Construction safety monitoring method and equipment for heat distribution pipeline laying and medium

Abstract

The embodiment of the invention discloses a construction safety monitoring method and device for heat distribution pipeline laying and a medium, belongs to the technical field of pipeline laying, and solves the problem that a monitoring method for the heat distribution pipeline laying process is low in precision. Real-time construction data in the heat distribution pipeline laying process are input into the heat distribution pipeline laying digital twinborn model, and the influence condition of the real-time construction data on the digital twinborn model is obtained; based on the influence condition, performing multi-level division on a space region corresponding to the digital twin model for laying the heat distribution pipeline to obtain a plurality of non-uniform sub-region units; based on historical heat distribution pipeline laying data and real-time construction data, risk data prediction is carried out on the heat conduction risk and the load risk of each non-uniform subarea unit; and through a time sliding window, time sequence construction feature extraction and time sequence risk prediction feature extraction are performed on each sub-region unit, and the extracted features are compared with a preset safety evaluation index system, so that construction safety monitoring of heat distribution pipeline laying is realized.

Description

Technical Field

[0001] This application relates to the field of pipeline laying monitoring technology, and in particular to a construction safety monitoring method, equipment and medium for laying thermal pipelines. Background Technology

[0002] In the laying of heating pipelines, the safety of pipeline operation and the standardization of the construction process are directly related to the stable operation of the heating system. Construction risks during pipeline laying can easily lead to safety accidents such as pipeline rupture and leakage. Therefore, it is necessary to conduct precise safety monitoring of the entire construction process.

[0003] In existing technologies, the construction safety of heating pipelines is often monitored by combining manual inspections with fixed thresholds. However, manual inspections are time-consuming and have blind spots for pipelines that are deeply buried or cross complex geological areas. Early hazards such as minor leaks and water ingress into the insulation layer are difficult to detect in time and are usually dealt with passively after an accident occurs.

[0004] To compensate for the shortcomings of manual inspections, some thermal pipeline monitoring systems have introduced sensor data collection. However, existing sensor monitoring relies solely on fixed threshold comparisons for data processing, making it difficult to obtain the hidden risk correlations behind parameter fluctuations. This results in low accuracy of early warnings and high false alarm and missed alarm rates, making it difficult to provide reliable support for construction safety decisions. Summary of the Invention

[0005] This application provides a construction safety monitoring method, equipment, and medium for laying thermal pipelines, which solves the following technical problems: existing sensor monitoring methods rely solely on fixed threshold comparisons for data processing during the laying of thermal pipelines, making it difficult to obtain the hidden risk correlations behind parameter fluctuations, resulting in low early warning accuracy and high false alarm and missed alarm rates.

[0006] The embodiments of this application adopt the following technical solutions: This application provides a method for monitoring the construction safety of thermal pipeline laying. The method includes: inputting real-time construction data during the thermal pipeline laying process into a digital twin model of the thermal pipeline laying; based on the data transmission relationship, obtaining the impact of real-time construction data on the digital twin model of the thermal pipeline laying; based on the impact, dividing the spatial region corresponding to the digital twin model of the thermal pipeline laying into multiple non-uniform sub-region units; based on historical thermal pipeline laying data and real-time construction data, predicting the heat conduction risk and load risk of each non-uniform sub-region unit; wherein the risk data includes at least one of the risk prediction level and the core risk prediction cause; and through a time sliding window, extracting time-series construction features and time-series risk prediction features from the real-time construction data and risk data corresponding to each sub-region unit, respectively, and comparing the extracted features with a pre-set safety evaluation index system to achieve construction safety monitoring of the thermal pipeline laying.

[0007] In one implementation of this application, before obtaining the impact of real-time construction data on the digital twin model of thermal pipeline laying based on the construction data transmission relationship, the method further includes: obtaining the graph model nodes corresponding to the digital twin model of thermal pipeline laying based on the multi-dimensional basic data corresponding to the thermal pipeline laying area; extracting the phased related sub-models sequentially from the graph model nodes based on different construction stages in the thermal pipeline laying process; wherein the output node of the previous related sub-model serves as the input node of the next related sub-model; adjusting the edge weights of the phased related sub-models based on the real-time construction data corresponding to different construction stages to obtain the construction transmission relationship; wherein the real-time construction data is related to at least one of the following: pipeline laying equipment, pipeline monitoring data, operation data, environmental dynamic data, and process requirements.

[0008] In one implementation of this application, based on the impact status, the spatial region corresponding to the digital twin model of the thermal pipeline laying is divided into multiple levels to obtain multiple non-uniform sub-region units. Specifically, this includes: initially dividing the spatial region corresponding to the digital twin model of the thermal pipeline laying based on geological data and thermal pipeline data obtained during the thermal pipeline laying process; determining multiple impact features based on the impact status; wherein the multiple impact features include at least one of thermal impact features and mechanical disturbance features; configuring corresponding regional densification threshold groups for each impact feature; wherein the regional densification threshold groups include different densification thresholds and densification algorithms corresponding to each densification threshold; comparing the real-time quantized values ​​of each impact feature with the corresponding regional densification threshold groups; and based on the comparison results, performing multi-level division processing on the spatial region corresponding to the initially divided digital twin model of the thermal pipeline laying to obtain multiple non-uniform sub-region units; wherein each non-uniform sub-region unit corresponds to a different risk sensitivity level.

[0009] In one implementation of this application, based on historical thermal pipeline laying data and its impact, the heat conduction risk and load risk of each non-uniform sub-region unit are predicted. Specifically, this includes: collecting historical heat conduction data and historical load data corresponding to the historical thermal pipeline laying process; the historical heat conduction data and historical load data are the corresponding data obtained within the most recent preset time period; fusing the historical heat conduction data, historical load data, and the correlation coefficient between the current thermal pipeline laying and historical risk events to generate composite features; and predicting the core risk prediction causes corresponding to each non-uniform sub-region unit by using the conditional probabilities and composite features between preset Bayesian network nodes; based on historical... Using heat conduction data and historical load data, initial heat conduction risk thresholds and initial load risk thresholds are determined, and an initial threshold matrix is ​​constructed. The initial threshold matrix is ​​optimized based on a preset risk-sensitive phase factor, and the risk threshold corresponding to each sub-region unit is determined through the optimized matrix. Event labels are set for historical heat conduction data and historical load data, and a heat conduction risk probability prediction model and a load value prediction model are constructed based on the event labels. The prediction results output by the heat conduction risk probability prediction model and the load value prediction model are compared with the risk thresholds, and the risk prediction level corresponding to each non-uniform sub-region unit is determined based on the difference between the prediction results and the risk thresholds.

[0010] In one implementation of this application, an initial heat conduction risk threshold and an initial load risk threshold are determined based on historical heat conduction data and historical load data, and an initial threshold matrix is ​​constructed. Specifically, this includes: inputting historical heat conduction data and historical load data into preset analytical expressions for heat conduction risk threshold and load risk threshold, respectively; generating an initial heat conduction risk threshold correspondence table and an initial load threshold correspondence table based on the calculation results; determining multiple first indicator sets based on historical heat conduction data; wherein each of the multiple first indicator sets includes at least one of a peak heat flux density index, a thermal diffusivity attenuation index, and a temperature gradient uniformity index; determining multiple second indicator sets based on historical load data; wherein each of the multiple second indicator sets includes at least one of a peak soil pressure index, a stress concentration factor index, and a load duration index; constructing a matrix framework using the first indicator sets as the matrix row dimension and the second indicator sets as the matrix column dimension; and filling the data from the initial heat conduction risk threshold correspondence table and the initial load threshold correspondence table into the corresponding element positions in the matrix framework to generate the initial threshold matrix.

[0011] In one implementation of this application, the initial threshold matrix is ​​optimized based on a preset risk-sensitive phase factor. Specifically, this includes: determining key parameters affecting threshold offset based on the risk characteristics of heat conduction risk and load risk; obtaining the risk-sensitive phase factor based on the correlation between the risk offset and the key parameters; determining the sub-regional units corresponding to heat conduction risk and load risk respectively, and matching the corresponding risk thresholds in the initial threshold matrix through the sub-regional units; inputting the risk-sensitive phase factor and the corresponding risk thresholds into a preset threshold offset calculation function, and optimizing the initial threshold matrix based on the calculation results.

[0012] In one implementation of this application, temporal construction feature extraction and temporal risk prediction feature extraction are performed on each non-uniform sub-region unit. The extracted features are compared with a pre-set safety evaluation index system to achieve construction safety monitoring. Specifically, this includes: dividing the construction data and risk data corresponding to each non-uniform sub-region unit into multiple datasets according to different construction types; dividing each dataset into multiple phased data temporal subsets based on the construction stages corresponding to each construction type; extracting risk features from multiple phased data temporal subsets through a sliding time window and comparing the extracted risk features with the pre-set safety evaluation index system; triggering a working condition safety upgrade command when any risk feature value does not meet the safety threshold corresponding to the safety evaluation index system; and determining the mapping relationship between the working condition type and construction risk based on the working condition safety upgrade command to obtain construction risk information.

[0013] In one implementation of this application, based on the working condition safety upgrade instruction, the mapping relationship between the working condition type and the construction risk is determined to obtain construction risk information. Specifically, this includes: obtaining instruction element information corresponding to the working condition safety upgrade instruction; wherein, the instruction element information includes at least one of the following: current construction stage, construction type, actual feature value, and safety threshold; determining the working condition type based on the difference between the actual feature value and the safety threshold; performing feature matching in the historical working condition information database based on the working condition type and the instruction element information to obtain historical similar working condition information; and determining the construction risk information based on the mapping relationship between the historical similar working condition information and the construction risk information database.

[0014] This application provides a construction safety monitoring method for laying thermal pipelines, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to: input real-time construction data during the laying process of the thermal pipeline into a digital twin model of the thermal pipeline; based on the data transmission relationship, obtain the impact of the real-time construction data on the digital twin model of the thermal pipeline; based on the impact, divide the spatial region corresponding to the digital twin model of the thermal pipeline into multiple non-uniform sub-region units; based on historical thermal pipeline laying data and real-time construction data, predict the heat conduction risk and load risk of each non-uniform sub-region unit; wherein the risk data includes at least one of the risk prediction level and the core risk prediction cause; and through a time sliding window, extract time-series construction features and time-series risk prediction features from the real-time construction data and risk data corresponding to each sub-region unit, respectively, and compare the extracted features with a preset safety evaluation index system to achieve construction safety monitoring of the thermal pipeline.

[0015] This application provides a non-volatile computer storage medium storing computer-executable instructions. These instructions are configured to: input real-time construction data during the laying of a thermal pipeline into a digital twin model of the thermal pipeline; based on the data transmission relationship, obtain the impact of the real-time construction data on the digital twin model of the thermal pipeline; based on the impact, divide the spatial region corresponding to the digital twin model of the thermal pipeline into multiple levels to obtain multiple non-uniform sub-region units; based on historical thermal pipeline laying data and real-time construction data, predict the heat conduction risk and load risk of each non-uniform sub-region unit; wherein the risk data includes at least one of the risk prediction level and the core risk prediction cause; through a time sliding window, extract time-series construction features and time-series risk prediction features from the real-time construction data and risk data corresponding to each sub-region unit, respectively; and compare the extracted features with a pre-set safety evaluation index system to achieve construction safety monitoring of the thermal pipeline laying.

[0016] The above-mentioned technical solutions adopted in this application embodiment can achieve the following beneficial effects: This application embodiment inputs real-time construction data into the model and simulates the impact situation. Based on the transmission relationship of construction data, it dynamically presents the chain impact of construction operations on pipelines and the surrounding environment, solving the problem that traditional monitoring is difficult to predict risk transmission. By dividing non-uniform sub-region units, it achieves a differentiated division of high-risk areas with refined data and low-risk areas with simplified data, avoiding the defects of insufficient monitoring accuracy or waste of resources in uniform sub-region units. In addition, this application embodiment predicts the heat conduction and load risks of non-uniform sub-region units based on historical data, and outputs results by combining risk level and core causes. It improves the accuracy of risk prediction through historical experience data, predicts the source of hidden dangers in advance, and achieves real-time risk verification through a dynamic indicator system, solving the problem of high false alarm and missed alarm rates in traditional fixed threshold early warning, and improving the accuracy and timeliness of construction safety monitoring. Attached Figure Description To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings: Figure 1 A flowchart illustrating a construction safety monitoring method for laying thermal pipelines, provided in this application embodiment; Figure 2 This is a structural schematic diagram of a construction safety monitoring device for laying thermal pipelines, provided as an embodiment of this application.

[0017] Figure label: 200: Construction safety monitoring equipment for laying thermal pipelines; 201: Processor; 202: Memory. Detailed Implementation

[0018] This application provides a method, equipment, and medium for monitoring the construction safety of thermal pipeline laying.

[0019] To enable those skilled in the art to better understand the technical solutions in this application, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this application.

[0020] The technical solutions proposed in the embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0021] Figure 1 A flowchart illustrating a construction safety monitoring method for laying thermal pipelines, as provided in this application embodiment, is shown below. Figure 1 As shown, the construction safety monitoring method for laying heating pipelines includes the following steps: Step 101: Input the real-time construction data during the laying of the heating pipeline into the digital twin model. Based on the data transmission relationship, simulate the impact of the real-time construction data on the digital twin model of the heating pipeline laying.

[0022] In one implementation of this application, geological data can be acquired by geological data acquisition sensors installed in the area where the heating pipeline is laid. This geological data includes soil stratification information of the construction area, such as the distribution depth and thickness of soft soil, sand, and rock; soil physical and mechanical parameters, such as thermal conductivity, bearing capacity, and compression modulus; and groundwater depth and historical geological disaster records obtained through geological exploration drilling, in-situ testing, and regional geological databases. Secondly, in this embodiment, pipeline data obtained from pipeline design drawings, material supplier technical documents, and construction specifications includes pipeline material, diameter, wall thickness, insulation layer parameters, and pipeline joint types. Simultaneously, environmental data of the surrounding area, such as the location of surrounding buildings and the distribution of underground pipelines, can be obtained through on-site surveying, ensuring that the data covers core dimensions such as geology, pipelines, and environment, thereby forming a structured multidimensional basic database.

[0023] Furthermore, based on the acquired pipeline data, a geometric model of the pipeline body is constructed in 3D modeling software to generate the pipeline diameter, wall thickness, insulation layer morphology, and joint structure. Simultaneously, the pipeline laying path is located according to the construction drawings. Secondly, a geological geometric model of the construction area is constructed, establishing the spatial distribution of different geological layers based on soil stratification information and marking the specific locations of geological risk zones. Finally, parameters such as soil thermal conductivity and bearing capacity are mapped to the corresponding areas of the geological model, and parameters such as the thermal conductivity and tensile strength of the pipeline material are assigned to the pipeline model. This digital twin reflects the interaction between the pipeline and the geology in the actual scenario.

[0024] Furthermore, based on the geometric and physical models, 3D models of construction equipment and geometric shapes of temporary construction facilities are added to the model to represent construction scene elements. Construction stages, such as excavation, welding, and backfilling, are preset in the model according to the pipeline laying sequence in the construction plan. This ensures that the model can simulate the heat transfer process of the pipeline, ultimately forming a digital twin model.

[0025] In one implementation of this application, based on multi-dimensional basic data corresponding to the thermal pipeline laying area, graph model nodes corresponding to the digital twin model of the thermal pipeline laying are obtained. Based on different construction stages in the thermal pipeline laying process, stage-specific associated sub-models are sequentially extracted from the graph model nodes, wherein the output node of the previous associated sub-model serves as the input node of the next associated sub-model. Based on the real-time construction data corresponding to different construction stages, the edge weights of the stage-specific associated sub-models are adjusted to obtain the construction transmission relationship, wherein the real-time construction data is related to at least one of the following: pipeline laying equipment, pipeline monitoring data, operational data, environmental dynamic data, and process requirements.

[0026] Specifically, core elements are extracted from the digital twin model of the heating pipeline laying process. These core elements can include geological nodes, pipeline component nodes, construction equipment nodes, and environmental nodes. Corresponding attribute parameters are set for each node; for example, soil bearing capacity is set for geological nodes, and material strength is set for pipeline nodes, to construct a graph model node network. Based on the construction process of the heating pipeline laying, namely trench excavation, pipeline welding, backfilling and compaction, and trial operation, core related nodes for each stage are selected from the graph model nodes. For example, the related sub-model for the excavation stage includes nodes such as excavator, trench slope, and soft soil layer; the related sub-model for the welding stage includes nodes such as welding machine, pipeline joint, and heat-affected zone. Through the actual interaction between nodes, such as the disturbance of the slope caused by excavation and the impact of welding on the pipeline joint, the node connection relationships within the sub-model are determined, forming a phased related sub-model. The output nodes of the related sub-model of the previous construction stage are used as the input nodes of the related sub-model of the next stage.

[0027] Furthermore, for each construction stage, corresponding real-time construction data is acquired, including pipeline laying equipment data, pipeline monitoring data, operational data, environmental dynamic data, and process requirements. For example, for the characteristics of thermal pipelines, pipeline characteristic parameters are extracted using corresponding data acquisition sensors. These extracted parameters include at least one of welding current, cooling rate, and weld strength. Based on process requirements, core process characteristic parameters are extracted, including at least backfill moisture content, layer thickness, and compaction threshold.

[0028] Furthermore, this application embodiment includes a mapping rule table for real-time construction data parameter values ​​and edge weights. In this table, for each associated parameter, a reasonable value range for the parameter is determined based on industry standards and historical construction data. The parameter values ​​are then divided into different intervals, corresponding to the basic range of edge weights. For example, when the welding current is 180-220A, the basic weight of the "welding machine-pipe joint" edge is set to 0.6-0.8. The initial weight of each edge is calculated based on the real-time parameter acquisition data of the current construction stage. For example, the welding current corresponding to the "welding machine-pipe joint" edge is 230A, which is in the 220-250A range. Therefore, the initial weight is 0.85 obtained through the mapping rule table. The weights of all edges are assigned through mapping, thereby obtaining the construction transmission association relationship. Furthermore, the collected real-time construction data parameter values ​​are compared with the target values ​​of the process requirements to calculate the parameter deviation rate. For example, if the measured backfill compaction degree is 90% and the standard value is 92%, the deviation rate is -2.17%. If the deviation rate is within ±5%, the weights are slightly adjusted according to the deviation ratio. For example, if the compaction degree deviation is -2.17%, the edge weight of the roller-backfill layer is reduced from 0.8 to 0.78. If the deviation rate exceeds ±5%, a significant weight adjustment is triggered, reducing the corresponding edge weight to the lower limit of the foundation range to reflect the weakening of the node correlation strength due to insufficient construction quality.

[0029] In one implementation of this application, based on the current construction stage, a corresponding associated sub-model is extracted from the digital twin model. The currently acquired real-time construction data is input into this associated sub-model, and the model is dynamically updated based on the construction data transmission relationship corresponding to the associated sub-model. For example, welding current data is input into the welding machine node, and compaction data is input into the backfill layer node. When the measured backfill compaction degree is 90% (lower than the standard 92%), the bearing capacity parameter of the backfill layer node in the model is simultaneously lowered, thereby simulating the change in the vertical settlement of the pipeline. Based on the simulation results, the impact status is output from different dimensions such as heat conduction influence and mechanical disturbance influence. Specifically, the heat conduction influence requires labeling the temperature field distribution and heat-affected zone range of the pipeline and surrounding soil; the mechanical disturbance influence requires presenting the soil stress concentration area, pipeline deformation location, and deformation amount; and the construction progress influence requires combining equipment status data to determine whether there is a risk of construction delay. The impact status is intuitively displayed through the visualization interface of the digital twin model using color gradients and dynamic curves.

[0030] Step 102: Based on the impact situation, the spatial region corresponding to the digital twin model of the thermal pipeline laying is divided into multiple levels to obtain multiple non-uniform sub-region units.

[0031] In one implementation of this application, the spatial region corresponding to the digital twin model of the heating pipeline is initially divided based on geological data and heating pipeline data obtained during the pipeline laying process. Multiple influence characteristics are determined based on the impact conditions; these characteristics include at least one of thermal influence characteristics and mechanical disturbance characteristics. For each influence characteristic, a corresponding regional densification threshold group is configured; this threshold group includes different densification thresholds and corresponding densification algorithms for each threshold. The real-time quantized values ​​of each influence characteristic are compared with the corresponding regional densification threshold group. Based on the comparison results, the spatial region corresponding to the initially divided digital twin model of the heating pipeline is further divided into multiple non-uniform sub-region units; each non-uniform sub-region unit corresponds to a different level of risk sensitivity.

[0032] Specifically, regarding geological data, this application employs ground-penetrating radar to detect soil stratification, including the type, moisture content, thermal conductivity, and compressive strength of each soil layer. For heating pipeline data, information such as pipe material type, diameter, thermal conductivity, weld location, and design pressure capacity is collected. Based on the soil homogeneity and stratification characteristics in the geological data, the digital twin model of the heating pipeline is divided into different geological zones along the horizontal direction. The boundary of each geological zone is consistent with the actual soil stratification boundary. Combining the pipeline laying path and key node distribution in the heating pipeline data, the model is divided into several pipe segment units along the pipeline route. For low-risk areas with homogeneous soil, straight pipelines, and no key nodes, the initial unit size can be larger, for example, 3m × 3m × 2m. For potentially high-risk areas with heterogeneous soil, pipelines containing welds, or bends, the initial unit size can be smaller, for example, 2m × 2m × 1.5m. Simultaneously, the corresponding geological parameters and pipeline parameters are bound to the attribute information of each initial unit.

[0033] Furthermore, in this embodiment, temperature gradient, heat flux density, and thermal diffusivity are obtained by using temperature sensors embedded on the surface of the pipeline and in the surrounding soil. Strain gauges are attached to the pipeline welds and bends to obtain pipeline stress values, soil pressure sensors are used to obtain soil compression loads, and displacement sensors are used to monitor the displacement between the pipeline and the soil. These methods are then used to extract mechanical disturbance characteristics such as stress concentration factor, peak soil pressure, and displacement rate.

[0034] Furthermore, based on historical data, this application embodiment sets three levels of encryption thresholds, including a low-risk threshold, a medium-risk threshold, and a high-risk threshold. The corresponding encryption algorithms are as follows: The low-risk threshold uses an equidistant partitioning algorithm, that is, maintaining the initial grid size and only optimizing the smoothness of the grid boundaries. The medium-risk threshold uses a quartering encryption algorithm, that is, dividing the initial grid into two equal parts along the length, width, and height directions to form eight sub-region units. The high-risk threshold uses an adaptive interpolation encryption algorithm, that is, based on the temperature gradient distribution, increasing the partition density in areas with drastic gradient changes and decreasing the partition density in areas with gentle gradients. For the pipe stress in the mechanical disturbance characteristics, three levels of densification thresholds are set according to the yield strength of the pipe material, including low-risk threshold, medium-risk threshold and high-risk threshold. The corresponding densification algorithms are as follows: the low-risk threshold adopts the boundary optimization algorithm, that is, only the fit between the mesh boundary and the pipe outline is adjusted; the medium-risk threshold adopts the octet densification algorithm, that is, the initial mesh is divided into 3 equal parts along the length, width and height directions to form 27 sub-region units; the high-risk threshold adopts the finite element subdivision densification algorithm, that is, based on the stress distribution cloud map, a high-density tetrahedral region is automatically generated in the stress concentration area.

[0035] Furthermore, the quantified values ​​of each influencing feature are acquired in real time via IoT sensors, preprocessed by the edge computing module, and then transmitted to the digital twin model data processing center. For each initial sub-region unit, the real-time quantified values ​​of its corresponding influencing features are extracted and compared one by one with the regional encryption threshold group for that feature. For example, if the real-time quantified value of the temperature gradient of an initial sub-region unit is 4℃ / m, after comparison with the temperature gradient regional encryption threshold group (low risk ≤2℃ / m, medium risk 2-5℃ / m, high risk >5℃ / m), the temperature gradient of this unit is determined to be in the medium-risk range. At the same time, the real-time quantified value of the pipeline stress of this unit is 160MPa, after comparison with the pipeline stress regional encryption threshold group (low risk ≤120MPa, medium risk 120-180MPa, high risk >180MPa), the pipeline stress is determined to be in the medium-risk range. The comparison results of all influencing features are recorded to determine the risk level and corresponding encryption requirements of each initial sub-region unit under each feature dimension. This application employs a priority stacking method to perform multi-level partitioning of the initially divided digital twin model. First, the risk priority of each influencing feature is determined. For each initial sub-region unit, the comparison results of all influencing features are integrated, and the encryption algorithm and precision corresponding to the highest priority risk level are selected for partitioning. Simultaneously, each partitioned non-uniform sub-region unit is labeled with a risk sensitivity level tag and associated with the corresponding quantified value of the influencing feature and the encryption basis, ultimately forming multiple non-uniform sub-region units that cover the entire construction area and are risk-adapted.

[0036] This application's embodiments can accurately capture subtle risk changes in high-risk areas by combining initial partitioning with impact feature grading thresholds and encryption algorithms, while maintaining coarser regional units in low-risk areas, saving computing power and data resources and balancing monitoring accuracy and resource efficiency. Secondly, it allows the partitioned non-uniform sub-region units to be bound with labels of different risk sensitivity levels, intuitively distinguishing risk scenarios and quickly locating risk sources, shortening risk response time. Furthermore, through continuous comparison of real-time quantized values ​​and threshold groups, the grid density can be dynamically adjusted according to the construction progress and environmental changes, ensuring the model always adapts to actual working conditions, avoiding risk omissions or misjudgments caused by static partitioning, and improving the adaptability of the digital twin model to complex construction scenarios.

[0037] Step 103: Based on historical thermal pipeline laying data and the aforementioned impact conditions, perform risk data prediction for the heat conduction risk and load risk of each non-uniform sub-region unit.

[0038] In one implementation of this application, historical heat conduction data and historical load data corresponding to the historical laying process of thermal pipelines are collected; the historical heat conduction data and historical load data are the corresponding data obtained within the most recent preset time period. The historical heat conduction data, historical load data, and the correlation coefficient between the current thermal pipeline laying and historical risk events are fused to generate composite features. By pre-setting the conditional probabilities and composite features between Bayesian network nodes, the core risk prediction causes corresponding to each non-uniform sub-region unit are predicted. Based on the historical heat conduction data and historical load data, initial heat conduction risk thresholds and initial load risk thresholds are determined, and an initial threshold matrix is ​​constructed. The initial threshold matrix is ​​optimized based on a pre-set risk-sensitive phase factor, and the risk threshold corresponding to each sub-region unit is determined through the optimized matrix. Event labels are set for the historical heat conduction data and historical load data, and heat conduction risk probability prediction models and load value prediction models are constructed based on the event labels. The prediction results output by the heat conduction risk probability prediction model and the load value prediction model are compared with the risk thresholds, and the risk prediction level corresponding to each non-uniform sub-region unit is determined based on the difference between the prediction results and the risk thresholds.

[0039] Specifically, historical heat transfer and load data for similar projects are extracted from the thermal pipeline engineering database. Historical heat transfer data includes the temperature distribution, heat loss rate, and heat-affected zone range of the pipeline and surrounding soil for different geological types and construction stages. Historical load data includes the earthwork pressure, equipment compaction load, and pipeline self-weight experienced by the pipeline during construction. Risk event records for corresponding historical projects are also collected simultaneously. The collected historical heat transfer and load data are standardized to unify data dimensions. Then, the similarity between the current project and historical projects is calculated. For example, the similarity between the geological type of the current project and historical projects, and the similarity between the current construction technology and historical construction technologies, are calculated. Based on the similarity, a correlation coefficient is determined; a higher correlation coefficient indicates greater reference value of the historical data for the current project. The standardized historical data is then weighted and fused with the correlation coefficient. For example, the correlation coefficient between the current construction and historical events is calculated to be 0.85, with a weight of 0.4 for heat conduction data, 0.4 for load data, and 0.2 for the correlation coefficient. A composite feature is generated by weighted summation. For example, the composite feature value of a certain construction section is "0.6 (standardized heat conduction data) × 0.4 + 0.7 (standardized load data) × 0.4 + 0.85 (correlation coefficient) × 0.2 = 0.67". Each composite feature corresponds to a non-uniform sub-region unit.

[0040] Furthermore, a pre-built Bayesian network model is constructed. The network nodes of this model include heat conduction anomalies, excessive loads, pipeline material defects, changes in geological conditions, and core risk factors. The network is trained using historical risk event data to determine conditional probability tables between nodes. For example, the conditional probability of heat conduction anomalies leading to weld heat loss as a core risk factor is 0.7, and the conditional probability of excessive loads leading to soil compression deformation as a core risk factor is 0.65. Subsequently, the generated composite features are assigned to the corresponding input nodes of the Bayesian network according to node type. For example, the heat conduction-related component of the composite features is input to the heat conduction anomaly node, and the load-related component is input to the excessive load node. Combined with the pre-built conditional probability tables, a Bayesian inference algorithm is used to calculate the posterior probability of each core risk factor. For each non-uniform sub-region unit, its corresponding composite feature data is extracted and input into the model. The factor with the highest posterior probability is selected as the core risk prediction factor for that unit.

[0041] Furthermore, historical heat conduction data and historical load data are input into the preset analytical expressions for heat conduction risk threshold and load risk threshold, respectively. Based on the calculation results, initial heat conduction risk threshold correspondence tables and initial load threshold correspondence tables are generated. Multiple first indicator sets are determined based on historical heat conduction data; each of these first indicator sets includes at least one of the following: peak heat flux density index, thermal diffusivity attenuation index, and temperature gradient uniformity index. Multiple second indicator sets are determined based on historical load data; each of these second indicator sets includes at least one of the following: peak soil pressure index, stress concentration factor index, and load duration. A matrix framework is constructed using the first indicator sets as the matrix row dimension and the second indicator sets as the matrix column dimension. Data from the initial heat conduction risk threshold correspondence tables and initial load threshold correspondence tables are then filled into the corresponding element positions in the matrix framework to generate an initial threshold matrix.

[0042] Specifically, this application embodiment includes a preset heat conduction risk threshold analytical formula and a preset load risk threshold analytical formula. Taking heat flux density threshold calculation as an example, the set risk threshold analytical formula is as follows: Heat flux density threshold = First calibration coefficient × Thermal conductivity of pipe material × Temperature difference between inside and outside the pipe / Pipe wall thickness - Second calibration coefficient × Thermal diffusivity attenuation rate × Pipe operating time; Taking soil pressure threshold calculation as an example, the analytical formula for the load risk threshold is set as follows: Soil pressure threshold = third calibration coefficient × pipe material yield strength / (1 + fourth calibration coefficient × pipe stress concentration factor).

[0043] Historical heat conduction data is input one by one into the pre-defined heat conduction risk threshold formula. All heat conduction-related thresholds are organized in the format of indicator type-historical data batch-threshold result to generate an initial heat conduction risk threshold correspondence table. The table must include indicators such as peak heat flux density threshold, thermal diffusivity decay threshold, and temperature gradient uniformity threshold. Each threshold must be labeled with the historical data time period and parameters on which the calculation was based. Similarly, historical load data is input into the pre-defined load risk threshold formula to organize load-related thresholds and generate an initial load threshold correspondence table, including indicators such as peak soil pressure threshold, stress concentration factor threshold, and load duration threshold, ensuring that each threshold corresponds to historical data.

[0044] Furthermore, each indicator from the first indicator set serves as the row dimension of the matrix, namely peak heat flux density, thermal diffusivity decay rate, and temperature gradient uniformity. Each indicator from the second indicator set serves as the column dimension of the matrix, namely peak soil pressure, stress concentration factor, and load duration. The matrix framework is a two-dimensional table, with row headings indicating the name and unit of the first indicator, and column headings indicating the name and unit of the second indicator. Additionally, an explanatory section is added below the matrix framework, indicating the applicable scenarios, historical data sources, and calculation basis of the indicators, ensuring the standardization and traceability of the matrix framework.

[0045] The threshold data for each first indicator is extracted from the initial heat conduction risk threshold correspondence table, and a matrix framework is matched along the row dimension. Similarly, the threshold data for each second indicator is extracted from the initial load threshold correspondence table, and a framework is matched along the column dimension. For each element in the matrix, the matched heat conduction threshold and load threshold are entered based on the scenario correlation between the row and column indicators. For example, for the element "Peak heat flux density (row) - Peak soil pressure (column)," if it corresponds to a clay layer scenario, a heat flux density threshold of 12 W / m² and a soil pressure threshold of 100 kPa are entered; if it corresponds to a sandy soil layer scenario, a heat flux density threshold of 14 W / m² and a soil pressure threshold of 120 kPa are entered, thus obtaining the initial threshold matrix.

[0046] After obtaining the initial threshold matrix, since this initial threshold matrix is ​​constructed based on historical heat conduction and historical load data, but its essence is a solidified presentation of static historical experience, while heat conduction, load, environment, construction, etc. during the laying of thermal pipelines are all dynamic data, the initial matrix is ​​difficult to fully adapt to the actual scenario, and optimization is needed to realize the transformation of static thresholds into dynamic and accurate thresholds.

[0047] Specifically, the threshold deviation of heat conduction risk is mainly affected by changes in material thermal properties and environmental thermal interference. Key parameters include: thermal conductivity of pipe materials, soil moisture content, ambient temperature fluctuation range, and pipe insulation layer thickness. The threshold deviation of load risk is related to geological conditions, pipe structure, and construction intensity. Key parameters include: soil compressibility modulus, pipe burial depth, pipe yield strength, and construction machinery tonnage. In this embodiment, the deviation is the deviation rate between the actual threshold and the initial threshold. The risk deviation under different key parameter values ​​is statistically analyzed using historical data, and the correlation between key parameters and risk deviation is constructed using a multiple linear regression model. The specific process is as follows: using each key parameter, such as thermal conductivity decay rate and moisture content, as independent variables and risk deviation as the dependent variable, a regression equation is obtained by fitting using the least squares method, such as heat conduction risk deviation = 0.3 × thermal conductivity decay rate + 0.5 × moisture content - 0.2 × ambient temperature fluctuation range. The coefficients of each key parameter in the regression equation are standardized to obtain the weight of each parameter. Then, combined with the real-time quantified value of the parameter, the risk-sensitive phase factor is calculated by weighted summation. The phase factor ranges from -1 to 1. A positive value indicates that the threshold needs to be adjusted upwards, while a negative value indicates that it needs to be adjusted downwards. The larger the absolute value, the more significant the offset effect.

[0048] Furthermore, based on the non-uniform partitioning results of the digital twin model of the thermal pipeline, the risk attributes of each sub-region unit were determined. Specifically, the heat conduction risk sub-region units were divided according to heat flux density levels and temperature gradient ranges. The unit boundaries were aligned with the actual thermal impact range using the temperature sensor locations and monitoring data marked in the model. The load risk sub-region units were divided according to soil type and construction intensity. The unit boundaries were determined by combining soil layer distribution and construction machinery operation trajectories. Each sub-region unit was assigned a unique identifier and associated with its core parameters. The element values ​​corresponding to each sub-region unit were matched in the initial threshold matrix to extract the corresponding initial heat conduction threshold and initial load threshold.

[0049] Furthermore, in this embodiment, the preset threshold offset calculation function is: initial threshold × (1 + risk-sensitive phase factor × k), where k is a scene correction coefficient. For heat conduction risk, k ranges from 0.8 to 1.2, and for load risk, k ranges from 0.6 to 1.0. The calculated risk-sensitive phase factor and the matched initial threshold are substituted into the function to obtain the optimized threshold. Based on this function, all thresholds in the initial threshold matrix are calculated, updating the risk threshold data corresponding to each sub-region unit.

[0050] Subsequently, this embodiment of the application also needs to predict the risk level. The specific process is as follows: Historical heat conduction data is labeled with event tags based on whether heat conduction risk events have occurred. Key features from the historical heat conduction data, such as the temporal change rate of heat flux density, temperature gradient, and thermal diffusivity, are selected as model input variables. A logistic regression algorithm is used to construct a heat conduction risk probability prediction model. The model is trained using 70% of the historical labeled data and validated using 30% of the data to ensure the model accuracy meets the requirements. Similarly, historical load data is labeled with event tags based on whether load exceedance events have occurred. Key features from the load data, such as peak soil pressure, pipeline stress change, and load duration, are selected as input variables. A random forest algorithm is used to construct a load value prediction model. This model is also trained and validated using historical labeled data to ensure the model prediction error meets the requirements. The heat conduction risk probability prediction model is applied to each non-uniform sub-region unit. The current heat conduction characteristic data of that unit is input to obtain the heat conduction risk probability prediction result. The load value prediction model is input into the current load characteristic data of that unit to obtain the load prediction value. Subsequently, the risk threshold corresponding to the unit is extracted from the optimized threshold matrix. The prediction result is compared with the threshold, the difference is calculated, and the risk prediction level standard is set based on the difference. For example, a difference in heat conduction risk probability < 0 and a difference in load value < 0 indicate low risk; any difference between 0 and 0.2 (or 0-5 kPa) indicates medium risk; and any difference > 0.2 (or > 5 kPa) indicates high risk. This yields the risk prediction level corresponding to each non-uniform sub-region unit.

[0051] Step 104: Based on real-time construction data and risk data, extract time-series construction features and time-series risk prediction features for each non-uniform sub-region unit. Compare the extracted features with the pre-set safety evaluation index system to achieve construction safety monitoring.

[0052] In one implementation of this application, construction data and risk data corresponding to each non-uniform sub-region unit are divided into multiple datasets according to different construction types. Based on the construction stages corresponding to each construction type, each dataset is divided into multiple stage-specific time-series subsets. Risk features are extracted from these multiple stage-specific time-series subsets using a sliding time window, and the extracted risk features are compared with a pre-set safety evaluation index system. If any risk feature value does not meet the safety threshold corresponding to the safety evaluation index system, a work condition safety upgrade command is triggered. Based on the work condition safety upgrade command, the mapping relationship between the work condition type and construction risk is determined to obtain construction risk information.

[0053] Specifically, the construction and risk data for each sub-region unit are categorized according to their respective construction types, ensuring that each dataset corresponds to a single construction type and that the data is precisely correlated with the sub-region unit. Each dataset is then split into phased time-series subsets based on the construction stage. For example, trench excavation is divided into different stages such as layout and positioning, trench excavation, and slope support; pipeline welding is divided into different stages such as pre-weld preparation, welding execution, and post-weld inspection. Based on real-time construction logs and data collection timestamps, datasets for each construction type are split according to construction stages and organized into time-series subsets. Each time-series subset is labeled with the corresponding sub-region unit number and stage duration, ensuring that the data time sequence is consistent with the construction stage progress.

[0054] Furthermore, matching window parameters are set for each phased time-series subset, and the mean, standard deviation, and rate of change of risk parameters within the window are extracted. The extracted features are compared with the corresponding indicators in the pre-set safety evaluation index system, and the compliance status of each feature value is recorded to generate a feature comparison report. When the feature value of any phased time-series subset of a sub-region unit exceeds the safety threshold, the system automatically triggers a working condition safety upgrade command. The command content must include the sub-region unit number that triggered the anomaly, the construction type and stage, the name and actual value of the anomaly feature, and the corresponding safety threshold. At the same time, the upgrade command level is determined according to the degree of feature value deviation to ensure that the command can accurately locate the anomaly scenario and risk level.

[0055] Furthermore, key elements are extracted from the triggered instructions to obtain the construction type corresponding to the current construction stage. The construction type corresponds to core categories such as trench excavation and pipeline welding. The actual characteristic parameter values ​​of the triggered anomaly and the standard values ​​of the corresponding features in the pre-set safety evaluation index system are also obtained. The working condition type is determined based on the difference between the actual characteristic value and the safety threshold. Based on the magnitude of the difference, the working condition type is divided into four levels: normal, attention, warning, and emergency. The historical working condition information database corresponding to this application embodiment contains core fields such as construction type, construction stage, working condition type, abnormal characteristic value, and difference rate for historical working conditions, and is associated with corresponding handling records and risk results. Using the construction type, construction stage, working condition type, and abnormal characteristic type of the current instruction as matching dimensions, a feature similarity algorithm is used to calculate the matching degree between the current working condition and historical working conditions, and 2-3 historical similar working condition information with the highest matching degree are selected.

[0056] Furthermore, risk records corresponding to similar historical working conditions are retrieved from the construction risk information database. This database stores a mapping relationship between historical working conditions, risk types, risk levels, scope of impact, core causes, and disposal recommendations. Based on this mapping relationship, construction risk information is matched. This construction risk information is structured construction risk information that includes: risk type, risk level, scope of impact, core causes, and targeted disposal recommendations.

[0057] Figure 2 This is a structural schematic diagram of a construction safety monitoring device for laying thermal pipelines, provided as an embodiment of this application. Figure 2 As shown, the construction safety monitoring device 200 for laying thermal pipelines includes: at least one processor 201; and a memory 202 communicatively connected to the at least one processor 201. The memory 202 stores instructions executable by the at least one processor 201, which, when executed by the at least one processor 201, enable the at least one processor 201 to: input real-time construction data during the laying process into a digital twin model of the thermal pipeline; based on the data transmission relationship, obtain the impact of the real-time construction data on the digital twin model of the thermal pipeline; and based on the impact, input the data into the digital twin model of the thermal pipeline. The spatial region corresponding to the twin model is divided into multiple non-uniform sub-region units. Based on historical thermal pipeline laying data and real-time construction data, risk data prediction is performed on the heat conduction risk and load risk of each non-uniform sub-region unit. Among them, the risk data includes at least one of the risk prediction level and the core risk prediction cause. Through a time sliding window, time-series construction feature extraction and time-series risk prediction feature extraction are performed on the real-time construction data and risk data corresponding to each sub-region unit. The extracted features are compared with the pre-set safety evaluation index system to realize the construction safety monitoring of thermal pipeline laying.

[0058] This application provides a non-volatile computer storage medium storing computer-executable instructions. These instructions are configured to: input real-time construction data during the laying of a thermal pipeline into a digital twin model of the thermal pipeline; based on the data transmission relationship, obtain the impact of the real-time construction data on the digital twin model of the thermal pipeline; based on the impact, divide the spatial region corresponding to the digital twin model of the thermal pipeline into multiple levels to obtain multiple non-uniform sub-region units; based on historical thermal pipeline laying data and real-time construction data, predict the heat conduction risk and load risk of each non-uniform sub-region unit; wherein the risk data includes at least one of the risk prediction level and the core risk prediction cause; through a time sliding window, extract time-series construction features and time-series risk prediction features from the real-time construction data and risk data corresponding to each sub-region unit, respectively; and compare the extracted features with a pre-set safety evaluation index system to achieve construction safety monitoring of the thermal pipeline laying.

[0059] The various embodiments in this application are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the embodiments of apparatus, devices, and non-volatile computer storage media are basically similar to the method embodiments, so the descriptions are relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0060] The above descriptions are merely embodiments of this application and are not intended to limit the scope of this application. For those skilled in the art, various modifications and variations can be made to the embodiments of this application. These modifications or substitutions do not cause the essence of the corresponding technical solutions to depart from the spirit and scope of the technical solutions in the embodiments of this application.

Claims

1. A method for monitoring the construction safety of thermal pipeline laying, characterized in that, The method includes: Real-time construction data during the laying of thermal pipelines is input into the digital twin model of thermal pipeline laying. Based on the data transmission relationship, the impact of the real-time construction data on the digital twin model of thermal pipeline laying is obtained. Based on the aforementioned impact, the spatial region corresponding to the digital twin model of the thermal pipeline laying is divided into multiple levels to obtain multiple non-uniform sub-region units. Based on historical thermal pipeline laying data and the real-time construction data, risk data prediction is performed on the heat conduction risk and load risk of each of the non-uniform sub-region units; wherein, the risk data includes at least one of the risk prediction level and the core risk prediction cause. By using a time-sliding window, the real-time construction data and risk data corresponding to each sub-region unit are subjected to time-series construction feature extraction and time-series risk prediction feature extraction, respectively. The extracted features are compared with a preset safety evaluation index system to achieve construction safety monitoring of thermal pipeline laying.

2. The construction safety monitoring method for laying thermal pipelines according to claim 1, characterized in that, Before obtaining the impact of the real-time construction data on the digital twin model of the thermal pipeline laying based on the construction data transmission relationship, the method further includes: Based on the multi-dimensional basic data corresponding to the thermal pipeline laying area, the graph model nodes corresponding to the digital twin model of the thermal pipeline laying are obtained; Based on the different construction stages in the laying of thermal pipelines, stage-related sub-models are extracted sequentially from the nodes of the graph model; wherein, the output node of the previous related sub-model serves as the input node of the next related sub-model. Based on the real-time construction data corresponding to different construction stages, the edge weights of the stage-specific association sub-model are adjusted to obtain the construction transmission association relationship; wherein, the real-time construction data is related to at least one of the following: pipeline laying equipment, pipeline monitoring data, operation data, environmental dynamic data, and process requirements.

3. The construction safety monitoring method for laying thermal pipelines according to claim 1, characterized in that, Based on the aforementioned impact conditions, the spatial region corresponding to the digital twin model of the thermal pipeline laying is divided into multiple levels to obtain multiple non-uniform sub-region units, specifically including: Based on the geological data and thermal pipeline data obtained during the laying of the thermal pipeline, the spatial region corresponding to the digital twin model of the thermal pipeline laying is initially divided. Based on the aforementioned impact conditions, multiple impact characteristics are identified; wherein, the multiple impact characteristics include at least one of thermal impact characteristics and mechanical disturbance characteristics; For each of the aforementioned influencing features, a corresponding regional encryption partitioning threshold group is configured; wherein, the regional encryption partitioning threshold group includes different encryption thresholds, and also includes encryption partitioning algorithms corresponding to each of the aforementioned encryption thresholds; The real-time quantized values ​​of each of the aforementioned impact features are compared with the corresponding regional encryption and division threshold groups. Based on the comparison results, the spatial region corresponding to the initially divided digital twin model of thermal pipeline laying is divided into multiple levels to obtain multiple non-uniform sub-region units. Each of the aforementioned non-uniform sub-region units corresponds to a different degree of risk sensitivity.

4. The construction safety monitoring method for laying thermal pipelines according to claim 1, characterized in that, The method of predicting the heat conduction risk and load risk of each non-uniform sub-region unit based on historical thermal pipeline laying data and real-time construction data specifically includes: Collect historical heat conduction data and historical load data corresponding to the historical laying process of thermal pipelines; the historical heat conduction data and historical load data are the corresponding data obtained within the most recent preset time period; The historical heat conduction data, the historical load data, and the correlation coefficient between the current thermal pipeline laying and historical risk events are fused to generate composite features; By pre-setting the conditional probabilities between Bayesian network nodes and the composite features, the core risk prediction factors corresponding to each of the non-uniform sub-region units are predicted. Based on the historical heat conduction data and the historical load data, the initial heat conduction risk threshold and the initial load risk threshold are determined, and an initial threshold matrix is ​​constructed. The initial threshold matrix is ​​optimized based on a preset risk-sensitive phase factor, and the risk threshold corresponding to each sub-region unit is determined by the optimized matrix. Event tags are set for the historical heat conduction data and the historical load data respectively, so as to construct a heat conduction risk probability prediction model and a load value prediction model based on the event tags respectively; The prediction results output by the heat conduction risk probability prediction model and the load value prediction model are compared with the risk threshold. Based on the difference between the prediction results and the risk threshold, the risk prediction level corresponding to each of the non-uniform sub-region units is determined.

5. The construction safety monitoring method for laying thermal pipelines according to claim 4, characterized in that, Based on the historical heat conduction data and the historical load data, the initial heat conduction risk threshold and the initial load risk threshold are determined, and an initial threshold matrix is ​​constructed, specifically including: Input the historical heat conduction data and the historical load data into the preset heat conduction risk threshold analytical formula and the preset load risk threshold analytical formula respectively, and generate the initial heat conduction risk threshold correspondence table and the initial load threshold correspondence table based on the calculation results; Based on the historical heat conduction data, a plurality of first index sets are determined; wherein, the plurality of first index sets include at least one of the following: peak heat flux density index, thermal diffusivity attenuation index, and temperature gradient uniformity index; Based on the historical load data, multiple sets of second indicators are determined; wherein, the multiple sets of second indicators include at least one of the following: peak soil pressure index, stress concentration factor index, and load duration. A matrix framework is constructed using the first set of indicators as the row dimension and the second set of indicators as the column dimension. The data from the initial heat conduction risk threshold correspondence table and the initial load threshold correspondence table are respectively filled into the corresponding element positions in the matrix frame to generate the initial threshold matrix.

6. The construction safety monitoring method for laying thermal pipelines according to claim 4, characterized in that, The optimization of the initial threshold matrix based on a pre-set risk-sensitive phase factor specifically includes: Based on the risk characteristics of heat conduction risk and load risk, the key parameters affecting the threshold offset are determined, and the risk-sensitive phase factor is obtained based on the correlation between the risk offset and the key parameters. Determine the sub-region units corresponding to the heat conduction risk and the load risk respectively, and match the corresponding risk thresholds in the initial threshold matrix through the sub-region units; The risk-sensitive phase factor and the corresponding risk threshold are input into a preset threshold offset calculation function, and the initial threshold matrix is ​​optimized based on the calculation results.

7. The construction safety monitoring method for laying thermal pipelines according to claim 1, characterized in that, The process of extracting temporal construction features and temporal risk prediction features for each of the aforementioned non-uniform sub-region units, and comparing the extracted features with a pre-set safety evaluation index system to achieve construction safety monitoring, specifically includes: According to different construction types, the construction data and risk data corresponding to each of the non-uniform sub-region units are divided into multiple datasets; Based on the construction stages corresponding to each of the construction types, each of the datasets is divided into multiple time-series subsets of stage data. Risk features are extracted from multiple time-series subsets of the aforementioned phased data using a sliding time window, and the extracted risk features are compared with a pre-set security evaluation index system. If any risk characteristic value fails to meet the safety threshold corresponding to the safety evaluation index system, a working condition safety upgrade command is triggered. Based on the aforementioned working condition safety upgrade instruction, the mapping relationship between working condition type and construction risk is determined to obtain construction risk information.

8. The construction safety monitoring method for laying thermal pipelines according to claim 7, characterized in that, Based on the work condition safety upgrade instruction, the mapping relationship between work condition type and construction risk is determined to obtain construction risk information, specifically including: Obtain the instruction element information corresponding to the work condition safety upgrade instruction; wherein, the instruction element information includes at least one of the following: current construction stage, construction type, actual characteristic value, and safety threshold. The operating condition type is determined based on the difference between the actual feature value and the safety threshold. Based on the operating condition type and the instruction element information, feature matching is performed in the historical operating condition information database to obtain historical similar operating condition information. Based on the mapping relationship between the historical similar working condition information and the construction risk information database, the construction risk information is determined.

9. A construction safety monitoring device for laying thermal pipelines, characterized in that, The device includes a memory for storing computer program instructions and a processor for executing the program instructions, wherein when the computer program instructions are executed by the processor, the device is triggered to perform the method described in any one of claims 1-8.

10. A non-volatile computer storage medium storing computer-executable instructions, characterized in that, The computer-executable instructions are capable of performing the method described in any one of claims 1-8.

Images