Detachable slurry shield slurry film forming and curing test device and intelligent evaluation method

Abstract

The application provides a detachable slurry shield slurry film forming and curing test device and an intelligent evaluation method, and relates to the technical field of slurry film forming, and comprises: a one-dimensional permeation column assembly, the one-dimensional permeation column assembly comprises an upper top plate, an upper permeation column body, a detachable permeation column body, a lower permeation column body, a lower bottom plate and a pressure gauge; a pressurizing assembly, the pressurizing assembly is in communication with a pressurizing through hole of the upper top plate; a pore water pressure monitoring assembly, the pore water pressure monitoring assembly is arranged in the one-dimensional permeation column assembly; a filtrate collecting assembly; a pressure maintaining observation assembly, a detachable permeation column placing ring is arranged in a test box body of the pressure maintaining observation assembly; and an information processing assembly, the information processing assembly is electrically connected with the pore water pressure monitoring assembly and a weighing mechanism respectively. The application improves the test efficiency and realizes accurate evaluation on the slurry film forming process and plugging effect.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of mud film forming, in particular to a detachable slurry shield mud film forming and curing test device and an intelligent evaluation method. BACKGROUND

[0002] In the technical field of slurry shield tunnel pressure opening and cutter changing operation, ensuring the integrity and stability of the air-tight mud film at the working face is the core link to ensure construction safety. In current engineering practice, indoor tests are mainly used to simulate the mud film forming process to optimize construction parameters. However, the existing test methods mostly use fixed permeation column devices for one-way film forming test. After completing a single test, the device needs to be disassembled and refilled with soil, which makes it difficult to realize continuous multiple parallel tests. Moreover, the removed mud film is easily damaged during the transfer process due to mechanical disturbance, resulting in distorted test data of the subsequent pressure maintaining performance. In addition, the data collection of the traditional test device mostly relies on manual recording and dispersed processing, and there is a lack of systematic correlation analysis between the parameters such as pressure and filtration loss obtained by different sensors, which cannot effectively reflect the internal law of the dynamic change of formation pore water pressure and mud permeability during the mud film forming process, and it is even more difficult to quantitatively evaluate the film forming quality, which to some extent restricts the improvement of mud film quality control accuracy in slurry shield construction.

[0003] Based on the above-mentioned shortcomings of the prior art, there is an urgent need for a detachable slurry shield mud film forming and curing test device and an intelligent evaluation method. SUMMARY

[0004] The purpose of the present application is to provide a detachable slurry shield mud film forming and curing test device and an intelligent evaluation method to improve the above-mentioned problems. In order to achieve the above-mentioned purpose, the technical solution adopted by the present application is as follows:

[0005] In a first aspect, the present application provides a detachable slurry shield mud film forming and curing test device, comprising:

[0006] A one-dimensional permeation column assembly, the one-dimensional permeation column assembly comprising an upper top plate, an upper permeation column body, a detachable permeation column body, a lower permeation column body, a lower bottom plate and a pressure gauge, the upper top plate and the lower bottom plate being connected by a plurality of connecting rods, the upper permeation column body, the detachable permeation column body and the lower permeation column body being coaxially arranged and sealingly connected between the upper top plate and the lower bottom plate along a center line, the pressure gauge being arranged on the upper top plate;

[0007] A pressurizing assembly in communication with the pressurizing through hole of the upper top plate;

[0008] A pore water pressure monitoring assembly arranged in the one-dimensional permeation column assembly;

[0009] A filtrate collection assembly, comprising a filtrate outlet pipe, a filtrate collection bucket, and a weighing mechanism, is disposed below the one-dimensional permeation column assembly. The filtrate outlet pipe is connected to a through hole in the lower base plate, the filtrate collection bucket is disposed below the filtrate outlet pipe, and the weighing mechanism is disposed below the filtrate collection bucket.

[0010] A pressure-holding observation assembly, wherein a detachable permeation column placement ring is provided inside the test chamber of the pressure-holding observation assembly;

[0011] An information processing component is provided, which is electrically connected to the pore water pressure monitoring component and the weighing mechanism, respectively.

[0012] Furthermore, the upper permeation column and the detachable permeation column are connected at the connection end with an upper groove, and the lower permeation column and the detachable permeation column are connected at the connection end with a lower groove. Both the upper groove and the lower groove are provided with sealing rings, and the two ends of the connecting rod are respectively connected to the upper top plate and the lower bottom plate and fixed by nuts.

[0013] Furthermore, the pressurization assembly includes an air compressor, a pressurization pipe, and a voltage regulator. The air compressor is connected to one end of the pressurization pipe, and the other end of the pressurization pipe is connected to a through hole in the upper top plate. The voltage regulator is mounted on the pressurization pipe.

[0014] Furthermore, the pressure holding observation assembly also includes an environmental control element and an upper cover plate, the upper cover plate being connected to the test chamber via a locking nut.

[0015] Furthermore, a screen is provided at one end of the detachable permeation column near the lower permeation column.

[0016] Secondly, this application also provides an intelligent evaluation method, including:

[0017] Obtain raw data, including pore water pressure time series data, test pressure data, mud ratio parameters, and filtration loss data during the film formation test;

[0018] Abnormal data filtering is performed on the original data. Abnormal data points that deviate from the main distribution are identified and removed based on the normal distribution characteristics to obtain the cleaned pore water pressure dataset.

[0019] Based on the pore water pressure dataset, data scale is unified. By calculating the relative positional relationship between each data point and the overall dataset, the magnitude difference of sensor readings at different depths is eliminated, and standardized pore water pressure data is obtained.

[0020] Based on the standardized pore water pressure data, temporal feature separation and dimensionality reduction are performed to obtain temporal features and spatial dimensionality reduction features.

[0021] Based on the sequence characteristics and spatial dimensionality reduction characteristics, a comprehensive evaluation of film formation quality is conducted to obtain a comprehensive evaluation result of mud film integrity and formation sealing effect.

[0022] The beneficial effects of this invention are as follows:

[0023] This invention enables non-destructive removal and batch maintenance of mud films through a detachable permeable column structure. It also combines multi-source data intelligent processing technology to perform fusion analysis on the temporal characteristics and spatial dimensionality reduction characteristics of pore water pressure, thereby improving experimental efficiency while achieving accurate evaluation of the mud film formation process and sealing effect. Attached Figure Description

[0024] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 This is a schematic diagram of the detachable slurry shield tunneling mud film formation and curing test device described in an embodiment of the present invention;

[0026] Figure 2 This is a cross-sectional view of the one-dimensional permeation column assembly;

[0027] Figure 3 This is a schematic diagram of the pressure holding observation component;

[0028] Figure 4 for Figure 3 AA view in the middle;

[0029] Figure 5 This is a schematic diagram of the detachable permeable column;

[0030] Figure 6 This is a flowchart of the intelligent evaluation method described in an embodiment of the present invention.

[0031] The diagram is labeled as follows: 1. One-dimensional permeation column assembly; 11. Top plate; 12. Upper permeation column; 13. Detachable permeation column; 131. Screen; 14. Lower permeation column; 15. Bottom plate; 16. Pressure gauge; 17. Connecting rod; 2. Pressurization assembly; 21. Air compressor; 22. Pressurization pipe; 23. Pressure stabilizer; 3. Pore water pressure monitoring assembly; 4. Filtrate collection assembly; 41. Outlet pipe; 42. Filtrate collection tank; 43. Weighing mechanism; 5. Pressure holding observation assembly; 51. Test chamber; 52. Environmental control element; 53. Top cover plate; 54. Detachable permeation column placement ring; 6. Information processing assembly. Detailed Implementation

[0032] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0033] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this invention, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0034] Example 1:

[0035] like Figures 1 to 4As shown, this embodiment provides a detachable slurry shield tunneling mud film formation and curing test device, which includes a one-dimensional permeation column assembly 1, a pressurization assembly 2, a pore water pressure monitoring assembly 3, a filtrate collection assembly 4, a pressure holding observation assembly 5, and an information processing assembly 6. The one-dimensional permeation column assembly 1 includes an upper top plate 11, an upper permeation column 12, a detachable permeation column 13, a lower permeation column 14, a lower bottom plate 15, and a pressure gauge 16. The upper top plate 11 and the lower bottom plate 15 are connected by multiple connecting rods 17. The upper permeation column 12, the detachable permeation column 13, and the lower permeation column 14 are arranged coaxially along the centerline between the upper top plate 11 and the lower bottom plate 15 and are sealed together. The pressure gauge 16 is located on the upper top plate 11. The purpose of this structural design is to simulate the actual geological environment, ensure that the mud film forms under controllable conditions, and facilitate the non-destructive removal of the mud film through the detachable permeation column 13, thereby improving the repeatability and efficiency of the test. The pressurizing component 2 is connected to the pressurizing through-hole of the upper top plate 11 to apply stable film-forming pressure, simulating the actual working conditions in shield tunneling and ensuring the authenticity of the mud infiltration process. The pore water pressure monitoring component 3 is installed inside the one-dimensional infiltration column component 1. It collects formation pore water pressure data in real time through pore water pressure monitoring elements, aiming to monitor the dynamic changes of water pressure during mud infiltration and provide a basis for analyzing formation response. The filtrate collection component 4 includes a filtrate outlet pipe 41, a filtrate collection tank 42, and a weighing mechanism 43. It is installed below the one-dimensional infiltration column component 1. The filtrate outlet pipe 41 is connected to the through-hole of the lower bottom plate 15. The filtrate collection tank 42 is installed below the filtrate outlet pipe 41 to collect filtrate. The weighing mechanism 43 is installed below the filtrate collection tank 42 to automatically count the filtrate loss. The pressure holding observation component 5 has a detachable infiltration column placement ring 54 in its test chamber 51, which is used to accommodate the detachable infiltration column 13 for batch curing after film formation. The information processing component 6 is electrically connected to the pore water pressure monitoring component 3 and the weighing mechanism 43 respectively, so as to realize the automatic acquisition and intelligent processing of data.

[0036] Preferably, the connecting end of the upper permeation column 12 and the detachable permeation column 13 is provided with an upper groove, and the connecting end of the lower permeation column 14 and the detachable permeation column 13 is provided with a lower groove. A sealing ring is provided in both the upper and lower grooves, and both ends of the connecting rod 17 are connected to the upper top plate 11 and the lower bottom plate 15 respectively and fixed by nuts. Further, the open end of the upper permeation column 12 is positioned above the detachable permeation column and presses against the sealing ring; the other end of the lower permeation column 14 is provided with a lower groove that fits the shape of the detachable permeation column 13 and is provided with a sealing ring; the open end of the lower permeation column 14 is positioned below the detachable permeation column and presses against the sealing ring; the open end of the lower permeation column 14 is positioned in the lower annular groove of the lower bottom plate 15 and presses against the sealing ring. Both the sealing ring of the detachable permeation column 13 and the sealing rings mentioned above are rubber sealing rings. This structural design aims to achieve an efficient sealing connection between the columns through the cooperation of the groove and the sealing ring, ensuring a stable internal pressure environment during the pressure test. At the same time, the fastening method of the connecting rod 17 and the nut provides uniform longitudinal pressure to the upper top plate 11 and the lower bottom plate 15, so that each column segment fits tightly under axial force. Thus, while ensuring the overall rigidity of the device, it supports the quick assembly, disassembly and reuse of the detachable permeable column 13, effectively improving the convenience of test operation and the reliability of assembly.

[0037] Preferably, the pressurization assembly 2 includes an air compressor 21, a pressurization pipe 22, and a pressure stabilizer 23. The air compressor 21 is connected to one end of the pressurization pipe 22, and the other end of the pressurization pipe 22 is connected to a through hole in the upper top plate 11. The pressure stabilizer 23 is installed on the pressurization pipe 22. The air compressor 21 provides a stable air source, which is delivered to the inside of the permeation column through the pressurization pipe 22. The pressure stabilizer 23 maintains a constant test pressure, aiming to accurately simulate the actual film-forming pressure conditions during shield tunneling, avoid pressure fluctuations from interfering with the mud permeation process, and thus ensure the reliability of the test data.

[0038] Preferably, such as Figure 5 As shown, the pressure holding observation assembly 5 also includes an environmental control element 52 and an upper cover plate 53, with the upper cover plate 53 connected to the test chamber 51 via a locking nut. This design precisely regulates the temperature and humidity conditions of the curing chamber through the environmental control element 52, and, in conjunction with the sealing and fixing of the upper cover plate 53, forms a controllable pressure holding environment. This aims to simulate the long-term curing conditions of mud film in actual engineering, effectively maintaining the integrity of the mud film and observing its stability evolution.

[0039] Preferably, a screen 131 is provided at one end of the detachable permeable column 13 near the lower permeable column 14. During the test, the screen 131 can both support the soil to prevent loss and ensure that the filtrate passes smoothly into the collection system. Its purpose is to maintain the integrity of the test stratum structure while accurately measuring the change in filtrate loss.

[0040] Example 2:

[0041] likeFigure 6 As shown, this embodiment provides an intelligent evaluation method, including steps S100 to S500.

[0042] Step S100: Obtain raw data, which includes pore water pressure time series data, test pressure data, mud ratio parameters and filtration loss data during the film formation test.

[0043] Understandably, pore water pressure time-series data is continuously recorded by pore water pressure monitoring elements buried at different depths within the removable permeable column, reflecting the dynamic changes in formation water pressure; test pressure data is monitored in real time by pressure gauges installed on the top plate, characterizing the stability of the film-forming pressure applied to the mud; mud mix parameters, including material properties such as clay concentration and additive ratio, are used to correlate mud properties with film-forming behavior; filtration loss data is automatically recorded by a weighing mechanism below the filtrate collection tank, quantifying the filtration loss efficiency of the mud by measuring the cumulative mass change of the filtrate during infiltration. These data collectively constitute the basic dataset for evaluating the mud film formation process, ensuring that subsequent analyses are supported by multi-dimensional parameters.

[0044] Step S200: Filter abnormal data based on the original data, identify and remove abnormal data points that deviate from the main distribution based on the normal distribution characteristics, and obtain the cleaned pore water pressure dataset.

[0045] It should be noted that, addressing the issue of sensors being susceptible to electromagnetic interference or mud particle impact in the high-pressure testing environment of slurry shield tunnels, the distribution range of the original data was defined based on the characteristics of normal distribution. By dynamically calculating the mean and standard deviation of the dataset, outliers caused by momentary equipment failures or external interference were identified and removed. This processing effectively filters out extreme points that do not conform to physical laws, retaining data sequences that truly reflect the permeability characteristics of the formation, and providing a reliable data foundation for subsequent analysis.

[0046] Step S300: Perform data scaling uniformity based on the pore water pressure dataset. By calculating the relative positional relationship between each data point and the overall dataset, eliminate the magnitude difference in sensor readings at different depths and obtain standardized pore water pressure data.

[0047] Understandably, this step takes into account the magnitude differences in pore water pressure sensors at different depths due to formation pressure gradients. By calculating the relative position of each data point to its corresponding depth level dataset, the raw readings are converted into standardized scores on a uniform scale. This processing eliminates the influence of depth on data distribution, making sensor readings at different depths comparable and avoiding analytical biases in subsequent models due to differences in the magnitude of input data.

[0048] Step S400: Perform temporal feature separation and dimensionality reduction processing based on standardized pore water pressure data to obtain temporal features and spatial dimensionality reduction features;

[0049] It should be noted that this step extracts features from both temporal and spatial dimensions: temporal feature separation decomposes non-stationary pressure signals to extract long-term trend terms reflecting the stabilization process of mud infiltration and residual terms capturing microscopic instability phenomena; spatial dimensionality reduction, on the other hand, maps high-dimensional features to low-dimensional space by preserving the local manifold structure for multi-dimensional pressure data from sensors at different depths at the same time, thus retaining key nonlinear characteristics of formation pressure transmission. Both steps characterize film formation properties from the perspectives of dynamic evolution and spatial distribution, respectively.

[0050] Step S500: Based on the sequence characteristics and spatial dimensionality reduction characteristics, a comprehensive evaluation of the film formation quality is conducted to obtain a comprehensive evaluation result of the mud film integrity and formation sealing effect.

[0051] Finally, step S500 constructs a multi-parameter collaborative evaluation model by integrating temporal and spatial dimensionality reduction features: trend features reflect the stability of mud film formation, residual features reveal the risk of local failure, and spatial features characterize the differences in formation permeability. Based on the interaction of these features, the sealing efficiency and long-term stability of the mud film are comprehensively judged, and a comprehensive evaluation result with both quantitative indicators and qualitative levels is finally output, providing a basis for engineering parameter optimization. Specifically, when conducting a comprehensive evaluation of film formation quality based on temporal and spatial dimensionality reduction features, the pressure stabilization rate, pressure mutation characteristics, and formation permeability characteristics are first integrated to form an input feature set; subsequently, the model is evaluated using the formula... Calculate the equivalent permeability coefficient of the formation, where, The equivalent permeability coefficient of the formation. This is the correction factor for the main grain size of the formation, selected based on the grain size distribution characteristics and main grain size information of the formation. This is a mud viscosity correction factor, designed to adjust the treatment method based on the viscosity and flow characteristics of different muds and specific experimental results. The spacing between the components of the pore water pressure testing assembly. The filtration loss is monitored by the weighing mechanism. for The head difference at corresponding intervals can be calculated by measuring the pore water pressure using a pore water pressure detection component. Let be the cross-sectional area of ​​the permeation column. On the right-hand side of the equation, The equivalent permeability coefficient is a correction value for the conversion between the actual and equivalent permeability coefficients, which can be obtained through permeability testing. This represents the actual permeability coefficient of the formation. Furthermore, Taylor's empirical formula is used... (In the formula, These are empirical parameters, determined based on actual conditions during experiments or engineering. (Reverse derivation of formation porosity) This allows for a quantitative evaluation of the formation porosity, and finally, by combining a gradient lifting regression model and a logistic regression classification model, a comprehensive evaluation result of mud film integrity and formation sealing effect is output.

[0052] Further, step S200 includes steps S210 to S230.

[0053] Step S210: Based on the pore water pressure time series data, instantaneous pulse noise is identified. By detecting abrupt changes in the pressure signal with a duration less than a preset value and an amplitude exceeding the range of physical laws, a set of abnormal data points caused by electromagnetic interference and particle impact is obtained.

[0054] Step S220: Calculate the dynamic distribution parameters based on the set of abnormal data points and the original data. After excluding the abnormal points, recalculate the mean and standard deviation of the dataset to obtain the dynamic distribution parameters adapted to the high-pressure test environment of the slurry shield tunnel.

[0055] Step S230: Perform three sigma criterion filtering based on dynamic distribution parameters and raw data. By removing outliers caused by equipment failure or environmental interference, the cleaned pore water pressure dataset is obtained.

[0056] Specifically, in the high-pressure test environment of slurry shield tunneling, pressure sensors are susceptible to electromagnetic interference or instantaneous pulse noise caused by mud particle impact. Step S210 first identifies outliers in the pore water pressure time series data. By detecting abrupt changes in the signal with extremely short durations and amplitudes exceeding the range of physical laws, the abnormal data set caused by external interference is accurately captured. Step S220, based on this, uses the identified outlier set to calculate the dynamic distribution parameters of the original data. By eliminating these outliers, the mean and standard deviation of the dataset are recalculated to obtain statistical parameters that better reflect the actual formation response characteristics, thus adapting to the dynamic changes in the high-pressure test environment. Step S230 finally applies the three-sigma criterion for filtering. Based on the dynamically calculated distribution parameters, the normal data fluctuation range is defined, and outliers caused by equipment failure or environmental interference are systematically eliminated. This ensures that the cleaned pore water pressure dataset truly reflects the physical laws of the mud infiltration process, providing a reliable basis for subsequent analysis. This progressive processing effectively overcomes the interference of data noise in high-pressure tests and improves the accuracy of film formation quality assessment.

[0057] Further, step S300 includes steps S310 to S330.

[0058] Step S310: Extract statistical features of depth stratification based on pore water pressure dataset. Calculate the mean and standard deviation of sensor data at different depth levels to obtain statistical parameters that reflect the inherent distribution characteristics of each depth level.

[0059] Step S320: Calculate the relative positional relationship based on statistical parameters and pore water pressure dataset. Obtain the standard score sequence that eliminates the influence of depth by subtracting the mean of the corresponding depth level from each data point and dividing by the standard deviation of that level.

[0060] Step S330: Perform cross-level data integration based on the standard score sequence. By mapping the standard scores of all depth levels to a unified numerical range, standardized pore water pressure data with uniform scale is obtained.

[0061] Specifically, step S310 first addresses the natural gradient differences in pore water pressure at different depths within the formation by extracting the mean and standard deviation of sensor data at each depth level through depth-stratified statistical feature extraction. This accurately captures the inherent distribution patterns of formation permeability characteristics at different depths. Based on this, step S320 calculates the relative positional relationships of the raw data using stratified statistical parameters. By subtracting the mean of the corresponding depth level from each data point and dividing by the standard deviation of that level, the raw pressure readings are converted into dimensionless standard scores. This process effectively eliminates the natural increase in water pressure caused by increasing depth, making data from different depths comparable. Finally, step S330 integrates cross-layer data, mapping the standard scores of each depth level to a unified numerical range to form a standardized dataset with consistent scale. This standardization process preserves the relative differences between sensor readings at different depths and provides a dimensionally unified input for subsequent machine learning models, avoiding model weight bias caused by differences in data magnitude.

[0062] Further, step S400 includes steps S410 to S430.

[0063] Step S410: Perform time series decomposition based on standardized pore water pressure data. By decomposing the non-stationary pressure signal into long-term trend terms, seasonal periodic terms, and residual terms, the time series characteristics are obtained.

[0064] Step S420: Based on the standardized pore water pressure data, a multidimensional dataset is constructed and processed. By integrating the pressure readings of sensors at different depths at the same time, a spatial dataset reflecting the formation pressure distribution characteristics is obtained.

[0065] Step S430: Perform nonlinear manifold learning processing based on the spatial dataset. By maintaining the pressure correlation between measurement points at different depths, obtain low-dimensional spatial features that characterize the permeability of the formation.

[0066] Specifically, in step S410, considering the non-stationary characteristics of the pressure signal during the mud-water shield membrane formation process, the STL time-series decomposition algorithm is preferably used to separate the standardized pore water pressure data into long-term trend terms, seasonal periodic terms, and residual terms. The core advantage of this algorithm is its ability to effectively decompose complex non-stationary pressure signals into these terms. During the mud film formation process, the trend term smoothly reflects the process of mud infiltration and gradual stabilization in the formation, and is used to extract the pressure stabilization rate to characterize the film formation efficiency. The residual term, on the other hand, can remove the influence of periodic vibrations from the air compressor, accurately capture non-random pressure mutations, and thus identify micro-instability phenomena such as mud film rupture or formation fracturing. Building upon this foundation, step S420 constructs and processes a multidimensional dataset, simultaneously integrating readings from multiple sensors arranged along the formation depth at the same time to form a high-dimensional dataset reflecting the spatial distribution characteristics of pressure. This process overcomes the limitations of single-point analysis, providing a data foundation for revealing the pressure transmission patterns within the formation. Step S430 further employs nonlinear manifold learning to process the spatial dataset. By preserving the local neighborhood structure of pressure correlations between measuring points at different depths, it maps high-dimensional spatial features to low-dimensional representations, such as using t-SNE feature reduction. Compared to traditional linear dimensionality reduction methods, t-SNE is more suitable for handling the nonlinear characteristics of formation pressure transmission in this experiment. It can significantly reduce computational complexity while maximally preserving the local manifold structure in the high-dimensional data, i.e., maintaining the pressure correlations between measuring points at different depths, thereby retaining key information reflecting formation permeability characteristics in low-dimensional space.

[0067] Further, step S500 includes steps S510 to S530.

[0068] Step S510: Perform feature fusion processing based on time-series characteristics. By fusing pressure stability rate characteristics, pressure abrupt change characteristics, and formation permeability characteristics, an input feature set is obtained.

[0069] Step S520: Continuous index prediction is performed based on the input feature set. The integrity of the mud film and the formation porosity are nonlinearly fitted and predicted based on the preset gradient boosting regression model to obtain quantitative evaluation results.

[0070] Step S530: Determine the quality level based on the quantitative assessment results, analyze the feature weights based on the preset logistic regression classification model and map them into a probability distribution to obtain the comprehensive assessment result of the mud film quality.

[0071] Preferably, step S510 first integrates the pressure stability rate characteristics and pressure mutation characteristics obtained from time-series analysis with the formation permeability characteristics extracted by spatial dimensionality reduction to construct an input feature set that can comprehensively reflect the dynamic characteristics of the film formation process. This fusion process effectively integrates the evolution law in the time dimension and the distribution characteristics in the spatial dimension. Step S520 uses a gradient boosting regression model to predict continuous indicators based on this feature set. By iteratively training multiple weak learners to approximate the residuals, nonlinear fitting of parameters that are difficult to directly observe, such as mud film integrity and formation pore filling rate, is achieved. This model is particularly suitable for handling the complex nonlinear relationship of mud filling in porous media. Step S530 finally uses a logistic regression classification model to determine the quality level of the quantitative prediction results. The linear combination of feature inputs is mapped to probability values ​​through the Sigmoid function, and the film formation effect is divided into different levels according to the probability distribution. This model can not only output the probability that the mud film belongs to a specific quality level, but also intuitively reflect the degree of influence of each feature on the rating through the weight coefficient, and finally form a comprehensive evaluation system that combines quantitative indicators and qualitative levels. Specifically, the comprehensive evaluation system based on a logistic regression classification model established in this invention classifies the film-forming effect into different levels according to probability distribution. For example, Level I represents extremely rapid pressure stabilization and no increase in deep pore water pressure, corresponding to excellent film-forming efficiency and formation sealing effect; Level II represents moderate pressure stabilization speed, which can form effective support; Level III represents continuous pressure fluctuation or sudden changes, indicating high risk in the project. The comprehensive evaluation results directly guide engineers to adjust the mud specific gravity or viscosity parameters to meet the strict requirements for the safety of pressurized opening in actual engineering projects.

[0072] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. An intelligent evaluation method, using a detachable slurry shield tunneling mud film formation and curing test device, characterized in that, The device includes: A one-dimensional permeation column assembly (1) includes an upper top plate (11), an upper permeation column (12), a detachable permeation column (13), a lower permeation column (14), a lower bottom plate (15), and a pressure gauge (16). The upper top plate (11) and the lower bottom plate (15) are connected by multiple connecting rods (17). The upper permeation column (12), the detachable permeation column (13), and the lower permeation column (14) are arranged coaxially along the center line and sealed between the upper top plate (11) and the lower bottom plate (15). The pressure gauge (16) is arranged on the upper top plate (11). A pressurizing assembly (2) is connected to a pressurizing through hole in the upper top plate (11); A pore water pressure monitoring component (3) is disposed within the one-dimensional permeation column component (1); The filtrate collection assembly (4) includes a filtrate outlet pipe (41), a filtrate collection bucket (42), and a weighing mechanism (43), and is located below the one-dimensional permeation column assembly (1). The filtrate outlet pipe (41) is connected to the through hole of the lower base plate (15), the filtrate collection bucket (42) is located below the filtrate outlet pipe (41), and the weighing mechanism (43) is located below the filtrate collection bucket (42). Pressure holding observation assembly (5), wherein a detachable permeation column placement ring (54) is provided inside the test chamber (51) of the pressure holding observation assembly (5); Information processing component (6), which is electrically connected to the pore water pressure monitoring component (3) and the weighing mechanism (43) respectively; The method includes: Obtain raw data, including pore water pressure time series data, test pressure data, mud ratio parameters, and filtration loss data during the film formation test; Abnormal data filtering is performed on the original data. Abnormal data points that deviate from the main distribution are identified and removed based on the normal distribution characteristics to obtain the cleaned pore water pressure dataset. Based on the pore water pressure dataset, data scale is unified. By calculating the relative positional relationship between each data point and the overall dataset, the magnitude difference of sensor readings at different depths is eliminated, and standardized pore water pressure data is obtained. Based on the standardized pore water pressure data, temporal feature separation and dimensionality reduction are performed to obtain temporal features and spatial dimensionality reduction features. Based on the temporal characteristics and spatial dimensionality reduction characteristics, a comprehensive evaluation of film formation quality is conducted to obtain a comprehensive evaluation result of mud film integrity and formation sealing effect; Specifically, based on the standardized pore water pressure data, temporal feature separation and dimensionality reduction processing are performed to obtain temporal features and spatial dimensionality-reduced features, including: Time series decomposition is performed on the standardized pore water pressure data. By decomposing the non-stationary pressure signal into long-term trend terms, seasonal periodic terms, and residual terms, time series characteristics are obtained. Based on the standardized pore water pressure data, a multidimensional dataset is constructed and processed. By integrating the pressure readings of sensors at different depths at the same time, a spatial dataset reflecting the formation pressure distribution characteristics is obtained. Based on the spatial dataset, nonlinear manifold learning is performed to obtain low-dimensional spatial features characterizing the formation permeability by maintaining the pressure correlation between measurement points at different depths. The film formation quality is comprehensively evaluated based on the sequence characteristics and spatial dimensionality reduction characteristics to obtain a comprehensive evaluation result of mud film integrity and formation sealing effect, including: Based on the aforementioned time-series characteristics, feature fusion processing is performed to obtain the input feature set by fusing pressure stability rate characteristics, pressure abrupt change characteristics, and formation permeability characteristics. Based on the input feature set, continuous index prediction is performed, and nonlinear fitting prediction of mud film integrity and formation porosity is performed based on the preset gradient enhancement regression model to obtain quantitative evaluation results. The quality level is determined based on the quantitative assessment results. The feature weights are analyzed and mapped to a probability distribution based on a preset logistic regression classification model to obtain a comprehensive assessment result of the mud film quality.

2. The intelligent evaluation method according to claim 1, characterized in that: The upper permeation column (12) and the detachable permeation column (13) are connected at the upper groove, and the lower permeation column (14) and the detachable permeation column (13) are connected at the lower groove. Both the upper groove and the lower groove are provided with sealing rings. The two ends of the connecting rod (17) are connected to the upper top plate (11) and the lower bottom plate (15) respectively and are fixed by nuts.

3. The intelligent evaluation method according to claim 1, characterized in that: The pressurization assembly (2) includes an air compressor (21), a pressurization pipe (22), and a voltage regulator (23). The air compressor (21) is connected to one end of the pressurization pipe (22), and the other end of the pressurization pipe (22) is connected to a through hole in the upper top plate (11). The voltage regulator (23) is mounted on the pressurization pipe (22).

4. The intelligent evaluation method according to claim 1, characterized in that: The pressure holding observation assembly (5) also includes an environmental control element (52) and an upper cover plate (53), which is connected to the test chamber (51) by a locking nut.

5. The intelligent evaluation method according to claim 1, characterized in that: The detachable permeation column (13) is provided with a screen (131) at one end near the lower permeation column (14).

6. The intelligent evaluation method according to claim 1, characterized in that, Based on the original data, abnormal data filtering is performed. Abnormal data points deviating from the main distribution are identified and removed based on the normal distribution characteristics, resulting in a cleaned pore water pressure dataset, including: Based on the pore water pressure time series data, instantaneous pulse noise is identified. By detecting abrupt changes in the pressure signal with a duration less than a preset value and an amplitude exceeding the range of physical laws, a set of abnormal data points caused by electromagnetic interference and particle impact is obtained. Dynamic distribution parameters are calculated based on the set of abnormal data points and the original data. By excluding abnormal points, the mean and standard deviation of the dataset are recalculated to obtain dynamic distribution parameters adapted to the high-pressure test environment of slurry shield tunneling. Based on the dynamic distribution parameters and the original data, the three sigma criterion is used for filtering. By removing outliers caused by equipment failure or environmental interference, the cleaned pore water pressure dataset is obtained.

7. The intelligent evaluation method according to claim 1, characterized in that, Based on the pore water pressure dataset, data scaling is performed. By calculating the relative positional relationship of each data point to the overall dataset, the magnitude differences in sensor readings at different depths are eliminated, resulting in standardized pore water pressure data, including: Based on the pore water pressure dataset, depth-stratified statistical features are extracted. By calculating the mean and standard deviation of sensor data at different depth levels, statistical parameters reflecting the inherent distribution characteristics of each depth level are obtained. The relative positional relationship is calculated based on the statistical parameters and pore water pressure dataset. By subtracting the mean of the corresponding depth level from each data point and dividing by the standard deviation of that level, a standard score sequence that eliminates the influence of depth is obtained. Cross-level data integration is performed based on the standard score sequence. By mapping the standard scores of all depth levels to a unified numerical range, standardized pore water pressure data with uniform scale is obtained.

Images