Method for analyzing stability of slurry retaining wall of slotted hole of diaphragm wall

Abstract

The invention relates to the technical field of intelligent civil engineering, in particular to a stability analysis method for a slurry retaining wall of a slotted hole of an anti-seepage wall. The method comprises the following steps: acquiring slotted hole geological data and slurry characteristic data, and performing pressure simulation according to the slotted hole geological data and the slurry characteristic data to obtain slotted hole pressure data; performing slurry pressure analysis according to the slotted hole pressure data to obtain slurry pressure data, and performing slotted hole wall surface stress calculation according to the slurry pressure data to obtain slotted hole wall surface stress data; performing instability area processing according to the groove hole wall surface stress data to obtain groove hole wall surface instability area data; and performing slot stability analysis according to the slot wall surface instability area data and the slot wall surface stress data to obtain slot stability data. According to the method, the high-risk area in the slotted hole is identified, dynamic monitoring of instability is provided, constructors can take preventive reinforcement measures in the construction process conveniently, and the construction safety is ensured.

Description

Technical Field

[0001] The invention relates to the field of intelligent civil engineering technology, and in particular to a method for analyzing stability of mud wall protection for slot holes in an anti-seepage wall. Background Art

[0002] During the construction of the anti-seepage wall slot, mud is usually used as a wall protection material to maintain the stability of the slot and prevent landslides. However, due to the complexity of factors such as geological conditions, slot depth, and mud characteristics, the pressure and stress distribution on the slot wall at different depths and areas is often uneven, which can easily lead to stress concentration, and then cause instability or even collapse. During the construction process, if an unstable area appears on the slot wall, it will seriously affect the construction safety and project quality, leading to construction accidents, delays in construction period, and increased project costs. How to promptly detect the stability problem of the mud wall protection of the anti-seepage wall slot has become a problem. Summary of the invention

[0003] In order to solve the above technical problems, the present invention proposes a stability analysis method for mud wall protection of anti-seepage wall slots to solve at least one of the above technical problems.

[0004] The present application provides a method for analyzing the stability of a slurry wall of an anti-seepage wall slot, the method comprising:

[0005] S1. Obtaining slot geological data and mud property data, and performing pressure simulation according to the slot geological data and mud property data to obtain slot pressure data;

[0006] S2. Perform mud pressure analysis based on the slot hole pressure data to obtain mud pressure data, and perform slot hole wall stress calculation based on the mud pressure data to obtain slot hole wall stress data;

[0007] S3, performing instability area processing according to the slot hole wall stress data to obtain the slot hole wall instability area data;

[0008] S4. Performing slot stability analysis based on the slot wall instability area data and slot wall stress data to obtain slot stability data.

[0009] The present invention can dynamically process slot geological and mud data, and through real-time pressure simulation and stress analysis, it is suitable for the constantly changing geological environment during the construction process. : By obtaining accurate slot geological data and mud property data and performing pressure simulation on them, the obtained slot pressure data is more in line with the actual working conditions. Through mud pressure analysis based on pressure data and slot wall stress calculation, it is ensured that the pressure and stress distribution inside and around the slot accurately reflects the actual situation. The unstable area processing step can identify possible unstable areas through the stress distribution and concentration of the slot wall, thereby achieving early warning. Stability analysis based on unstable area data and slot wall stress data can provide the overall safety factor and stability evaluation of the slot, which is convenient for construction decision-making.

[0010] Optionally, S1 includes:

[0011] S11, obtaining slot geological data and mud property data through a geological engineering database or a geological engineering input interface;

[0012] S12, performing mud property modeling according to the mud property data to obtain a mud property clustering model;

[0013] S13, performing geological dynamic feature conversion according to the slot geological data to obtain geological dynamic feature sequence data;

[0014] S14, performing virtual reality formation modeling according to the mud property clustering model and the quality dynamic characteristic sequence data to obtain a three-dimensional formation characteristic model;

[0015] S15, calculating the mud characteristic distribution depth according to the three-dimensional formation characteristic model to obtain mud characteristic distribution depth data;

[0016] S16. Perform slot hole dynamic pressure simulation on the mud characteristic distribution depth data to obtain slot hole pressure data.

[0017] In the present invention, multi-source slot geological data and mud property data are obtained through a geological engineering database or a data input interface, so that the initial data source is rich and accurate. The mud property modeling step generates a mud property clustering model, which provides a classification basis for the similarities and differences of different mud property data in depth and distribution, and adopts appropriate simulation strategies for different property distributions to improve the accuracy of pressure and stress analysis. The geological dynamic feature conversion step generates dynamic feature sequence data, so that the time-varying nature of the slot geological characteristics can be quantified, which can reflect the real geological conditions changes during the simulation process, and help improve the adaptability of the model to geological changes, thereby supporting more accurate construction plan formulation. The three-dimensional stratigraphic characteristic model generated by virtual reality stratigraphic modeling provides intuitive three-dimensional visualization for the characteristic distribution of mud and slots, so that engineering personnel can more clearly understand the geological structure and mud distribution around the slot. Through the mud property distribution depth calculation, the distribution of the physical properties of the mud at different depths of the slot is analyzed to ensure that the pressure data in the subsequent simulation steps is more accurate. The slot pressure data generated by the slot dynamic pressure simulation more realistically reflects the effect of mud at different depths on the slot wall through dynamic simulation of real-time pressure changes.

[0018] Optionally, the mud property modeling includes:

[0019] Extract mud characteristic features according to the mud characteristic data to obtain mud characteristic feature data;

[0020] Performing feature coding according to the mud characteristic feature data to obtain mud characteristic coding data;

[0021] Perform preliminary clustering on the mud characteristic coding data to obtain mud characteristic clustering data;

[0022] According to the clustering data of mud characteristics, clustering scale is adjusted and clustering boundary is multi-scaled to obtain multi-scale clustering data of mud characteristics;

[0023] Performing label association on mud characteristic data according to multi-scale clustering data of mud characteristics to obtain mud characteristic clustering label association data;

[0024] The model is constructed based on the associated data of mud property clustering labels to obtain the mud property clustering model.

[0025] In the present invention, key features are extracted from mud property data to provide refined and important data input for subsequent coding and clustering, effectively reduce redundant information, and increase data processing efficiency. Feature coding makes the information of data in high-dimensional feature space more regular and standardized by quantifying mud properties. Through preliminary clustering, the coded data is grouped according to the similarity of characteristics to obtain mud property clustering data. The preliminary clustering effectively reduces the data dimension, provides simplified basic data for subsequent clustering scale adjustment, and helps to improve analysis efficiency. Through adaptive adjustment of clustering scale and multi-scale processing of boundaries, cluster analysis can adapt to different density distributions of mud data. Multi-scale clustering can refine the clustering boundaries of dense areas, reduce boundary fuzziness and errors, and thus provide more scientific results for clustering of different mud properties. The label association step forms mud property clustering label association data by assigning labels to cluster data. Label association enhances the interpretability of the model, so that construction personnel can understand the uses and effects of different mud properties according to labels, and provide clear classification guidance for the construction process. The model building steps based on cluster label associated data form a clustering model of mud characteristics, providing a complete data model including multi-dimensional characteristics of mud, cluster labels and associated relationships.

[0026] Optionally, the feature encoding includes:

[0027] Performing local linear embedding processing on the mud characteristic feature data to obtain mud characteristic feature mapping data;

[0028] Perform characteristic abstraction and stratification according to the mud characteristic feature mapping data to obtain mud abstract characteristic tensor data;

[0029] Perform pre-coding processing on the mud abstract characteristic tensor data to obtain mud compression characteristic data;

[0030] The mud compression characteristic data is recursively encoded to obtain mud characteristic encoding data.

[0031] In the present invention, mud characteristic data is mapped from a high-dimensional space to a low-dimensional feature space by local linear embedding, and the local geometric structure of the characteristic data is retained, which not only reduces the computational complexity, but also ensures the intrinsic characteristic relationship of the data, and provides more refined input data for subsequent characteristic abstraction layering. The characteristic abstraction layering extracts the core information of the mud characteristics through layer-by-layer abstraction, so that the data has higher representativeness. Each layer of abstraction process will gradually retain the elements closely related to the mud characteristics, so as to summarize the characteristics at multiple levels, so that the characteristic data can more completely represent the multi-dimensional characteristics of the mud. The pre-coding process compresses the mud abstract characteristic tensor, removes redundant and noise information, and generates mud compressed characteristic data, which significantly improves the purity and compactness of the data, ensures that the recursive coding process only needs to process the core features, improves the coding efficiency and reduces the computational overhead. Recursive coding converts the mud compressed characteristic data into a structured mud characteristic code to form an efficient data representation that can be used for further analysis. The recursive coding process can identify and capture the deep relationship of the characteristic data, and the generated coded data has uniqueness and high discrimination, which is suitable for subsequent mud clustering and analysis.

[0032] Optionally, the cluster scale adjustment and cluster boundary multi-scaling include:

[0033] Initialize the clustering scale according to the clustering data of mud characteristics to obtain the initial scale data of mud characteristics;

[0034] Density center identification is performed based on initial scale data of mud characteristics to obtain cluster center data of mud characteristics;

[0035] The scale of the cluster center data of mud characteristics is gradually expanded to obtain the expanded scale data of mud characteristics;

[0036] According to the mud property extended scale data, the cluster center change rate is calculated and the cluster change characteristics are extracted to obtain the mud property cluster center change rate data and the mud property cluster change characteristic data;

[0037] According to the mud property cluster center change rate data and the mud property cluster change characteristic data, the boundary of the mud property extended scale data is adjusted to obtain the cluster property multi-scale adjustment data;

[0038] The multi-scale adjustment data of clustering characteristics are used to refine dense areas and obtain multi-scale clustering data of mud characteristics.

[0039] In the present invention, by dynamically adjusting the clustering scale and boundary, the method can more accurately identify the dense areas and differences of mud characteristic data, effectively reduce errors, avoid the over-clustering or under-clustering problems caused by traditional clustering methods, and make the clustering model have higher adaptability and accuracy in diversified mud characteristic data. Multi-scale clustering technology supports differentiated processing of different density areas, so as to adapt to the complex distribution of mud data, and can form refined clusters in dense areas, while maintaining reasonable scales and boundaries in sparse areas, enhancing the versatility and adaptability of clustering. Through layer-by-layer expansion and boundary adjustment, it supports multi-level analysis from macro to micro, which helps to observe the changes and laws of mud characteristic distribution at different scales. Multi-level characteristics provide a more comprehensive understanding of mud characteristics for engineering construction, which is convenient for optimization decision-making. The cluster center change rate and cluster change characteristic data enhance the physical interpretability of the clustering results, so that engineering personnel can understand the law of mud characteristic clustering according to the adjustment of characteristic change rate and cluster boundary, and support more reasonable mud distribution control and management.

[0040] Optionally, S2 includes:

[0041] S21, performing pressure multi-dimensional deconstruction map processing according to the slot pressure data to obtain mud multi-dimensional pressure deconstruction map data;

[0042] S22, performing slot hole depth pressure correlation analysis based on the mud multi-dimensional pressure deconstruction map data to obtain mud pressure data;

[0043] S23, identifying pressure sensitive points according to the mud pressure data to obtain mud pressure sensitive point data;

[0044] S24, performing stress conduction mapping on mud pressure sensitive point data to obtain mud stress conduction mapping data;

[0045] S25, performing slot stress response simulation according to the mud stress conduction mapping data to obtain slot stress response data;

[0046] S26. Screen the stress concentration area according to the slot hole stress response data to obtain the slot hole wall stress data.

[0047] In the present invention, by multi-dimensional deconstruction of slot pressure data, mud multi-dimensional pressure deconstruction map data is generated, which shows the distribution of pressure in different dimensions and different positions, decomposes complex pressure data, highlights key features, and helps to identify the distribution pattern of pressure. By analyzing the correlation of pressure data at different depths, the law of pressure change with depth is identified. This analysis helps to understand the differences in mud pressure at different depth layers, so that the construction party can monitor the pressure abnormality points at a specific depth to prevent the risk of instability caused by pressure abnormalities. Identifying sensitive points in mud pressure data can quickly locate areas with concentrated pressure, effectively indicate potential weak links and unstable areas in the slot, provide an accurate data source for the calculation of stress conduction and concentrated areas, and improve the safety of the model. Through stress conduction mapping of pressure sensitive points, the pressure conduction path and range of mud in the slot are simulated. Stress conduction mapping can accurately show the situation of pressure transmission from sensitive points to surrounding areas, provide real data basis for slot stress response simulation, and help evaluate the effect of mud wall protection. Based on stress conduction mapping, stress response simulation of slot wall is performed. This step simulates the effect of mud pressure on different areas of the slot wall, helps identify high stress areas and locations of stress concentration, and provides key guidance for construction parties on mud configuration and pressure control. The stress concentration area screening step analyzes the stress response data, screens out the stress concentration areas on the slot wall, generates slot wall stress data, and provides accurate pressure distribution areas, making it easier for construction parties to take additional stabilization measures in these high stress areas to ensure the stability of the structure.

[0048] Optionally, the stress transmission mapping includes:

[0049] According to the mud pressure sensitive point data, the mechanical characteristics of the sensitive point are analyzed to obtain the mechanical characteristics matrix data of the sensitive point;

[0050] Initialize the stress field according to the mechanical property matrix data of the sensitive points to obtain the initial stress field distribution map data;

[0051] Conducting adjacent pressure conduction simulation on the initial stress field distribution map data to obtain pressure conduction sequence data;

[0052] Optimize the conduction path according to the pressure conduction sequence data to obtain optimized conduction sequence data;

[0053] Perform geological stress attenuation simulation based on optimized conduction sequence data to obtain stress attenuation distribution matrix data;

[0054] According to the stress attenuation distribution matrix data, the initial stress field distribution map data is adjusted for stress conduction boundary to obtain stress conduction boundary data;

[0055] The mud stress conduction mapping data is obtained by mapping and reconstructing the optimized conduction sequence data, stress attenuation distribution matrix data and stress conduction boundary data.

[0056] In the present invention, through the analysis of mechanical properties of sensitive points, initialization of stress fields and optimization of conduction paths, the real pressure conduction and stress distribution are simulated at different depths and environments, so that the generated stress mapping is more accurate and detailed, fully reflecting the stress conduction of actual working conditions. Through fine processing, including stress decay simulation and boundary adjustment, the present invention can dynamically adjust the stress changes in different regions. The refined stress control provides a deep understanding of the stress distribution on the wall of the slot hole, which helps to discover potential instability risks and prevent them. By analyzing the mechanical properties of sensitive points, the generated stress conduction mapping data has physical meaning, which is convenient for interpreting the transmission and changes of pressure in different regions. At the same time, the present invention can adapt to various geological conditions and mud property data, making the model more universal. Conductive path optimization and stress decay simulation enable the method to quickly identify areas of high stress concentration, thereby providing key data for instability analysis and construction decisions. Timely identification of high stress areas can help the construction party take preventive measures to ensure construction safety. The optimization and reconstruction of conduction data provide dynamic response capabilities for the mud wall protection during construction, so that the stress conduction model during construction can adapt to changes in working conditions in real time, improve the practicality of the model, and help to manage and adjust stress during construction.

[0057] Optionally, S3 includes:

[0058] S31, performing stress distribution anomaly detection according to the slot wall stress data to obtain stress anomaly point data;

[0059] S32, performing instability area analysis based on stress abnormal point data to obtain instability area data of the slot wall;

[0060] S33, performing instability trend analysis based on the instability area data of the slot hole wall to obtain instability trend data of the slot hole wall;

[0061] S34, performing instability region depth identification on the slot wall instability trend data to obtain instability region identification data;

[0062] S35. Dynamically map the unstable region according to the unstable region identification data to obtain the unstable region data of the slot wall.

[0063] In the present invention, through stress anomaly detection and unstable area analysis, high-risk areas can be identified before the instability trend appears, early warning and prompting construction personnel to take preventive measures. Early identification helps to reduce the risk of instability, avoid accidents, and improve construction safety. Through instability trend analysis and dynamic mapping, the instability of the slot wall can be monitored and predicted in real time, so that the change of instability trend can be tracked in time. Dynamic analysis supports engineering personnel to quickly adjust construction strategies when the instability trend changes to ensure project safety. The present invention can accurately identify the position and depth of the unstable area on the slot wall. Accurate positioning helps the construction party to take targeted reinforcement measures, thereby reducing the scope of instability and optimizing the construction plan. The present invention can adapt to complex construction environments and provide support for multi-level instability analysis. No matter how the pressure and stress distribution of the slot wall changes, the evolution and expansion of the unstable area can be effectively identified. The unstable area mapping data generated by the dynamic mapping step provides key data for dynamic instability control. By real-time monitoring of the changes in the unstable area, engineering personnel can continuously monitor and adjust, reduce the probability of instability, and improve the overall safety of the mud wall.

[0064] Optionally, S4 includes:

[0065] S41, extracting stability characteristics according to the instability area data of the slot hole wall and the stress data of the slot hole wall to obtain stability characteristic data;

[0066] S42, calculating the stability index according to the stability characteristic data to obtain the slot area stability index data;

[0067] S43, performing a multi-factor cross analysis based on the slot hole wall instability area data and the slot hole wall stress data to obtain a multi-factor cross analysis matrix data;

[0068] S44, performing stress instability risk assessment according to the slot area stability index data and the multi-factor cross analysis matrix data to obtain instability risk assessment matrix data;

[0069] S45. Calculate the slot safety factor according to the instability risk assessment matrix data to obtain slot stability data.

[0070] The present invention generates stability characteristic data by extracting the stability characteristics of the instability area and stress data of the slot wall, and provides a core characteristic index for stability analysis. According to the stability characteristic data, the stability index of the slot area is calculated, so that the stability assessment result can be presented quantitatively. The stability index provides an intuitive numerical reference for engineering personnel, enabling them to quickly judge the safety status of different areas and facilitate the implementation of targeted protection measures. By performing multi-factor cross-analysis on the instability area data and stress data of the slot wall, a cross-analysis matrix is ​​generated, which integrates multiple influencing factors, helps to identify the relationship between different factors and their comprehensive impact on the stability of the slot, and helps to more comprehensively evaluate the safety of the slot wall. Based on the stability index and the cross-analysis matrix, a stress instability risk assessment is performed to effectively quantify the instability risk of the slot wall. The instability risk assessment matrix can provide the construction party with detailed high-risk areas and potential instability factors, and guide it to take additional protection and monitoring measures in dangerous areas. The slot safety factor is calculated using the instability risk assessment matrix, and the generated slot stability data can fully reflect the stability of the slot wall under the current construction conditions.

[0071] The purpose of the present invention is to combine slot geology and mud property data to perform pressure simulation, so that the initial pressure data has a real engineering background, ensuring the physicality and rationality of the pressure calculation. Through multi-dimensional pressure analysis and slot wall stress calculation, the pressure and stress distribution of the slot are gradually refined, which can not only identify the high stress area of ​​the slot wall, but also effectively capture the stress concentration phenomenon in the mud wall protection system, which is helpful for early identification of potential instability risks. Using slot wall stress data for anomaly detection, it is possible to quickly capture the pressure concentration area and conduct instability risk assessment on these areas. By performing dynamic trend analysis on the unstable area, the evolution characteristics of the unstable area at different depths and different time nodes can be captured, providing the construction party with the expansion direction and rate information of instability. On the basis of the identification of the unstable area, further depth identification is performed to accurately locate the range and depth of the unstable area, providing a clear target area for the implementation of protective measures. By extracting the stability characteristics in the unstable area and stress data, the stability index is calculated to generate stability indexes for different slot areas. Through the multi-factor cross-analysis matrix, the slot wall instability area data, stress data, stability index and other multi-dimensional factors are integrated into the same framework, and the mutual influence between the factors is systematically analyzed. Based on the instability risk assessment matrix, the slot safety factor is calculated to obtain the overall stability data of the slot. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting implementations made with reference to the following drawings:

[0073] Figure 1 A flowchart showing the steps of a method for analyzing the stability of a mud wall protection wall of an anti-seepage wall slot hole according to an embodiment is shown;

[0074] Figure 2 A flowchart showing the steps of a slot pressure simulation method according to an embodiment is shown;

[0075] Figure 3 A flowchart showing a method for calculating slot wall stress according to an embodiment of the present invention is provided;

[0076] Figure 4 A flowchart showing a method for processing an unstable region on a slot wall according to an embodiment of the present invention is provided;

[0077] Figure 5 A flowchart showing a method for analyzing the stability of a slot according to an embodiment of the present invention is provided;

[0078] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0079] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0080] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0081] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0082] The construction party first obtained the stratum depth, density and physical properties of the slot through geological survey and laboratory data. Based on the above data, pressure simulation was performed through finite element analysis. The simulation results obtained the pressure distribution of the slot at different depths. For example: depth 0-10 meters: pressure is 0.6MPa, depth 10-20 meters: pressure is 0.9MPa, depth 20-30 meters: pressure is 1.2MPa, depth 30-50 meters: pressure is 1.5MPa.

[0083] The slot pressure data is multi-dimensionally deconstructed to form a multi-dimensional pressure deconstruction map of the mud. For example, the area with a depth of 0-10 meters is divided into several grid nodes, and the pressure distribution of each node generates a multi-dimensional deconstruction map. The pressure data at different depths are correlated to identify the pressure change relationship at each depth. For example, the pressure in the area with a depth of 10-20 meters increases greatly, forming an area with high pressure correlation. By analyzing the locations where the pressure changes significantly in the deconstruction map, the mud pressure sensitive points are identified. For example, the pressure at 20 meters and 30 meters increases significantly and is marked as a pressure sensitive point. According to the pressure sensitive point data, stress conduction mapping is performed. By analyzing the conduction path of the mud at these sensitive points, a pressure conduction sequence is formed. For example, the pressure at the sensitive point at 20 meters is transmitted to a depth of 30 meters, and an attenuation model is generated, indicating that the attenuation coefficient of the pressure from 20 meters to 30 meters is 0.85. The stress conduction mapping is applied to the entire slot wall to simulate the wall stress response distribution. For example, the stress generated at 20 meters is distributed as 1.1MPa on the wall, while it is distributed as 0.9MPa at 30 meters. Based on the stress response data, the high stress areas on the wall are screened out and the stress data on the slot wall are obtained. For example, the depth of 20-30 meters is the stress concentration area, and the stress value is close to 1.2MPa.

[0084] The slot wall stress data was tested to identify abnormally high stress points. For example, the stress value at a depth of 30 meters reached 1.2MPa, exceeding the safety threshold and was determined to be an unstable abnormal point. The abnormal point was taken as the center point, and its influence range was analyzed to determine the unstable area. The depth of 28-35 meters was determined to be the unstable area. The time change trend of the unstable area was analyzed to obtain the instability trend. The analysis shows that the unstable area has a tendency to expand upward, with an expansion rate of 0.5 meters per day. Based on the instability trend data, the depth range of the unstable area was further determined, and the identification data was obtained: the 28-35 meter area is the main instability depth, and the expansion trend reaches a depth of 40 meters. The dynamic changes of the unstable area are generated into a mapping diagram to show the expansion of the unstable area at different depths and time points.

[0085] The key features that affect the stability of the slot wall are extracted from the unstable area and stress data, including the depth range, stress concentration, etc. The stability features obtained include stress concentration coefficient, expansion rate, etc. Based on the stability feature data, the stability index of the slot area is calculated. The stability index at a depth of 30 meters is 0.65 (lower than the safety threshold of 0.75), indicating that there is an instability risk. A cross-analysis of the unstable area data, stress data and stability index shows that stress concentration and formation density are key influencing factors, and the generated cross-analysis matrix reflects the comprehensive relationship between the various factors. The instability risk of the slot is evaluated through the instability risk assessment matrix. The assessment shows that the instability risk at a depth of 30 meters is high, with a risk factor of 0.8. Based on the risk assessment data, the overall safety factor of the slot is calculated. The calculated safety factor is 1.2 (lower than the standard value of 1.5), indicating that reinforcement measures need to be taken.

[0086] See also Figures 1 to 5 The present application provides a method for analyzing the stability of a mud wall of an anti-seepage wall slot, the method comprising:

[0087] S1. Obtaining slot geological data and mud property data, and performing pressure simulation according to the slot geological data and mud property data to obtain slot pressure data;

[0088] In one embodiment, it is first necessary to obtain slot geological data through on-site geological surveys, laboratory tests, etc., including information such as slot depth, soil layer structure, and groundwater level. In addition, it is necessary to collect experimental analysis data of the mud used for wall protection to obtain the physical and chemical properties of the mud, such as density, viscosity, rheology, etc. The fluidity and permeability of the mud affect its pressure distribution in the slot. Next, pressure simulation is performed using the slot geological data and mud property data. The pressure simulation can analyze the pressure exerted by the mud on the slot wall layer by layer in a layered and segmented manner. In the calculation process, the mud density is associated with the slot depth through the control variable method to simulate the slot pressure data at different depths. During the pressure simulation process, the support effect of the soil layer around the slot and the simulation of the hydrostatic pressure effect of the mud are also added.

[0089] The geological data of the slot are as follows: slot depth: 20 meters, soil layer structure: 0-10 meters is clay layer, 10-20 meters is sand layer, groundwater level: 8 meters, mud characteristic data: mud density: 1.2g / cm 3 , Mud viscosity: 20mPa×s. According to the slot geological data and mud property data, simulate the pressure inside the slot. It can be assumed that the pressure distribution of the slot is a hydrostatic pressure calculation model. When the mud density is known, the mud pressure at different depths of the slot is calculated using the following formula:

[0090] P = ρ × g × h;

[0091] Where: P is the slot mud pressure (Pa), ρ is the mud density (1.2g / cm 3 =1200kg / m 3 ), g is the acceleration due to gravity (9.81 m / s 2 ), h is the mud depth. For example, to calculate the mud pressure of a slot with a depth of 10 meters: P = 1200 × 9.81 × 10 = 117,720 Pa.

[0092] S2. Perform mud pressure analysis based on the slot hole pressure data to obtain mud pressure data, and perform slot hole wall stress calculation based on the mud pressure data to obtain slot hole wall stress data;

[0093] In one embodiment, the mud pressure analysis needs to consider the change of mud pressure with depth and the influence of the slot structure on pressure transmission. The mud pressure data is finely decomposed using the hierarchical analysis method to ensure that the mud pressure data at different depths is accurate. Then, the stress distribution of the slot wall is calculated based on the mud pressure data. The calculation is accumulated layer by layer through the stress formula to simulate the stress conditions of the slot wall at different positions. The stress calculation involves the friction between the mud pressure and the slot wall, as well as the lateral pressure exerted by the mud on the slot wall at different depths. Through this layer-by-layer decomposition and accumulation method, complete slot wall stress data is obtained, providing a basis for unstable area analysis.

[0094] The mud pressure at different depths of the slot is analyzed in detail. Considering the influence of soil structure at different depths on mud pressure, the mud pressure can be adjusted with depth. The total stress on the slot wall is calculated, such as the effective wall stress of the slot (taking soil resistance into account) is: w =P―σ t , where σ w is the wall stress, P is the mud pressure, σ t The resistance stress provided by the soil layer. In the clay layer, the soil layer resistance stress is 20,000Pa, and in the sand layer it is 10,000Pa: Clay layer at a depth of 10 meters: σ w =117,720-20,000=97,720Pa. In the sand layer at a depth of 15 meters: P=1200×9.81×15=176,580Pa, σ w =176,580-10,000=166,580 Pa. The distribution of slot wall stress is obtained by calculation.

[0095] S3, performing instability area processing according to the slot hole wall stress data to obtain the slot hole wall instability area data;

[0096] In one embodiment, the slot wall stress data obtained in step S2 is used to analyze and identify potential unstable areas of the slot wall. First, a stress threshold for the stability of the slot wall needs to be set, and the threshold is set according to the bearing capacity of the soil layer and the supporting capacity of the mud. When the stress value of the slot wall exceeds the threshold, instability will occur. By marking the area exceeding the stability threshold in the slot wall stress data, the unstable area can be identified. In the specific operation process, the slot wall stress data is layered according to the depth, and the stress value is compared with the stability threshold layer by layer to determine the unstable area of ​​each layer, forming the slot wall unstable area data, which includes information such as the depth, range, and stress magnitude exceeding the threshold of the unstable area. The wall stress data is compared with the instability threshold of the slot. If the instability threshold is 150,000Pa, when the slot wall stress exceeds this value, it is considered that there is an instability risk. At a depth of 10 meters, the wall stress is 97,720Pa, which is lower than the instability threshold and there is no instability risk. At a depth of 15 meters, the wall stress is 166,580Pa, which exceeds the instability threshold and there is a risk of instability. Therefore, the sand layer area at a depth of 15 meters is marked as an unstable area.

[0097] S4. Performing slot stability analysis based on the slot wall instability area data and slot wall stress data to obtain slot stability data.

[0098] In one embodiment, the obtained instability area data and wall stress data are used to evaluate the overall stability of the slot. First, based on the range and depth of the instability area, the influence of the instability area on the overall stability of the slot wall is calculated. The stress of the instability area can be compared with the stress of the surrounding stable area by a regional weighted analysis method to calculate the comprehensive stability index of the slot wall at different depths. Then, the stability status of the entire slot is judged by comprehensively analyzing the stability index of each layer. If it is found that the stability index of certain areas is low, corresponding support or reinforcement measures need to be taken during the construction process to ensure the safety of the slot. The formed slot stability data includes the overall stability index of the slot and the possibility of instability at each depth level, providing a scientific basis for risk control during construction.

[0099] The overall stability of the slot is evaluated based on the unstable area data and the overall wall stress distribution. The unstable area is weighted to calculate the overall stability index. The overall stability formula of the slot is:

[0100]

[0101] Among them, S is the stability index, ranging from 0 to 1, the closer to 1, the more stable, σ max is the instability threshold of 150,000Pa, σ wis the actual stress value of each layer, and n is the total number of layers. Substituting the data, at a depth of 10 meters: S 10 =(150,000-97,720) / 150,000≈0.347. At a depth of 15 meters: S 15 =(150,000-166,580) / 150,000≈-0.110 (indicating instability), calculate the overall stability index: S total =(0.347+(-0.110)) / 2≈0.118. This result indicates that the slot is at risk of instability and needs to be reinforced or supported.

[0102] Optionally, S1 includes:

[0103] S11, obtaining slot geological data and mud property data through a geological engineering database or a geological engineering input interface;

[0104] In one embodiment, the geological data and mud property data of the slot are obtained through an existing geological engineering database interface or an engineering input interface. The slot geological data includes information such as soil layer structure, groundwater level, pore pressure, etc. The data is stored in a standardized format in the database and extracted through SQL queries or API calls. In addition, the mud property data mainly covers the density, viscosity, rheology, etc. of the mud, which can be measured experimentally or retrieved from the database. The required data is loaded into the modeling platform through the interface.

[0105] S12, performing mud property modeling according to the mud property data to obtain a mud property clustering model;

[0106] In one embodiment, the modeling process of mud property data is intended to cluster and classify mud properties at different depths to determine the behavior patterns of mud at different depths. First, the mud data is standardized to normalize parameters such as density and viscosity to improve accuracy in the multivariate clustering process. Next, the mud properties are classified in an iterative manner to assign muds with similar properties to the same category. After clustering is completed, the system generates a mud property clustering model that displays the distribution of mud properties at different depths.

[0107] S13, performing geological dynamic feature conversion according to the slot geological data to obtain geological dynamic feature sequence data;

[0108] In one embodiment, the slot geological data needs to be converted into dynamic feature sequence data suitable for calculation. First, the geological data is preprocessed, such as filtering abnormal data, removing noise, and filling missing data by interpolation. Then, the geological data is segmented according to the changes in the soil layer structure and the increment of the slot depth. Then, the segmented geological feature data is used to generate a dynamic feature sequence according to the depth. For example, the conversion point between the clay layer and the sand layer will be marked, and a set of sequence data increasing according to the depth will be generated, showing the soil layer type and corresponding characteristics at each depth, providing the necessary geological dynamic feature information for three-dimensional stratigraphic modeling.

[0109] In one embodiment, the slot geological data is converted into a dynamic feature sequence to represent the depth and characteristic changes of different geological layers. According to the slot depth, the slot is divided into several layers, and the soil layer type and depth range of each layer are marked. For example, 0-10 meters are classified as clay layer, 10-20 meters are classified as transition layer, and 20-30 meters are classified as sand layer. The geological layer information and groundwater level at each depth are integrated into a set of dynamic sequence data, including soil layer type, groundwater level influence, pore pressure, etc. The obtained dynamic feature sequence data are as follows: Depth 0-10 meters: soil layer type "clay layer", small groundwater level influence, high pore pressure. Depth 10-20 meters: transition layer, medium groundwater level influence, medium pore pressure. Depth 20-30 meters: sand layer, large groundwater level influence, low pore pressure.

[0110] S14, performing virtual reality formation modeling according to the mud property clustering model and the quality dynamic characteristic sequence data to obtain a three-dimensional formation characteristic model;

[0111] In one embodiment, a mud property clustering model and geological dynamic feature sequence data are combined to construct a three-dimensional stratigraphic model. First, the slot depth is divided into multiple virtual layers, and the depth of each layer is determined according to the mud property distribution and the change of geological characteristics. Then, the mud property clustering model data is input on the modeling platform to assign values ​​to the mud properties of each layer; at the same time, the geological dynamic feature sequence data is imported into the model so that each layer contains soil layer information and corresponding mud properties. The three-dimensional model decomposes the slot into multiple three-dimensional cells through gridding, and each cell has independent mud and soil layer characteristics. After the model is visualized, the system outputs a three-dimensional stratigraphic feature model to achieve a virtual reproduction of the real environment of the slot.

[0112] S15, calculating the mud characteristic distribution depth according to the three-dimensional formation characteristic model to obtain mud characteristic distribution depth data;

[0113] In one embodiment, the mud property distribution depth is calculated based on the three-dimensional formation model to obtain mud property data at different depths. First, each cell in the three-dimensional model is scanned layer by layer to extract the corresponding mud property value and depth information. Then, these data are sorted by depth to generate a mud property distribution depth table. For example, if the mud density fluctuates at different depths, the model automatically marks these changes and saves the density values ​​at each depth as mud property distribution depth data. This data directly reflects the distribution of mud in the slot, providing a reliable basis for slot dynamic pressure simulation.

[0114] S16. Perform slot hole dynamic pressure simulation on the mud characteristic distribution depth data to obtain slot hole pressure data.

[0115] In one embodiment, the pressure data of the slot is obtained by simulating the pressure distribution of the mud in the slot. Based on the mud characteristic distribution depth data, the mud hydrostatic pressure at each depth is first calculated. For example, the mud pressure of each unit cell is calculated using the hydrostatic pressure formula, taking into account the changes in gravity acceleration and mud density. Then, the units at different depths in the slot model are iteratively calculated to obtain the overall pressure distribution. The calculation results are integrated to generate the slot pressure data, forming a depth-pressure relationship curve, which lays the foundation for the mud wall stability analysis.

[0116] Optionally, the mud property modeling includes:

[0117] Extract mud characteristic features according to the mud characteristic data to obtain mud characteristic feature data;

[0118] In one embodiment, characteristic data for describing mud behavior are extracted by analyzing key variables in mud characteristic data. Mud characteristic data include information such as density, viscosity, rheological parameters (such as yield stress, plastic viscosity) and permeability of mud. Main variables reflecting mud characteristics are selected from mud data, including density, viscosity, yield stress, plastic viscosity and permeability, etc., to ensure that each variable can effectively characterize mud characteristics. Statistical information such as the mean, maximum, minimum and standard deviation of the selected variables are calculated, and these characteristics are organized into a structured data table, where each row represents the characteristics of mud at different depths and each column is a specific mud characteristic.

[0119] Performing feature coding according to the mud characteristic feature data to obtain mud characteristic coding data;

[0120] In one embodiment, each variable in the feature data is normalized so that the data is within the same dimensional range. For non-numeric variables (such as soil layer type), one-hot encoding is used to convert categorical variables into numeric types to adapt to the clustering algorithm. The encoded data is integrated into a new table to form a standardized feature encoding data table. At this time, each data point is represented in a numerical way to facilitate subsequent calculations.

[0121] Perform preliminary clustering on the mud characteristic coding data to obtain mud characteristic clustering data;

[0122] In one embodiment, the encoded mud characteristic data is preliminarily clustered to identify mud data points with similar characteristics. By calculating the similarity between the data points, a distance matrix of the mud characteristic data is constructed. The similarity can be calculated using the Euclidean distance, and the distance between each pair of data points represents the degree of similarity between the two. According to the distance matrix, data points with higher similarity are assigned to the same cluster. This preliminary clustering process can use an iterative allocation method to gradually assign all data points to different classes. The preliminary cluster number to which each data point belongs is recorded to form mud characteristic cluster data.

[0123] According to the clustering data of mud characteristics, clustering scale is adjusted and clustering boundary is multi-scaled to obtain multi-scale clustering data of mud characteristics;

[0124] In one embodiment, by adjusting the scale parameters of clustering and multi-scaling of boundaries, the clustering effect is further optimized to make the boundaries between each cluster group clearer. The preliminary clustering results are analyzed, and adjacent cluster groups that are close but not classified as the same are merged. The scale is adjusted to a larger range to ensure that similar data points are classified into the same cluster. The boundary data points of the cluster boundary are fuzzy processed so that each cluster has a multi-level boundary. For example, the area close to the core cluster point is set as a high confidence boundary, the area slightly farther is set as a medium confidence boundary, and the outermost boundary is set as a low confidence. The boundary and core area of ​​each cluster are recorded, and the boundary confidence of each layer is marked to form multi-scale clustering data of mud characteristics containing multi-scale information.

[0125] The multi-scale clustering data of mud properties is shown in Table 1 as follows:

[0126] Depth (m) Preliminary clustering Core Boundary Near the border Far Boundary 1-5 Category 1 1-3 meters 4-5 meters 10-15 meters 5-10 Category 2 5-8 meters 9-10 meters

[0127] Performing label association on mud characteristic data according to multi-scale clustering data of mud characteristics to obtain mud characteristic clustering label association data;

[0128] In one embodiment, multi-scale clustering data is labeled and associated with mud property data so as to assign clear labels to different cluster groups. Based on the multi-scale distribution results of clustering, each cluster is labeled. For example, a cluster with higher density and higher viscosity is labeled as "high density-high viscosity mud". The cluster number of each data point is converted to its corresponding cluster label. For example, if the cluster number "1" represents "high density-high viscosity mud", the cluster number "1" is converted to the corresponding text label. The features of each data point are integrated with the cluster label to form a mud property cluster label association data table, which contains the label information of each mud property data point.

[0129] The model is constructed based on the associated data of mud property clustering labels to obtain the mud property clustering model.

[0130] In one embodiment, a mud property clustering model is established based on label association data for predicting and classifying new data. Based on the existing label association data, a clustering model is trained to enable it to recognize mud property patterns of different labels. During the training process, the model learns the characteristic distribution of each cluster group so that it can make correct classifications when encountering new data. The trained model is verified, and a portion of the data is used as a test set to calculate the classification accuracy and robustness of the model. By adjusting parameters, it is ensured that the model can effectively distinguish the mud property categories. The trained and verified model is saved as a mud property clustering model, which can input new mud property data to automatically perform clustering and label classification.

[0131] Optionally, the feature encoding includes:

[0132] Performing local linear embedding processing on the mud characteristic feature data to obtain mud characteristic feature mapping data;

[0133] In one embodiment, the characteristic data of mud properties are standardized, and the data include density, viscosity, permeability and other characteristics, and each characteristic is converted into standardized data to eliminate the influence of different characteristic dimensions. A number of neighboring points k is selected (for example, 5), and the nearest neighbors of each data point are calculated, and each data point is connected with its neighbors to form a local neighborhood. For each data point, it is reconstructed in the neighborhood, and the degree of influence of other data points in the neighborhood on it in the local geometric structure is calculated. Specifically, the reconstruction weight of each point in the neighborhood is found by minimizing the reconstruction error. The reconstruction weight matrix is ​​input into the low-dimensional space so that it maintains the same local structural relationship as the high-dimensional space, thereby obtaining low-dimensional mud property feature mapping data. The mapping data can effectively retain the key structural information in the original mud properties, which is convenient for further feature abstraction.

[0134] Perform characteristic abstraction and stratification according to the mud characteristic feature mapping data to obtain mud abstract characteristic tensor data;

[0135] In one embodiment, the mud characteristic mapping data after dimensionality reduction is further abstracted in layers to form a multi-level abstract characteristic tensor of the mud. Based on the new dimension feature values ​​in the feature mapping data, it is divided into multiple levels, each corresponding to a different level of abstraction. For example, the characteristic mapping data has two main characteristic dimensions after dimensionality reduction, and they can be divided into a low characteristic layer, a medium characteristic layer, and a high characteristic layer according to the value range. The data of different layers are integrated into a multi-dimensional tensor structure. Each tensor layer represents the change of mud characteristics at different abstract levels. For example, the density characteristic can be in the first dimension of the tensor, and the viscosity characteristic can be in the second dimension, so as to construct a multi-dimensional abstract characteristic tensor of the mud. The characteristic values ​​of the corresponding level are filled in the tensor structure. Data points of different abstract levels such as depth, density, viscosity, etc. are filled in the corresponding tensor layers to obtain the abstract characteristic tensor data of the mud.

[0136] In one embodiment, the values ​​of feature 1 and feature 2 are divided into multiple abstract layers according to the eigenvalue range. For example, eigenvalues ​​0 to 0.5 are set as "low-level characteristics" and 0.5 to 1 are set as "high-level characteristics". These hierarchical characteristics are organized into a tensor structure. For example, feature 1 and feature 2 represent the density and permeability characteristics of mud, respectively, and a 2x3 tensor (2-dimensional feature x3 level) is constructed: level 1: low density-low permeability, level 2: high density-low permeability, level 3: high density-high permeability, and each data point is mapped to the corresponding tensor level according to its eigenvalue range, and the following mud abstract characteristic tensor data is obtained:

[0137] The mud abstract characteristic tensor data table 2 is as follows:

[0138] Tensor Hierarchy Feature 1 Density Feature 2 Permeability Level 1 0.3 0.7 Level 2 0.8 0.4 Level 3 0.5 0.9

[0139] Perform pre-coding processing on the mud abstract characteristic tensor data to obtain mud compression characteristic data;

[0140] In one embodiment, the mud abstract characteristic tensor data is preprocessed before encoding to compress features and reduce data redundancy to obtain streamlined mud compressed characteristic data. The mud characteristic tensor data is subjected to dimensionality reduction processing to remove redundant information. Principal component analysis (PCA) or other dimensionality reduction methods are used to compress dimensions with smaller variations in the tensor data to reduce the complexity of the characteristics. In the reduced dimensionality tensor data, the main features are selected based on the importance of the features. The variance of each feature at different levels can be calculated, and features with larger variations are retained, while unimportant features are removed. The reduced and selected features are reconstructed into a compressed feature matrix of the mud. This matrix has fewer characteristic dimensions, but retains the most important information about the mud characteristics.

[0141] In one embodiment, the characteristic tensor is reduced in dimension, and the dimension with the largest variation is retained. For example, the variance of each tensor level is calculated, and the dimension with the largest variance is retained to reduce the complexity of the feature dimension. The features with the largest variance in the tensor are selected for combination. If the variation of feature 1 is large, and the variation of feature 2 is small, feature 1 is mainly retained. The tensor data after dimension reduction is reorganized into a matrix to obtain the mud compression feature data.

[0142] The mud compression characteristic data is recursively encoded to obtain mud characteristic encoding data.

[0143] In one embodiment, the compressed feature data is recursively encoded to generate mud characteristic encoded data. The recursive encoder is initialized, and the number of recursive layers and the input dimension are set. For example, the number of encoding layers is set to 3 layers to gradually extract high-level features in the data. In each layer, the input mud compressed feature data is further encoded, and the high-dimensional feature space is mapped to the low-dimensional space layer by layer. The output of each layer is used as the input of the next layer, so that the high-level features in the data gradually appear. Recursive coding captures the complex patterns hidden in the mud characteristics through nonlinear transformation. The final encoding result, i.e., mud characteristic encoded data, is output at the last layer of recursive coding. This data contains mud characteristic data that has undergone multiple layers of abstraction and compression, and can effectively express the core characteristic pattern of the mud.

[0144] Optionally, the cluster scale adjustment and cluster boundary multi-scaling include:

[0145] Initialize the clustering scale according to the clustering data of mud characteristics to obtain the initial scale data of mud characteristics;

[0146] In one embodiment, the initial scale setting is performed using the clustering data of mud properties to determine the distribution range and density of the initial clusters. Based on the distance distribution from the data points in the cluster to the cluster center, the average distance is calculated as the initial scale radius of each cluster. The radius is set to the mean of the distance from all data points to the center point. For example, if the average distance from the center point in a cluster is 3 meters, the initial scale radius is 3 meters. The density of data points per unit area in each cluster is calculated to provide a basis for the subsequent identification of the density center. The density distribution is obtained by counting the density of the number of points in each initial scale. After obtaining the initial scale data of mud properties, the initial scale range setting of each cluster is completed, laying the foundation for the next step of density center identification.

[0147] Density center identification is performed based on initial scale data of mud characteristics to obtain cluster center data of mud characteristics;

[0148] In one embodiment, high-density center points within the cluster are identified, and these density center points will serve as the core for further expansion and boundary adjustment. Expand outward from each initial cluster center, and gradually calculate the change in point density at different distances. For example, the point density is calculated at multiple levels with the cluster center as a radius of 2 meters, 3 meters, 4 meters, etc. When the density at a certain distance reaches a peak, the distance is marked as a high-density center point as the basis for expansion. The density is highest at a radius of 3 meters in a cluster, so the density center radius is set to 3 meters. The density peak position of each cluster is marked as the cluster center, which serves as the core point for the next scale expansion.

[0149] The scale of the cluster center data of mud characteristics is gradually expanded to obtain the expanded scale data of mud characteristics;

[0150] In one embodiment, the scale of each cluster is gradually expanded to capture potential boundary features and the distribution of adjacent clusters. Starting from the cluster center, the radius is gradually expanded, and the expansion distance is increased to include more data points. For example, each time 1 meter is increased, the number of new data points within the expansion radius is calculated. After each expansion, the density of the new area is calculated, and the density distribution after each layer of expansion is recorded. If the density drops significantly after expansion, the expansion is stopped; otherwise, the expansion is continued until the density gradually stabilizes. The radius of each expansion and its density are recorded to form the expansion scale data. If the density is stable when the expansion radius reaches 6 meters, 6 meters is recorded as the expansion radius. The obtained mud property expansion scale data contains the density changes of each cluster from the center to the boundary.

[0151] According to the mud property extended scale data, the cluster center change rate is calculated and the cluster change characteristics are extracted to obtain the mud property cluster center change rate data and the mud property cluster change characteristic data;

[0152] In one embodiment, the expansion change of each cluster is analyzed to extract the core features of the change in cluster characteristics. At each expansion scale, the density change rate of each layer is calculated. The change rate formula can be set as: the ratio of the density difference of adjacent layers to the expansion radius. By observing the increase and decrease trend of the change rate, the boundary of significant density change is identified. The point with the most significant density change during the expansion process is recorded as the change feature of the cluster. For example, when the change rate is the highest peak when expanded to 3 meters, 3 meters is marked as the change feature. The change features of each cluster are organized into a table to form cluster center change rate data and cluster change feature data.

[0153] According to the mud property cluster center change rate data and the mud property cluster change characteristic data, the boundary of the mud property extended scale data is adjusted to obtain the cluster property multi-scale adjustment data;

[0154] In one embodiment, the boundaries of each cluster are adjusted according to the change rate and change characteristics to ensure that the boundaries reflect the true distribution of the clusters. Based on the change rate data, the boundaries of each cluster are identified. The location where the change rate is significantly reduced is used as the boundary stop point. For example, when the change rate is reduced to less than 10% when extended to 4 meters, 4 meters is marked as the boundary. Near the boundary, the boundary is corrected according to the distribution characteristics in the change characteristic data so that the boundary can cover the area where the density decreases more slowly. Through hierarchical boundary adjustment, boundary ranges of different radii are formed, such as "core boundary", "near boundary" and "far boundary", indicating the multi-scale distribution of multi-level boundaries.

[0155] In one embodiment, the cluster boundary is adjusted at multiple scales based on the aforementioned change rate and change characteristic data to make the boundary more consistent with the data distribution characteristics. Use the change rate data to mark the boundary. Assume that class A sets the main boundary at 5 meters where the change rate drops significantly, and class B sets the main boundary at 6 meters. Add a layer of far boundary outside the main boundary to form a multi-scale boundary. For example, class A has a core boundary at 5 meters and an extended boundary at 6 meters. For areas where the density change decreases slowly, the boundary is allowed to extend moderately. For example, in areas where the change rate is below -0.1, the boundary extends 1 meter. The obtained cluster characteristic multi-scale adjustment data is: Class A: core boundary 5 meters, extended boundary 6 meters, Class B: core boundary 6 meters, extended boundary 7 meters.

[0156] The multi-scale adjustment data of clustering characteristics are used to refine dense areas and obtain multi-scale clustering data of mud characteristics.

[0157] In one embodiment, the dense area is refined according to the multi-scale adjustment data, and the dense subgroups in the cluster are identified to obtain more accurate multi-scale clustering data. Within the multi-scale boundary, the density value of each sub-area is calculated, and the area with a density higher than a certain threshold is identified as a dense area. For example, the area with a density greater than 0.8 within a radius of 3 meters is regarded as a core dense area. The interior of the dense area is further divided to form several subgroups. Similar dense data points are clustered into a group to refine the internal structure of the dense area. The dense subgroup in each cluster is marked as a "core dense subgroup", and the "extended boundary subgroup" far away from the core is identified to form multi-scale clustering data of mud properties.

[0158] In one embodiment, within the core boundary of each cluster, the density of dense subgroups is calculated, and areas above a density threshold are identified as dense subgroups. For example, within a 5-meter radius of class A, the density is greater than 1.0 point / meter 2The dense area is marked as a dense area. The dense area is refined and the area with density higher than the threshold is divided into subgroups. There are two high-density subgroups at the radius of 3 meters and 4 meters in the dense subgroup of class A. The dense subgroup of class A is marked as the core subgroup, and the extended boundary is the secondary dense area. The dense area of ​​class B is refined into multi-level subgroups such as core subgroups and secondary core subgroups. The obtained multi-scale clustering data of mud characteristics are as follows: Class A: core subgroup (density 1.2 points / m 2 , radius 3-4 meters), secondary dense area (5 meters), extended boundary (6 meters), Category B: core subgroup (density 0.9 points / m 2 , radius 4-5 meters), secondary core subgroup (6 meters), and extended boundary (7 meters).

[0159] Optionally, S2 includes:

[0160] S21, performing pressure multi-dimensional deconstruction map processing according to the slot pressure data to obtain mud multi-dimensional pressure deconstruction map data;

[0161] In one embodiment, the distribution of mud pressure at different depths and directions is visualized by multi-dimensionally deconstructing the slot pressure data to obtain a multi-dimensional pressure deconstruction diagram of the mud. The slot pressure data is sorted, and the pressure values ​​at different depths and the data of the corresponding directions are classified according to the dimensions of depth, direction, etc. For example, the slot depth is 20 meters, which is divided into one layer per meter, and the pressure values ​​of each layer are measured in four directions. A three-dimensional graph is used to represent the depth, direction, and pressure value as a deconstruction diagram, with the depth as the vertical axis, the pressure value as the horizontal axis, and the pressure changes in different directions represented by colors or lines, to generate a multi-dimensional pressure deconstruction diagram of the mud, which intuitively shows the changes of pressure with depth and direction. For example, the pressure data collected in four directions (east, south, west, and north) at depths of 5 meters, 10 meters, 15 meters, and 20 meters are as follows: Depth 5 meters: East 20kPa, South 18kPa, West 19kPa, North 21kPa, Depth 10 meters: East 25kPa, South 23kPa, West 24kPa, North 22kPa, Depth 15 meters: East 30kPa, South 28kPa, West 29kPa, North 27kPa, Depth 20 meters: East 35kPa, South 33kPa, West 34kPa, North 32kPa. The data is displayed in a three-dimensional graph, with the vertical axis being depth, the horizontal axis being pressure, and the color representing direction. The generated graph can help observe the distribution characteristics of pressure at different depths and directions.

[0162] S22, performing slot hole depth pressure correlation analysis based on the mud multi-dimensional pressure deconstruction map data to obtain mud pressure data;

[0163] In one embodiment, a correlation analysis is performed on the depth variation of the mud pressure to determine the regular distribution of pressure at different depth levels. The pressure value of each depth layer is extracted from the deconstruction map data, the pressure difference between adjacent depth layers is calculated, and a depth-pressure variation correlation curve is constructed. For example, the pressure change is calculated for every 1 meter of depth, and the average pressure and variation trend of each depth layer are recorded. By calculating the standard deviation, mean and other statistics of the pressure change of each layer, the law of pressure variation in depth is determined. The layers with large pressure variation amplitudes are marked as pressure fluctuation areas, and the layers with smaller changes are marked as stable areas. The analysis results are organized into a table to show the distribution characteristics of pressure at different depths, so as to identify sensitive points and conduction characteristics.

[0164] In one embodiment, according to the pressure deconstruction diagram, the change of pressure in the depth direction is analyzed to identify the correlation of pressure. The pressure difference of adjacent layers is calculated for each depth layer to construct a depth-pressure change curve. For example: depth 5 meters to 10 meters: pressure change in the east direction = 25-20 = 5kPa, depth 10 meters to 15 meters: pressure change in the east direction = 30-25 = 5kPa, depth 15 meters to 20 meters: pressure change in the east direction = 35-30 = 5kPa. Calculate the average value and variance of the pressure change at each depth to determine the law of pressure change between different depths. If the variance is less than a certain threshold, it means that the pressure is relatively stable within the depth range.

[0165] S23, identifying pressure sensitive points according to the mud pressure data to obtain mud pressure sensitive point data;

[0166] In one embodiment, the sensitive points of mud pressure at each depth are identified to determine the depth area that is most sensitive to pressure changes. A threshold range for pressure fluctuations is set, for example, a pressure change of more than 10% in adjacent layers is defined as a sensitive point. According to the rate of change of adjacent depths in the pressure data, areas with significant pressure changes are screened out and marked as sensitive points. For example, at depths of 5 meters, 10 meters, and 15 meters, the pressure change rate exceeds 10%, and these points are marked as pressure sensitive points. The position, pressure value, and change rate of each pressure sensitive point are recorded in a table to form a pressure sensitive point data table.

[0167] S24, performing stress conduction mapping on mud pressure sensitive point data to obtain mud stress conduction mapping data;

[0168] In one embodiment, the stress conduction path is simulated based on the pressure sensitive point data, and the propagation of pressure from the sensitive point to the surrounding area is analyzed. A stress conduction model is established with the sensitive point as the center. It is assumed that the pressure is gradually transmitted to the surrounding strata through the mud, and the pressure decreases layer by layer during the conduction process. The pressure conduction value in each direction around the sensitive point is calculated. According to the mud density and the formation characteristics, the stress conduction distance and the decrease rate are calculated. For example, the conduction pressure decreases by 20% for every 1 meter of expansion. The conduction path is added to the deconstruction diagram, and the pressure decrease value in each direction is marked. At the sensitive point at a depth of 5 meters, the pressure value transmitted to 7 meters decreases to 60% of the original value, forming a multi-level stress conduction map.

[0169] S25, performing slot stress response simulation according to the mud stress conduction mapping data to obtain slot stress response data;

[0170] In one embodiment, based on the data of stress conduction mapping, the overall stress response of the slot is simulated, and the stress state at different depths is analyzed. At each depth point in the slot, the conduction stresses in all directions are superimposed to calculate the total stress at that depth. Assuming that a certain depth is affected by stress conduction in multiple directions at the same time, the total value of the superimposed pressure is calculated. In the depth-stress curve, the total stress value after each layer of superposition is recorded to generate a stress response curve. For example, after the pressure of sensitive points at 5 meters, 10 meters, and 15 meters is superimposed, the overall slot wall shows a multi-level stress response. According to the stress response curve, the peak and valley values ​​of the stress change are analyzed, and the high-pressure area and low-pressure area of ​​the slot stress distribution are marked.

[0171] S26. Screen the stress concentration area according to the slot hole stress response data to obtain the slot hole wall stress data.

[0172] In one embodiment, stress concentration areas on the slot wall are identified based on stress response data, and high stress areas in the slot that may have a risk of instability are screened out. A stress concentration threshold is set. For example, if the stress value at a certain depth exceeds 1.5 times the average stress, it is marked as a stress concentration area. Based on the stress response data, the depths where the stress value exceeds the threshold are screened out. Assuming that the stress values ​​at 10 meters and 15 meters exceed the threshold, they are marked as stress concentration areas. The location, stress value, conduction path and other information of all concentrated areas are integrated to form slot wall stress data, which shows the distribution of high stress on the slot wall.

[0173] Optionally, the stress transmission mapping includes:

[0174] According to the mud pressure sensitive point data, the mechanical characteristics of the sensitive point are analyzed to obtain the mechanical characteristics matrix data of the sensitive point;

[0175] In one embodiment, the mechanical properties of the mud pressure sensitive points are analyzed to form mechanical property matrix data including soil layers and mud properties around the sensitive points. Mechanical properties such as the position, pressure value, surrounding mud density, viscosity, etc. of each sensitive point are extracted from the pressure sensitive point data. The mechanical property parameters of each sensitive point are calculated according to the properties of each sensitive point, such as the yield stress, shear modulus, and permeability of the mud. For example, the mud density of a sensitive point is 1.2 g / cm 3 , the viscosity is 30mPa×s, then the shear modulus and other mechanical parameters of the point are calculated. The mechanical properties of all sensitive points are organized into a matrix, the rows of the matrix represent different sensitive points, and the columns represent the different mechanical properties of each sensitive point. The formed matrix data can represent the mechanical state of the sensitive point. The obtained sensitive point mechanical property matrix data includes the density, viscosity, shear modulus and other properties of the sensitive point, providing comprehensive mechanical information for each sensitive point.

[0176] Initialize the stress field according to the mechanical property matrix data of the sensitive points to obtain the initial stress field distribution map data;

[0177] In one embodiment, based on the mechanical properties of the sensitive points, the stress field of the entire slot is initialized to obtain an initial stress field distribution diagram. Using the data in the mechanical property matrix, the initial stress distribution of the sensitive point and its surroundings is calculated according to the pressure and shear modulus. For example, the initial pressure of a sensitive point is 20kPa, and the shear modulus is 1MPa. The initial stress field of the surrounding area is estimated based on its position and mechanical properties. The slot space is divided into several small grid units, and the stress value of each unit is obtained by interpolation of the mechanical properties of adjacent sensitive points. An initial stress field covering the entire slot is generated by the interpolation method. A three-dimensional distribution diagram is formed to display the stress values ​​at each depth and direction, and to intuitively display the initial distribution of stress in the slot. The obtained initial stress field distribution diagram data shows the initial distribution of stress in the slot.

[0178] Conduct adjacent pressure conduction simulation on the initial stress field distribution map data to obtain pressure conduction sequence data;

[0179] In one embodiment, based on the initial stress field distribution diagram, the conduction process of pressure from a sensitive point to an adjacent area is simulated. Starting from the sensitive point, the conduction path is extended to the adjacent area layer by layer, and each layer conduction decreases a certain pressure. If the decreasing coefficient is 0.9, the pressure decays to 90% of the pressure of the previous layer when each layer is expanded. For example, if the initial pressure of the sensitive point is 20kPa, the adjacent conduction pressure of the first layer is 20kPa×0.9=18kPa, and the second layer is 18kPa×0.9=16.2kPa, which decreases successively. The pressure values ​​of the adjacent layers of each sensitive point are recorded layer by layer to generate pressure conduction sequence data to show the conduction process from the sensitive point to the adjacent area. The obtained pressure conduction sequence data shows the conduction process of decreasing pressure from the sensitive point to the adjacent area, which provides a basis for optimizing the conduction path.

[0180] Optimize the conduction path according to the pressure conduction sequence data to obtain optimized conduction sequence data;

[0181] In one embodiment, based on the preliminary pressure conduction sequence, the conduction path is optimized to make the pressure conduction more reasonable. The effectiveness of each conduction path is analyzed, and the paths with higher pressure and shorter conduction paths are retained to reduce invalid long-distance conduction. For example, if the pressure of a certain path decays to below 5kPa within two layers, it is considered an invalid path. The path is re-optimized based on the screening results, and the pressure attenuation coefficient of each node in the path is adjusted to retain the effective conduction path. For example, the attenuation coefficient of the short-distance path is set to 0.95 to retain more pressure values, form the optimized conduction sequence data, and record the pressure attenuation of each path and the conduction path length. The obtained optimized conduction sequence data includes the optimized pressure conduction path, so that the pressure is better distributed on the effective path.

[0182] Perform geological stress attenuation simulation based on optimized conduction sequence data to obtain stress attenuation distribution matrix data;

[0183] In one embodiment, the attenuation of pressure in a geological environment is simulated to form a distribution matrix of stress attenuation. The stress attenuation model is set according to the stratum structure. For example, the stress attenuation is slower in the clay layer, and the attenuation coefficient is set to 0.9; the attenuation is faster in the sand layer and is set to 0.7. According to the path in the optimized conduction sequence, the pressure attenuation is calculated layer by layer along the path direction. After the path passes through 2 meters in the clay layer, the pressure value decays from 20kPa to 20×0.9×0.9=16.2kPa. The pressure attenuation value of each layer in each conduction path is recorded to form a stress attenuation distribution matrix, which shows the gradual attenuation of stress in different strata.

[0184] According to the stress attenuation distribution matrix data, the initial stress field distribution map data is adjusted for stress conduction boundary to obtain stress conduction boundary data;

[0185] In one embodiment, based on the stress attenuation distribution, the stress conduction boundary is adjusted to ensure the rationality of pressure conduction. The boundary threshold of pressure conduction is set, and the area below 2kPa no longer conducts pressure conduction, and it is marked as a boundary area. According to the attenuation matrix, all areas below the boundary threshold are screened out and set as the conduction termination point. If the pressure decays to 1.8kPa after a path passes through 5 meters in the sand layer, 5 meters is used as the conduction boundary. The end point position of each path is recorded to form stress conduction boundary data to display the boundary conditions of the conduction path.

[0186] The mud stress conduction mapping data is obtained by mapping and reconstructing the optimized conduction sequence data, stress attenuation distribution matrix data and stress conduction boundary data.

[0187] In one embodiment, the aforementioned data are integrated to complete the stress conduction mapping of the mud. The optimized conduction sequence, stress attenuation matrix and conduction boundary data are integrated into the same mapping framework, and the pressure values ​​of each path and boundary are reconstructed together. A three-dimensional stress conduction map is generated based on the integrated data to show the stress conduction process of the sensitive point in the entire slot. The area with high stress is marked as a high conduction area, and the area attenuated to the boundary is marked as a low conduction area. The area with large pressure fluctuations in the mapping map is smoothed to enhance the readability of the conduction mapping.

[0188] Optionally, S3 includes:

[0189] S31, performing stress distribution anomaly detection according to the slot wall stress data to obtain stress anomaly point data;

[0190] In one embodiment, by analyzing the slot wall stress data, abnormal points in the stress distribution are detected to identify areas of stress concentration or sharp changes. A threshold for determining stress anomalies is set based on the mean and standard deviation of the slot wall stress data. For example, if the average stress is 30 kPa and the standard deviation is 5 kPa, the threshold is set to 1.5 times the mean value (i.e., 45 kPa) or twice the standard deviation (i.e., 40 kPa), and all points above the threshold are abnormal points. Check the stress value of each wall point one by one, and mark the points above the abnormal threshold. If 50 kPa and 55 kPa are detected at two adjacent points at a certain depth, the two points are marked as abnormal points. Record all points marked as abnormal, including depth position, stress value, and surrounding stress environment for subsequent analysis.

[0191] S32, performing instability area analysis based on stress abnormal point data to obtain instability area data of the slot wall;

[0192] In one embodiment, based on the detected stress anomalies, possible instability areas are further analyzed to determine the specific distribution range of the instability areas in the slot wall. According to the distribution of stress anomalies, areas with continuous anomalies or high stress values ​​exceeding a certain range are set as unstable areas. If three or more anomalies appear within a continuous depth of 5 meters, the depth range is regarded as a potential unstable area. The abnormal points that meet the conditions are clustered into unstable areas. For example, if there are multiple anomalies within a depth range of 10 meters to 15 meters, this depth range is clustered into an unstable area. The location, depth range and stress value of the unstable area are recorded for trend analysis, and the obtained data on the instability area of ​​the slot wall indicates the specific location and stress characteristics of the unstable area.

[0193] S33, performing instability trend analysis based on the instability area data of the slot hole wall to obtain instability trend data of the slot hole wall;

[0194] In one embodiment, the distribution characteristics and stress change trends of the unstable area are analyzed to predict the future expansion or contraction trend of the unstable area. The stress change trend of each unstable area is extracted, including the stress increase and decrease rate of each abnormal point. Assuming that in the unstable area at a depth of 12 meters, the stress gradually increases from 30kPa to 50kPa in the past period of time, the increase is 20kPa. Linear fitting is performed on the stress change rates of adjacent points in the unstable area to determine the overall stress change trend of the area. If the trend is positive, that is, the stress continues to rise, then the area has a tendency to expand instability; if it is negative, the instability is weakened. The analyzed trend results are organized into a table to record the stress change rate and instability expansion direction of each unstable area.

[0195] S34, performing instability region depth identification on the slot wall instability trend data to obtain instability region identification data;

[0196] In one embodiment, by analyzing the instability trend data, the depth range of the instability area is identified, and the potential expansion depth of the instability area is determined. The depth range of the instability area is determined according to the stress increase and decrease in the instability trend data. For example, if the stress increase from 12 meters to 15 meters is 20kPa, and the stress increase from 15 meters to 18 meters is 10kPa, then 12 meters to 18 meters is marked as the instability range. If the trend analysis shows that the area is still expanding, a certain expansion range is reserved below the existing instability depth. For example, if the stress from 15 meters to 18 meters continues to increase, the instability depth is further extended to 20 meters. The depth range, current expansion situation and extension depth of each instability area are recorded as a data table to form instability area identification data.

[0197] S35. Dynamically map the unstable region according to the unstable region identification data to obtain the unstable region data of the slot wall.

[0198] In one embodiment,

[0199] Optionally, S4 includes: generating a dynamic mapping diagram of the unstable area according to the depth range and expansion trend of the identified unstable area to intuitively display the distribution of the unstable area. Integrate the depth range, stress value and trend in the unstable area identification data into the same mapping framework. Generate a three-dimensional dynamic mapping diagram of the slot wall based on the identification data, and mark the unstable area with color or line. Unstable areas of different depths and trends can be distinguished by color depth. For example, high stress concentration areas are marked with dark red, and unstable expansion areas are marked with orange. The mapping diagram is updated in real time, and the color and range of the unstable area are adjusted over time to reflect the dynamic changes of the instability trend. The generated dynamic mapping diagram of the unstable area of ​​the slot wall shows the distribution and expansion trend of the unstable area, providing visual reference data for stability monitoring and risk control of the slot.

[0200] S41, extracting stability characteristics according to the instability area data of the slot hole wall and the stress data of the slot hole wall to obtain stability characteristic data;

[0201] In one embodiment, stability features are extracted by analyzing the instability area data and wall stress data of the slot wall, laying the foundation for the subsequent stability index calculation. The depth, stress value, expansion trend and other characteristics of the instability area are analyzed. The depth of the instability area is 8-12 meters, the average stress is 45kPa, and the stress change trend is rising. The distribution and change law of the wall stress are analyzed, including the average stress, stress gradient and maximum stress value. For example, the average stress at each depth of the slot wall is 30kPa, and the stress gradient is 5kPa / meter. Various features are organized into data tables to form stability feature data, covering information such as the depth of the instability area, stress trend, average wall stress and gradient.

[0202] Stability characteristic data table 3 is as follows:

[0203] feature value Depth of unstable area 8-12 meters Mean stress 45kPa Stress change trend rise Average wall stress 30kPa Stress gradient 5kPa / m

[0204] S42, calculating the stability index according to the stability characteristic data to obtain the slot area stability index data;

[0205] In one embodiment, the stability index can be defined as the ratio of the stress in the unstable region to the average stress on the wall. The index formula is: S is the stability index data of the slot area, Q is the average wall stress, K is the average stress in the unstable area, and C is the stress gradient. Substitute the characteristic data into the formula. For example, the average wall stress is 30kPa, the stress in the unstable area is 45kPa, the stress gradient is 5kPa / m, and the stability index is 3.33.

[0206] S43, performing a multi-factor cross analysis based on the slot hole wall instability area data and the slot hole wall stress data to obtain a multi-factor cross analysis matrix data;

[0207] In one embodiment, multi-factor cross analysis is used to analyze the relationship between different factors that affect stability (such as stress gradient, stress change trend, etc.). Main factors such as wall stress, stress gradient, unstable area depth, stress trend, etc. are selected. Cross-comparison analysis is performed on the selected factors to construct a matrix. The rows and columns of the matrix represent different factors, for example:

[0208] The multi-factor cross analysis matrix table 4 is as follows:

[0209] factor Wall stress Stress gradient Depth of instability Stress Trend Wall stress high middle Low Stress gradient high high middle Depth of instability middle high high Stress Trend Low middle high

[0210] S44, performing stress instability risk assessment according to the slot area stability index data and the multi-factor cross analysis matrix data to obtain instability risk assessment matrix data;

[0211] In one embodiment, the instability risk of the slot is evaluated by combining the stability index and the cross analysis matrix to obtain risk assessment matrix data. The risk score is set according to the stability index range. An S value less than 3 is a high risk, 3-4 is a medium risk, and greater than 4 is a low risk. Based on the score of the stability index and the strength of the relationship in the cross matrix, the risk score of each factor is comprehensively calculated. For example, the stability index at a depth of 8-12 meters is 3.33, which is a medium risk. The instability risk assessment results of each depth are recorded to form the following matrix:

[0212] The instability risk assessment matrix table 5 is as follows:

[0213] Depth (m) Wall stress Stress gradient Depth of instability Stress Trend Overall risk assessment 8-12 middle high high middle medium

[0214] The resulting instability risk assessment matrix shows the contribution of each factor to the instability risk.

[0215] S45. Calculate the slot safety factor according to the instability risk assessment matrix data to obtain slot stability data.

[0216] In one embodiment, the safety factor can be defined as the ratio of the stability index to the total risk assessment score. The safety factor formula is: S 安 is the slot safety factor, S is the stability index, and M is the total risk assessment score. The slot safety factor is calculated based on the stability index and risk score matrix data. For example, the stability index of the 8-12 meter area is 3.33, the total risk assessment is "medium", and the corresponding score is 2. The safety factor is 1.665.

[0217] Therefore, from any point of view, the embodiments should be regarded as illustrative and non-restrictive, and the scope of the present invention is limited by the attached application documents rather than the above description, and it is intended that all changes falling within the meaning and scope of equivalent elements of the application documents are included in the present invention.

[0218] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A method for analyzing the stability of mud wall protection in anti-seepage wall slots, characterized in that: The method comprises: S1. Obtaining slot geological data and mud property data, and performing pressure simulation according to the slot geological data and mud property data to obtain slot pressure data; S2. Perform mud pressure analysis based on the slot hole pressure data to obtain mud pressure data, and perform slot hole wall stress calculation based on the mud pressure data to obtain slot hole wall stress data; S3, performing instability area processing according to the slot hole wall stress data to obtain the slot hole wall instability area data; S4. Performing slot stability analysis based on the slot wall instability area data and slot wall stress data to obtain slot stability data.

2. The method according to claim 1, characterized in that S1 includes: S11, obtaining slot geological data and mud property data through a geological engineering database or a geological engineering input interface; S12, performing mud property modeling according to the mud property data to obtain a mud property clustering model; S13, performing geological dynamic feature conversion according to the slot geological data to obtain geological dynamic feature sequence data; S14, performing virtual reality formation modeling according to the mud property clustering model and the quality dynamic characteristic sequence data to obtain a three-dimensional formation characteristic model; S15, calculating the mud characteristic distribution depth according to the three-dimensional formation characteristic model to obtain mud characteristic distribution depth data; S16. Perform slot hole dynamic pressure simulation on the mud characteristic distribution depth data to obtain slot hole pressure data.

3. The method according to claim 2, characterized in that The mud characteristics modeling includes: Extract mud characteristic features according to the mud characteristic data to obtain mud characteristic feature data; Performing feature coding according to the mud characteristic feature data to obtain mud characteristic coding data; Perform preliminary clustering on the mud characteristic coding data to obtain mud characteristic clustering data; According to the clustering data of mud characteristics, clustering scale is adjusted and clustering boundary is multi-scaled to obtain multi-scale clustering data of mud characteristics; Performing label association on mud characteristic data according to multi-scale clustering data of mud characteristics to obtain mud characteristic clustering label association data; The model is constructed based on the associated data of mud property clustering labels to obtain the mud property clustering model.

4. The method according to claim 3, characterized in that The feature codes include: Performing local linear embedding processing on the mud characteristic feature data to obtain mud characteristic feature mapping data; Perform characteristic abstraction and stratification according to the mud characteristic feature mapping data to obtain mud abstract characteristic tensor data; Perform pre-coding processing on the mud abstract characteristic tensor data to obtain mud compression characteristic data; The mud compression characteristic data is recursively encoded to obtain mud characteristic encoding data.

5. The method according to claim 3, characterized in that: The cluster scale adjustment and cluster boundary multi-scaling include: Initialize the clustering scale according to the clustering data of mud characteristics to obtain the initial scale data of mud characteristics; Density center identification is performed based on initial scale data of mud characteristics to obtain cluster center data of mud characteristics; The scale of the cluster center data of mud characteristics is gradually expanded to obtain the expanded scale data of mud characteristics; According to the mud property extended scale data, the cluster center change rate is calculated and the cluster change characteristics are extracted to obtain the mud property cluster center change rate data and the mud property cluster change characteristic data; According to the mud property cluster center change rate data and the mud property cluster change characteristic data, the boundary of the mud property extended scale data is adjusted to obtain the cluster property multi-scale adjustment data; The multi-scale adjustment data of clustering characteristics are used to refine dense areas and obtain multi-scale clustering data of mud characteristics.

6. The method according to claim 1, characterized in that S2 include: S21, performing pressure multi-dimensional deconstruction map processing according to the slot pressure data to obtain mud multi-dimensional pressure deconstruction map data; S22, performing slot hole depth pressure correlation analysis based on the mud multi-dimensional pressure deconstruction map data to obtain mud pressure data; S23, identifying pressure sensitive points according to the mud pressure data to obtain mud pressure sensitive point data; S24, performing stress conduction mapping on mud pressure sensitive point data to obtain mud stress conduction mapping data; S25, performing slot stress response simulation according to the mud stress conduction mapping data to obtain slot stress response data; S26. Screen the stress concentration area according to the slot hole stress response data to obtain the slot hole wall stress data.

7. The method according to claim 6, characterized in that The stress transmission mapping includes: According to the mud pressure sensitive point data, the mechanical characteristics of the sensitive point are analyzed to obtain the mechanical characteristics matrix data of the sensitive point; Initialize the stress field according to the mechanical property matrix data of the sensitive points to obtain the initial stress field distribution map data; Conduct adjacent pressure conduction simulation on the initial stress field distribution map data to obtain pressure conduction sequence data; Optimize the conduction path according to the pressure conduction sequence data to obtain optimized conduction sequence data; Perform geological stress attenuation simulation based on optimized conduction sequence data to obtain stress attenuation distribution matrix data; According to the stress attenuation distribution matrix data, the initial stress field distribution map data is adjusted for stress conduction boundary to obtain stress conduction boundary data; The mud stress conduction mapping data is obtained by mapping and reconstructing the optimized conduction sequence data, stress attenuation distribution matrix data and stress conduction boundary data.

8. The method according to claim 1, characterized in that S3 include: S31, performing stress distribution anomaly detection according to the slot wall stress data to obtain stress anomaly point data; S32, performing instability area analysis based on stress abnormal point data to obtain instability area data of the slot wall; S33, performing instability trend analysis based on the instability area data of the slot hole wall to obtain instability trend data of the slot hole wall; S34, performing instability region depth identification on the slot wall instability trend data to obtain instability region identification data; S35. Dynamically map the unstable region according to the unstable region identification data to obtain the unstable region data of the slot wall.

9. The method according to claim 1, characterized in that: S4 includes: S41, extracting stability characteristics according to the instability area data of the slot hole wall and the stress data of the slot hole wall to obtain stability characteristic data; S42, calculating the stability index according to the stability characteristic data to obtain the slot area stability index data; S43, performing a multi-factor cross analysis based on the slot hole wall instability area data and the slot hole wall stress data to obtain a multi-factor cross analysis matrix data; S44, performing stress instability risk assessment according to the slot area stability index data and the multi-factor cross analysis matrix data to obtain instability risk assessment matrix data; S45. Calculate the slot safety factor according to the instability risk assessment matrix data to obtain slot stability data.

Images