Aviation oil pipeline collapsible loess risk identification and inspection method

Abstract

The invention relates to the technical field of pipeline safety monitoring, in particular to an aviation oil pipeline collapsible loess risk identification and inspection method, which comprises the following steps of: dividing and numbering monitoring sub-regions, collecting humidity data in the monitoring sub-regions, calculating a soil humidity change value to identify pre-risk sub-regions, and constructing an effective stress network for the pre-risk sub-regions to identify the collapsible loess risk of an aviation oil pipeline. Analyzing the network structure entropy and the maximum node dynamic unbalance degree, obtaining a predicted collapsibility potential coefficient by using the trained prediction model, comprehensively calculating a collapsibility-induced risk value, collecting pipeline response data to calculate a pipeline response offset, and performing coupling correlation analysis on the collapsibility-induced risk value and the pipeline response offset to obtain the collapsibility-induced risk value. Therefore, the risk level is judged, and the corresponding unmanned aerial vehicle inspection frequency is dynamically matched, so that risk identification and quantitative evaluation from soil structure stability to pipeline mechanical response are realized, collapse disasters can be warned in advance, and intelligent decision support is provided for accurate scheduling of inspection resources.

Description

Technical Field

[0001] This invention relates to the field of pipeline safety monitoring technology, specifically to a method for risk identification and inspection of collapsible loess in aviation fuel pipelines. Background Technology

[0002] As a key infrastructure for national energy transportation, the long-term safe operation of aviation fuel pipelines is of paramount importance. When traversing the collapsible loess regions that are widely distributed in Northwest and North my country, the pipelines face unique engineering geological risks. Collapsible loess has high strength in its natural state, but its structure will be rapidly destroyed and significant subsidence will occur after it is soaked in water. This uneven settlement will generate huge additional stress on the buried pipeline, leading to pipe displacement, deformation or even rupture, causing serious leaks and safety accidents. Currently, risk management of aviation fuel pipelines in collapsible loess areas still relies mainly on manual inspections, periodic geological surveys, and experience-based judgments. This approach has significant limitations, such as long monitoring cycles, discontinuous spatial coverage, and difficulty in identifying latent or early risks. Although existing automated monitoring methods can collect single parameters such as pipeline strain, settlement, and water level in real time, they usually use fixed thresholds to trigger alarms, lack the ability to analyze the coupling relationship between parameters, and are also unable to characterize the nonlinear evolution characteristics of collapsible loess during the infiltration process. This fails to meet the need for dynamic risk identification and graded early warning. To address the aforementioned shortcomings, a technical solution is provided. Summary of the Invention

[0003] To address the aforementioned shortcomings of existing technologies, this invention provides a method for risk identification and inspection of collapsible loess in aviation fuel pipelines. This method effectively solves the problems of existing technologies, such as the limitations of single monitoring methods, the inability to perform correlation analysis on multi-source monitoring parameters, and the inability to dynamically identify and warn of potential settlement risks induced by collapsibility.

[0004] To achieve the above objectives, the present invention can be implemented through the following technical solutions: This invention provides a method for risk identification and inspection of collapsible loess in aviation fuel pipelines, comprising the following steps: The target monitoring area is determined and divided into monitoring sub-areas according to a preset length interval; Humidity data is collected in each monitoring sub-area to calculate soil moisture change values, thereby determining whether it is a pre-risk sub-area. If it is a pre-risk sub-area, a risk identification signal is triggered. Based on risk identification signals, an effective stress network is constructed, and the network structure index is analyzed. A prediction model is established based on the network structure index to obtain the predicted collapse potential coefficient. Finally, the collapse-induced risk value is calculated by combining the network structure index and the predicted collapse potential coefficient. Construct the pipeline response time-series matrix and calculate the pipeline response offset; The risk value induced by wet collapse in the pre-risk sub-region is coupled and correlated with the pipeline response offset to determine the risk status and match the corresponding UAV inspection frequency.

[0005] Furthermore, the method for calculating the soil moisture change value is as follows: Humidity data were collected for each monitoring sub-area within a unit time period. The humidity data included surface soil moisture and deep soil moisture. The humidity standard deviation value is obtained by weighting the difference between the humidity data at each monitoring time within a unit time period and the corresponding reference humidity data. The humidity data at each monitoring time within a unit time period is weighted and calculated with the humidity data at the corresponding initial monitoring time to obtain the initial humidity deviation value. The soil moisture change value is obtained by combining the standard deviation of humidity and the initial deviation of humidity.

[0006] Furthermore, the method for determining whether a region is a pre-risk sub-region is as follows: The soil moisture change value of each monitoring sub-region is compared with the preset reference moisture change range. If the soil moisture change value of a certain monitoring sub-region is outside the preset reference moisture change range, the monitoring sub-region is identified as a pre-risk sub-region.

[0007] Furthermore, the method for constructing the effective stress network is as follows: Each detection point deployed in the pre-risk sub-region is regarded as a node i in the network model, and the effective stress value of the node at monitoring time t is used as its node weight. Combined with the spatial geometric layout relationship between detection points, the connection edge between nodes is established. The connection strength can be determined according to the inverse distance function, thereby constructing the effective stress network. The effective stress is calculated by acquiring the total vertical stress data and pore water pressure of each node at the monitoring time.

[0008] Furthermore, the network structure metrics include network structure entropy and maximum node dynamic imbalance, wherein: Based on the effective stress distribution in the effective stress network, the probability of effective stress falling into different intervals is calculated to obtain the network structure entropy; The dynamic imbalance of a node is obtained based on the degree of deviation between the effective stress change rate of each node and the average change rate of adjacent nodes.

[0009] Furthermore, the method for solving the predicted collapsibility potential coefficient is as follows: Based on historical monitoring data, a time-series training dataset is constructed by correlating network structure indicators with actual observed settlement amounts. By using a long short-term memory network as a prediction model, the constructed time-series training dataset is input into the model for training. The model input is the network structure index sequence of the past T consecutive time steps, and the model output is the wet collapsibility potential coefficient corresponding to the future preset time step. During the real-time monitoring phase, based on the network structure indicators calculated during the current monitoring cycle, an input sequence consistent with that used in the model training phase is constructed. This input sequence is then fed into the trained prediction model to obtain the predicted collapsibility potential coefficient.

[0010] Furthermore, the formula for calculating the risk value induced by landslides is as follows: ; in, This represents the risk value induced by subsidence. For network structure entropy, The maximum dynamic imbalance of the node. To predict the collapsibility potential coefficient, This is the baseline value of network structure entropy under stable regional conditions. , , and All are weighting coefficients.

[0011] Furthermore, the method for calculating the pipeline response offset is as follows: Pipeline response data is collected within the pre-risk sub-region. The pipeline response data includes pipeline axial strain, circumferential strain, pressure difference between inside and outside the pipe wall, and pipe displacement. These data are then organized in chronological order to construct the corresponding pipeline response time series matrix. The pipeline response offset is calculated by adding the difference between the data in the pipeline response time series matrix and the corresponding reference data.

[0012] Furthermore, the method for determining the risk status and matching the corresponding drone inspection frequency is as follows: When the risk value of the wet collapse induced by the pre-risk sub-region continues to rise within a set short time window, and the pipeline response offset increases synchronously or with lag abnormally within that time window, the pre-risk sub-region is determined to be in a high-risk state. When the risk value of the wet collapse induced by the pre-risk sub-region continues to rise within a short time window, and the pipeline response offset does not increase synchronously or with lag, the pre-risk sub-region is determined to be in a potential risk state. When a pre-risk sub-region is determined to be in a high-risk or potential-risk state, the weighted difference between the risk value induced by wet collapse and the pipeline response offset is extracted as the risk level difference. The risk level is determined based on the risk level range in which the risk level difference is located, and the corresponding UAV inspection frequency is matched based on the risk level.

[0013] The technical solution provided by this invention has the following advantages compared with the known prior art: 1. This invention, by establishing a progressive monitoring and analysis chain from soil moisture changes to soil stress field structure to pipeline mechanical response, changes the traditional lag mode that relies on a single threshold alarm. In particular, by introducing indicators such as effective stress network model and structural entropy, it can capture early signals of subsidence development from the perspective of soil internal structural stability, thereby achieving sensitivity and foresight in risk identification. 2. This invention combines dynamic network analysis of soil structure with collapsibility potential prediction based on long short-term memory network. It not only assesses the current state but also quantifies the future short-term risk trend. Then, it couples and correlates this soil risk with the actual response offset of the pipeline, thereby realizing closed-loop verification from environmental causes to engineering effects. This makes the risk assessment conclusion more scientific and reliable, and effectively reduces false alarms and false negatives. 3. This invention transforms the complex results of multi-source data fusion analysis into intuitive risk levels and precisely matched drone inspection frequencies. This enables inspection resources to shift from the traditional fixed-cycle, full-coverage mode to a dynamic scheduling and key inspection mode based on real-time risk status, greatly improving the pertinence and operational efficiency of inspection work, thereby achieving a strategic upgrade from passive response to proactive prevention and precise control. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of the present invention 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 of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 This is a flowchart illustrating the overall process of the present invention. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0017] like Figure 1 As shown, the method for risk identification and inspection of collapsible loess in aviation fuel pipelines includes the following steps: Step 1: Based on the geological survey report along the pipeline route, the section of the pipeline crossing the collapsible loess layer is identified as the target monitoring area. The target monitoring area is then divided into sub-regions according to a pre-set length interval (e.g., every 1 kilometer), and each sub-region is sequentially numbered. The sub-region numbering is set as follows: ,and , To monitor the total number of sub-regions; Soil collectors and pipeline status collectors are deployed in each monitoring sub-area to collect monitoring data within the monitoring sub-area. The soil collectors may include humidity sensors, soil pressure cells, and pore water pressure gauges, while the pipeline status collectors may include strain gauges, pressure sensors, and displacement gauges. In one specific embodiment, the present invention divides the monitoring method into sub-regions according to preset length intervals, so as to form a uniform coverage layout. By combining dual-source data from soil collectors and pipeline status collectors, the accuracy and reliability of the monitoring results are improved, providing strong data support for subsequent analysis.

[0018] Step 2: Collect humidity data in each monitoring sub-area using a humidity sensor to obtain humidity data for each monitoring sub-area, including surface soil humidity and deep soil humidity; Among them, the surface soil moisture is real-time data obtained by the moisture sensors in the shallow surface layer (0-50 cm) within a preset range (e.g., 5 meters) directly above and on both sides of the pipeline. These data are most sensitive to rainfall, snowmelt or surface runoff. Deep soil moisture is real-time data obtained by moisture sensors buried near and below the pipeline burial depth. This data is used to determine whether moisture has infiltrated to the critical depth that affects the pipeline foundation. For each monitoring sub-region, the surface soil moisture and deep soil moisture within a unit time period are obtained. The surface soil moisture and deep soil moisture at each monitoring moment within the unit time period are then weighted and calculated with the corresponding reference surface soil moisture and reference deep soil moisture to obtain the standard deviation value of humidity. The specific calculation formula is as follows: Where t is the label of each monitoring time, and These represent the surface soil moisture and deep soil moisture at monitoring time t for the q-th monitoring sub-region, respectively. and These represent the reference surface soil moisture and reference deep soil moisture for the qth monitoring sub-region, respectively. and All are preset weighting coefficients; Simultaneously, the surface soil moisture and deep soil moisture at each monitoring time are weighted and calculated by comparing them with the surface soil moisture and deep soil moisture at the corresponding initial monitoring time to obtain the initial humidity deviation value. The specific calculation formula is as follows: ,in, and These represent the surface soil moisture and deep soil moisture at the initial monitoring time corresponding to the q-th monitoring sub-region, respectively. and All are preset weighting coefficients; Subsequently, the standard deviation of humidity and the initial deviation of humidity were weighted according to preset weights, and the weighted results were summed and averaged to obtain the soil moisture change value of each monitoring sub-region. The specific calculation formula is as follows: Where T is the total number of monitoring time markers. and All are preset weighting coefficients; The soil moisture change value of each monitoring sub-region is compared with the preset reference moisture change range. If the soil moisture change value of a monitoring sub-region is within the preset reference moisture change range, the monitoring sub-region is determined to be a normal sub-region. If the soil moisture change value of a monitoring sub-region is outside the preset reference moisture change range, the monitoring sub-region is determined to be a pre-risk sub-region, and a risk identification signal is triggered. Steps three and four are executed according to the triggered risk identification signal.

[0019] Step 3: Based on the risk identification signals, monitor and analyze the soil structure status of the pre-risk sub-regions. The specific analysis process is as follows: Each detection point deployed within the pre-risk sub-region is regarded as a node i in the network model, and the effective stress value of the node at monitoring time t is used as its node weight. Combined with the spatial geometric layout relationship between detection points (including the distance between detection points, the relative position distribution, and the topological characteristics of the soil structure), the connection edges between nodes are established. The connection strength can be determined according to the inverse distance function. Thus, an effective stress network that dynamically changes with time and has node weights and edge weights is constructed. The effective stress is calculated by obtaining the total vertical stress data of each node at the monitoring time using an earth pressure cell. The corresponding pore water pressure is simultaneously obtained through a pore water pressure gauge. Based on the combination relationship between the two, the effective stress of each node is calculated. The formula for its calculation is: ; Total vertical stress data refers to the total vertical stress at a node, which represents the total pressure generated by the soil's self-weight and all loads above it (including pipeline cover, additional ground loads, etc.) at that node. Pore ​​water pressure refers to the pressure exerted on the water in the pores of the soil at the joint. It reflects the mechanical action of groundwater or infiltration water on the soil skeleton. Based on the effective stress distribution in the effective stress network, the probability of effective stress falling into different intervals is statistically analyzed. Where k represents the interval number into which the effective stress is divided, i.e., the index of each discrete interval used to statistically analyze the probability distribution of the effective stress, and the network structure entropy is calculated. The specific calculation formula is as follows: The network structure entropy reflects the uniformity of the overall stress distribution. When the network structure entropy is stable and high, it indicates that the stress state of the soil is evenly distributed. When the network structure entropy drops rapidly in a short period of time, it indicates that the stress concentration trend of local nodes is significant, indicating that the soil structure may enter the early stage of reorganization or local yielding. For each node, the deviation between its effective stress change rate and the average change rate of its adjacent nodes is calculated to obtain the node's dynamic imbalance. The specific calculation formula is as follows: Where node j is the set of neighboring nodes of node i. Let i be the change in effective stress at node i within the time window. To monitor time intervals, This represents the average rate of change between adjacent nodes. The standard deviation of the rate of change of the state of adjacent nodes is used to identify points of abnormal stress change. The maximum value of the node dynamic imbalance is defined as the maximum node dynamic imbalance. Based on historical monitoring data, the network structure indicators at each monitoring time, including network structure entropy and maximum node dynamic imbalance, are synchronized and correlated with the actual settlement observed at the corresponding monitoring time. A time-series training dataset is constructed using a sliding time window, with the network structure indicator sequence as input and the collapsibility potential coefficient obtained by mapping the settlement as the supervision signal. Long Short-Term Memory Network (such as LSTM) is used as the prediction model. The constructed time-series training dataset is input into the model for training. The model input is the network structure index sequence of the past T consecutive time steps, and the model output is the wet collapsibility potential coefficient corresponding to the future preset time step. During the real-time monitoring phase, based on the network structure indicators calculated during the current monitoring cycle, an input sequence consistent with that used in the model training phase is constructed. This input sequence is then fed into the trained prediction model to obtain the predicted collapsibility potential coefficient. ; To comprehensively reflect the current state of the soil structure and its potential future risks, the network structure entropy is used. Maximum node dynamic imbalance and predicting collapsibility potential coefficient By performing comprehensive weighting, the risk value induced by landslides is obtained. The specific calculation formula is as follows: ,in, This is the baseline value of network structure entropy under stable regional conditions. , , and All are weighting coefficients, and satisfy the following conditions: .

[0020] Step 4: Based on the risk identification signals, monitor and analyze the pipeline response status of the pre-risk sub-regions. The specific analysis process is as follows: Pipeline response data is acquired in real time by pipeline status acquisition devices deployed in the pre-risk sub-region. The pipeline response data includes pipeline axial strain, circumferential strain, pressure difference between inside and outside the pipe wall and pipe displacement. The pipeline response data is organized in chronological order to construct the corresponding pipeline response time series matrix. Among them, the axial strain of the pipeline refers to the strain generated along the pipeline axis (i.e. the pipeline extension direction). It is a quantitative indicator of axial tensile or compressive deformation and is usually monitored in real time by strain gauges arranged along the axial direction. It reflects the stress changes of the pipeline in the longitudinal direction, including longitudinal tension, compression and stress disturbance caused by uneven settlement of the foundation. Circumferential strain refers to the strain generated along the circumferential direction of the pipeline (the circumferential direction perpendicular to the axis). It is used to reflect the deformation of the pipe wall circumferentially caused by external loads, internal pressure fluctuations, or changes in soil lateral pressure. Its monitoring value is often used to identify risks such as uneven pressure on the pipe wall, local squeezing, or concentration of external forces. The pressure difference between the inside and outside of the pipe wall refers to the pressure difference between the pressure inside the pipe wall and the soil pressure or groundwater pressure outside the pipe wall. It is measured in real time by pressure sensors installed inside and outside the pipe wall and is used to determine whether there is excessive internal pressure load, sudden increase in external soil pressure, or additional load caused by groundwater seepage in the pipeline. Pipe displacement refers to the offset of the pipeline in space, including vertical displacement, lateral displacement and comprehensive displacement relative to the original design position. It is usually measured by a displacement gauge and is used to identify the overall movement or local offset of the pipeline caused by foundation subsidence, lateral slippage or stratum disturbance. The pipeline response offset is calculated by adding the differences between the data in the pipeline response time series matrix and the corresponding baseline data. The offset is used to characterize the stress change of the pipeline under the condition of collapse. Step 5: Perform a coupled correlation analysis between the risk value induced by wet collapse in the pre-risk sub-region and the pipeline response offset to determine the risk status and guide the inspection strategy. The specific analysis process is as follows: When the risk value of the wet collapse induced by the pre-risk sub-region continues to rise within a set short time window, and the pipeline response offset increases synchronously or with lag abnormally within that time window, the pre-risk sub-region is determined to be in a high-risk state. When the risk value of the wet collapse induced by the pre-risk sub-region continues to rise within a short time window, and the pipeline response offset does not increase synchronously or with lag, the pre-risk sub-region is determined to be in a potential risk state. When a pre-risk sub-region is determined to be in a high-risk or potential-risk state, the weighted difference between the risk value induced by wet collapse and the pipeline response offset is extracted as the risk level difference. The risk level is then determined based on the risk level range in which the risk level difference is located. Subsequently, the determined risk level is matched with the corresponding UAV inspection frequency to achieve dynamic inspection scheduling for high-risk or potential-risk sub-regions.

[0021] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A method for risk identification and inspection of collapsible loess in aviation fuel pipelines, characterized in that, Includes the following steps: The target monitoring area is determined and divided into monitoring sub-areas according to a preset length interval; Humidity data is collected in each monitoring sub-area to calculate soil moisture change values, thereby determining whether it is a pre-risk sub-area. If it is a pre-risk sub-area, a risk identification signal is triggered. Based on risk identification signals, an effective stress network is constructed, and the network structure index is analyzed. A prediction model is established based on the network structure index to obtain the predicted collapse potential coefficient. Finally, the collapse-induced risk value is calculated by combining the network structure index and the predicted collapse potential coefficient. Construct the pipeline response time-series matrix and calculate the pipeline response offset; The risk value induced by wet collapse in the pre-risk sub-region is coupled and correlated with the pipeline response offset to determine the risk status and match the corresponding UAV inspection frequency.

2. The method for risk identification and inspection of collapsible loess in aviation fuel pipelines according to claim 1, characterized in that, The method for calculating soil moisture variation is as follows: Humidity data were collected for each monitoring sub-area within a unit time period. The humidity data included surface soil moisture and deep soil moisture. The humidity standard deviation value is obtained by weighting the difference between the humidity data at each monitoring time within a unit time period and the corresponding reference humidity data. The humidity data at each monitoring time within a unit time period is weighted and calculated with the humidity data at the corresponding initial monitoring time to obtain the initial humidity deviation value. The soil moisture change value is obtained by combining the standard deviation of humidity and the initial deviation of humidity.

3. The method for risk identification and inspection of collapsible loess in aviation fuel pipelines according to claim 1, characterized in that, The method for determining whether a region is a pre-risk sub-region is as follows: The soil moisture change value of each monitoring sub-region is compared with the preset reference moisture change range. If the soil moisture change value of a certain monitoring sub-region is outside the preset reference moisture change range, the monitoring sub-region is identified as a pre-risk sub-region.

4. The method for risk identification and inspection of collapsible loess in aviation fuel pipelines according to claim 1, characterized in that, The method for constructing an effective stress network is as follows: Each detection point deployed in the pre-risk sub-region is regarded as a node i in the network model, and the effective stress value of the node at monitoring time t is used as its node weight. Combined with the spatial geometric layout relationship between detection points, the connection edge between nodes is established. The connection strength can be determined according to the inverse distance function, thereby constructing the effective stress network. The effective stress is calculated by acquiring the total vertical stress data and pore water pressure of each node at the monitoring time.

5. The method for risk identification and inspection of collapsible loess in aviation fuel pipelines according to claim 1, characterized in that, The network structure metrics include network structure entropy and maximum node dynamic imbalance, wherein: Based on the effective stress distribution in the effective stress network, the probability of effective stress falling into different intervals is calculated to obtain the network structure entropy; The dynamic imbalance of a node is obtained based on the degree of deviation between the effective stress change rate of each node and the average change rate of adjacent nodes.

6. The method for risk identification and inspection of collapsible loess in aviation fuel pipelines according to claim 1, characterized in that, The method for calculating and predicting the collapsibility potential coefficient is as follows: Based on historical monitoring data, a time-series training dataset is constructed by correlating network structure indicators with actual observed settlement amounts. By using a long short-term memory network as a prediction model, the constructed time-series training dataset is input into the model for training. The model input is the network structure index sequence of the past T consecutive time steps, and the model output is the wet collapsibility potential coefficient corresponding to the future preset time step. During the real-time monitoring phase, based on the network structure indicators calculated during the current monitoring cycle, an input sequence consistent with that used in the model training phase is constructed. This input sequence is then input into the trained prediction model to obtain the predicted collapsibility potential coefficient.

7. The method for risk identification and inspection of collapsible loess in aviation fuel pipelines according to claim 1, characterized in that, The formula for calculating the risk value induced by landslides is as follows: ; in, This represents the risk value induced by subsidence. For network structure entropy, The maximum dynamic imbalance of the node. To predict the collapsibility potential coefficient, This is the baseline value of network structure entropy under stable regional conditions. , , and All are weighting coefficients.

8. The method for risk identification and inspection of collapsible loess in aviation fuel pipelines according to claim 1, characterized in that, The method for calculating the pipeline response offset is as follows: Pipeline response data is collected within the pre-risk sub-region. The pipeline response data includes pipeline axial strain, circumferential strain, pressure difference between inside and outside the pipe wall, and pipe displacement. These data are then organized in chronological order to construct the corresponding pipeline response time series matrix. The pipeline response offset is calculated by adding the difference between the data in the pipeline response time series matrix and the corresponding reference data.

9. The method for risk identification and inspection of collapsible loess in aviation fuel pipelines according to claim 1, characterized in that, The method for determining the risk status and matching the corresponding UAV inspection frequency is as follows: When the risk value of the wet collapse induced by the pre-risk sub-region continues to rise within a set short time window, and the pipeline response offset increases synchronously or with lag abnormally within that time window, the pre-risk sub-region is determined to be in a high-risk state. When the risk value of the wet collapse induced by the pre-risk sub-region continues to rise within a short time window, and the pipeline response offset does not increase synchronously or with lag, the pre-risk sub-region is determined to be in a potential risk state. When a pre-risk sub-region is determined to be in a high-risk or potential-risk state, the weighted difference between the risk value induced by wet collapse and the pipeline response offset is extracted as the risk level difference. The risk level is determined based on the risk level range in which the risk level difference is located, and the corresponding UAV inspection frequency is matched based on the risk level.

Images