition intervention feedforward scheduling module detects a construction condition intervention action, it calls the risk-scheduling multidimensional mapping matrix to calculate the expected risk fluctuation caused by the intervention action, and converts the expected risk fluctuation into a feedforward scheduling compensation signal and superimposes it onto the construction scheduling system.

2. The municipal construction safety dispatch and management system based on a cloud platform according to claim 1, characterized in that, The specific steps involved in establishing the pipeline attribute-detection characteristic-risk level benchmark database include: collecting standard detection data and disturbance accident sample data of underground concealed pipelines under different geological and construction conditions; collecting disturbance transmission data and pipeline failure morphology data during multiple sets of gradient construction disturbance tests; constructing a disturbance transmission-pipeline failure classification model; obtaining construction disturbance risk correction coefficients; and forming a physical model for advanced prediction of pipeline disturbance risk. S101. Prepare gradient burial depth and gradient pipe diameter samples for three typical types of pipelines: gas, electricity, and water supply and drainage, as well as gradient defect degree samples including three typical defects: loose joints, pipe wall corrosion, and burial depth deviation. S102. Perform standard operating condition simulation disturbance tests on all samples, record the disturbance magnitude and risk level corresponding to the pipeline failure critical state, and label the characteristic response data of the corresponding detection equipment. S103. The pipeline attribute parameters, defect type and defect severity parameters, corresponding detection feature data, and standard risk level of the sample are stored in a structured manner to form the pipeline attribute-detection feature-risk level benchmark database.

3. The municipal construction safety dispatch and management system based on a cloud platform according to claim 1, characterized in that, The specific steps involved in collecting standard detection data and disturbance accident sample data of underground concealed pipelines under different geological and construction conditions to establish a pipeline attribute-detection characteristic-risk level benchmark database; collecting disturbance transmission data and pipeline failure morphology data during multiple sets of gradient construction disturbance tests to construct a disturbance transmission-pipeline failure classification model, obtain construction disturbance risk correction coefficients, and form a physical model for advanced prediction of pipeline disturbance risk include: S11. Set up multiple gradient construction disturbance test groups, each group corresponding to the first level of disturbance intensity, and conduct simulated excavation disturbance tests on standard pipeline samples of different types and burial depths, and simultaneously collect soil stress, pipeline deformation and damage morphology data at different time nodes. S12. Perform segmented fitting on the disturbance propagation curve, identify the curve abrupt inflection point, and combine the pipeline damage morphology data of the corresponding nodes to mark the stable deformation stage before the inflection point as the disturbance propagation-dominant stage, and mark the sudden damage stage after the inflection point as the pipeline failure-dominant stage, thus forming the disturbance propagation-pipeline failure classification model. S13. Using the standard risk threshold in the pipeline attribute-detection feature-risk level benchmark database as a benchmark, fit the quantitative relationship between the deformation rate and the standard risk level in the disturbance propagation dominant stage under different disturbance intensities, and combine the inflection point time correction amount in the pipeline failure dominant stage to calculate the risk correction coefficient under each disturbance gradient. S14. By integrating pipeline attribute parameters, detection feature parameters, and the corrected disturbance magnitude, a physical model for advanced prediction of pipeline disturbance risk is constructed.

4. The municipal construction safety dispatch and management system based on a cloud platform according to claim 1, characterized in that, During the construction preparation and trial operation phase, pipeline detection data, construction process parameters, and work surface layout data for the current section are collected. Combined with the pipeline disturbance risk advance prediction physical model, a baseline risk value is calculated. Construction process parameters, equipment resource parameters, pipeline risk parameters, and scheduling strategies are bound together to construct a risk-scheduling multidimensional mapping matrix. When constructing the risk-scheduling multidimensional mapping matrix, a geological environment characteristic compensation step is also incorporated. Based on the soil type, groundwater level and surrounding building load parameters of the construction area, the soil disturbance transmission coefficient is calculated and used as a risk transmission correction factor to scale and correct the benchmark risk value in the risk-scheduling multidimensional mapping matrix, and then a corresponding matching scheduling strategy is applied.

5. The municipal construction safety dispatch and management system based on a cloud platform according to claim 1, characterized in that, The specific steps for calculating the real-time risk level and risk evolution trend during continuous construction, including online real-time collection of pipeline disturbance monitoring data, construction condition data, and personnel and equipment location data at the current work site, extraction of disturbance characteristic parameters, and calculation of the real-time risk level and risk evolution trend using the pipeline disturbance risk advanced prediction physical model, include: S31. During the current detection cycle, spatial location data of underground pipelines are collected using ground-penetrating radar and electromagnetic pipeline instrumentation, real-time disturbance data is collected using distributed optical fiber and soil stress gauge, and location data of construction machinery and personnel are collected using positioning equipment. S32. Perform spatiotemporal registration and feature extraction on multi-source detection data, identify pipeline types, and calculate disturbance intensity, disturbance distance, and deformation rate parameters to obtain disturbance characteristic parameters of the current working face; S33. Input the pipeline attribute parameters and disturbance characteristic parameters into the pipeline disturbance risk advance prediction physical model, and output the real-time risk level of the current working face and the risk evolution trend in the future period.

6. The municipal construction safety dispatch and management system based on a cloud platform according to claim 1, characterized in that, The calculation of the risk deviation of the real-time risk level relative to the preset safety threshold; when the risk deviation exceeds the dynamic adaptive early warning threshold, the risk-scheduling multidimensional mapping matrix is ​​invoked to calculate the adjustment amount of the construction scheduling strategy, which is then sent to the on-site construction execution terminal for closed-loop management, and the risk-scheduling multidimensional mapping matrix is ​​updated synchronously. When the risk deviation does not exceed the dynamic adaptive early warning threshold, the specific steps for smoothly updating the current monitoring data to the risk-scheduling multidimensional mapping matrix include: S41. Using a preset safety risk threshold as a benchmark value, calculate the ratio of the difference between the real-time risk level and the benchmark value to the benchmark value to obtain the risk deviation. S42. Determine whether the risk deviation is greater than the dynamic adaptive early warning threshold; if yes, proceed to S43; if no, proceed to S44. S43. Based on the risk impact weight of each scheduling element in the risk-scheduling multidimensional mapping matrix, prioritize the adjustment of high-weight core scheduling parameters, calculate the adjustment amount of the scheduling strategy for each element based on the risk deviation value and weight ratio, issue it to the corresponding field execution terminal for closed-loop management, and update the risk-scheduling multidimensional mapping matrix simultaneously. S44. Keep the current construction scheduling strategy unchanged, and use the preset exponential smoothing formula to smoothly update the current disturbance data, risk parameters and scheduling data to the risk-scheduling multidimensional mapping matrix.

7. The municipal construction safety dispatch and management system based on a cloud platform according to claim 6, characterized in that, Before the S43 statement is sent to the corresponding field execution terminal for closed-loop control, a scheduling boundary verification step is also included: S431. Read the target scheduling parameter value after the adjustment amount of the superimposed scheduling strategy; S432. Determine whether the target scheduling parameter value falls within the preset construction allowable range; If the value exceeds the upper limit of the interval, the scheduling policy adjustment will be forcibly truncated to the parameter value corresponding to the upper limit of the interval; if the value is below the lower limit of the interval, the scheduling policy adjustment will be forcibly truncated to the parameter value corresponding to the lower limit of the interval.

8. The municipal construction safety dispatch and management system based on a cloud platform according to claim 1, characterized in that, When a construction condition intervention action is detected, the risk-scheduling multidimensional mapping matrix is ​​invoked to calculate the expected risk fluctuation caused by the intervention action. The expected risk fluctuation is then converted into a feedforward scheduling compensation signal and superimposed onto the construction scheduling system. The intervention action includes cross-operation entry actions. The specific steps include: S51. Real-time monitoring of cross-operation entry signals and parameters of newly added work surfaces through the construction scheduling system; S52. Extract the disturbance superposition coefficient of the corresponding work process from the risk-scheduling multidimensional mapping matrix, calculate the expected risk increase caused by cross-operations, and use it as the expected risk fluctuation amount. S53. Generate a corresponding feedforward scheduling compensation signal based on the expected risk fluctuation amount, and adjust the excavation rate, equipment path and personnel on-duty configuration in advance to offset the risk fluctuation caused by cross-operations.

9. The municipal construction safety dispatch and management system based on a cloud platform according to claim 1, characterized in that, When a construction condition intervention action is detected, the risk-scheduling multidimensional mapping matrix is ​​called to calculate the expected risk fluctuation caused by the intervention action, and the expected risk fluctuation is converted into a feedforward scheduling compensation signal and superimposed on the construction scheduling system. The intervention action also includes a construction acceleration action. When a construction acceleration command is detected, the change in the magnitude of disturbance per unit time is calculated based on the excavation rate increase ratio. The change in the magnitude of disturbance is substituted into the risk-scheduling multidimensional mapping matrix to obtain the expected risk increase as the expected risk fluctuation, and corresponding speed reduction, current limiting, and encrypted monitoring feedforward compensation signals are generated and sent to the construction scheduling system.

Images