Rock mass chemical-dry-wet cycle coupling degradation test device and method and response model

Abstract

The invention discloses a rock mass chemical-dry-wet cycle coupling degradation test device and method and a response model.The test device comprises a cabin body, a cabin plate used for blocking the cabin body is detachably arranged at the bottom of the cabin body, and a sample to be tested is placed on the cabin plate; a spray head for spraying a chemical solution to a sample and a heating device for drying the sample are arranged at the top in the cabin body, a fan for air-drying the sample is arranged on the side wall in the cabin body, and an observation window for conveniently observing the condition of the sample is also formed in the side wall of the cabin body; the device has the beneficial effects that a high-simulation coupling environment is provided, precise atomization and programmed drying are combined, high-precision coupling simulation of two degradation factors of chemical corrosion and physical dry and wet in time, space and strength is realized, and a field environment is reflected more truly.

Description

Technical Field

[0001] This invention relates to the field of geotechnical engineering testing technology, and in particular to a rock mass chemical-dry-wet cycle coupled degradation test device, method and response model, which is especially suitable for the study of the durability, degradation mechanism and engineering life prediction of soft rocks such as red mudstone. Background Technology

[0002] Red mudstone is widely distributed in central and western my country and is a common foundation and slope medium for infrastructure construction. This rock mass exhibits significant characteristics of "softening upon contact with water and cracking upon loss of water." Under the influence of groundwater or rainwater containing salts (such as sulfates and chlorides), its deterioration process is rapidly accelerated due to the coupled effect of chemical corrosion and physical wetting / drying, leading to frequent engineering disasters. To study this complex process, existing technologies mainly employ the following two experimental methods: (1) Immersion-drying cycle method: The rock sample is immersed in a salt solution for a period of time, then taken out and placed in an oven to dry, and this cycle is repeated. This method has significant drawbacks: the immersion process does not match the natural precipitation infiltration or capillary water absorption process, and cannot simulate the dynamic front of water migration; the drying process is slow and uneven, and cannot accurately simulate the field conditions of sunlight or air drying; the entire cycle is long and inefficient, and it is difficult to achieve in-situ and continuous observation during the experiment.

[0003] (2) Simplified spray-drying method: This method uses a humidifier to simulate wetting, combined with a heating lamp or fan for drying. Although this method is closer to alternating wet and dry conditions, it generally suffers from problems such as difficulty in precisely controlling the droplet size and chemical concentration, uneven drying rate and temperature field, and lack of monitoring and feedback of environmental parameters (temperature, humidity, and chemical state of the solution). More importantly, existing devices generally lack a seamless interface with high-precision in-situ observation equipment (such as micro-CT and 3D laser scanners), making it impossible to obtain continuous, multi-scale evolution data of internal cracks from initiation to penetration without damaging the sample or interrupting the experiment.

[0004] The patent application number "CN201920518643.3" discloses "a multi-functional control device for revealing the mechanism of mudstone engineering strength degradation". Although it realizes the engineering simulation of mudstone, the simulation process cannot predict and evaluate the actual service life of mudstone. Summary of the Invention

[0005] The purpose of this invention is to provide a chemical-wet-dry cycle coupled degradation test device, method and response model for rock mass. It not only realizes high-precision coupled simulation of the two major degradation factors of chemical corrosion and physical wet-dry cycles in time, space and intensity, and more realistically reflects the field environment, but also establishes a response model, which upgrades this device from a simulation device to a prediction and evaluation platform.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: A rock mass chemical-wet-dry cycle coupled deterioration test device includes a chamber, the bottom of which is detachably equipped with a chamber plate for sealing the chamber. The chamber plate includes a plate body, a guide plate is provided on the plate body, and a liquid outlet is provided at the lowest point of the guide plate. The test sample is placed on the plate body. The top of the chamber body is equipped with a nozzle for spraying chemical solution onto the sample and a heating device for drying the sample. The side wall of the chamber body is equipped with a fan for air drying the sample. The side wall of the chamber body is also provided with an observation window for easy observation of the sample condition. The chamber is also equipped with a first sensor group for detecting the chamber temperature, chamber humidity, sample surface temperature, sample surface humidity, and sample mass, as well as a second sensor group for detecting the pH value, conductivity, and concentration of the chemical solution.

[0007] Preferably, this application also provides a rock mass chemical-wet-dry cycle coupled deterioration test method, based on the aforementioned rock mass chemical-wet-dry cycle coupled deterioration test device, including the following steps: S1: Sample preparation and pretreatment: The rock mass is processed into standard cylindrical or cubic samples, and its initial physical and mechanical parameters are measured. Then the samples are installed on the plate. S2: Experimental parameter setting: Different experimental parameters are set according to different research objectives; S3: Perform automated wet-dry cycle test: According to the test parameters set in step S2, automatically perform spray wetting, constant humidity standing, and drying operations. After completing one cycle, automatically enter the next cycle until the set number of cycles is reached. During this process, record the monitoring data of the first sensor group and the second sensor group. S4: Intermittent in-situ monitoring: Pause the test procedure at the set critical cycle points to acquire images of the sample; S5: Data post-processing and analysis: Combine the sample images collected in step S4 with the test parameters in step S2 to establish a response model.

[0008] Preferably, the test parameters in step S2 include, Chemical parameters: different chemical solutions and their concentrations, pH values, and spray duration and sequence of different chemical solutions; Dry-wet cycle parameters: duration of spray wetting phase, constant humidity settling phase, and drying phase in a complete cycle, sample surface temperature, sample surface humidity, and heating / drying rate; Total number of loops: N, the total number of complete loops required to complete the experiment.

[0009] Preferably, the key loop node in step S4 refers to the 10th / 20th / 50th complete loop.

[0010] Preferably, the image acquisition of the sample in step S4 includes: The surface image of the sample is acquired through the observation window to obtain the crack rate, crack width, and crack fractal dimension of the sample. The chamber plate and the sample on it were removed from the chamber and moved to a micro-CT device for internal scanning to obtain three-dimensional crack network data. The three-dimensional crack network data included the three-dimensional volume, surface area, connectivity and pore size distribution of the cracks inside the sample. After the scanning was completed, the chamber plate and the sample were reinstalled on the chamber.

[0011] Preferably, this application also provides a rock mass chemical-dry-wet cycle coupled deterioration response model, based on the aforementioned rock mass chemical-dry-wet cycle coupled deterioration test method, wherein the response model has an input X and an output Y, wherein the input X is the test parameter in step S2; The output Y is the sample response parameter, including the sample mass change and the sample image acquisition result in step S4.

[0012] Preferably, the method for constructing the response model includes the following steps: S6: Establish a quantitative functional relationship between rock mass performance parameters and the number of wet-dry cycles and the concentration of erosion solution; S7: Establish a multivariate machine learning prediction model, use the quantitative function in step S6 as features, construct a dataset, and use a supervised learning algorithm to train the prediction model so that the prediction model can achieve two functions: degradation prediction and state inversion. S8: Construct a multi-scale coupled constitutive model framework. Based on real crack network data obtained from CT scans, and combined with chemical-mechanical theory, establish a discrete element numerical model to link the evolution of microcracks with macroscopic mechanical properties through the equivalent continuous medium theory.

[0013] Preferably, the process of training the prediction model in step S7 is as follows: S71: After the device completes a set of comparative experiments covering different chemical concentrations and wet / dry regimes, it obtains an "input-response" database. S72: Use the database from step S71 to train the prediction model so that the prediction model learns to predict long-term key indicators that are difficult to obtain in real time or require destructive testing from readily available early or real-time data.

[0014] Preferably, the degradation prediction of the prediction model in step S7 is manifested as: after inputting the target environmental conditions, outputting the complete decay curve and final state of the rock mass mechanical properties; In step S7, the state inversion of the prediction model is manifested as follows: after inputting the early deterioration data of the rock mass monitored on site, the environmental erosion intensity it has experienced is estimated in reverse, so as to realize intelligent diagnosis of the engineering state.

[0015] Preferably, the rock mass chemical-wet-dry cycle coupled degradation response model has the following output: Output 1 (Material Property Envelope Chart): Generates an engineering chart with "solution concentration" and "number of wet-dry cycles" as coordinate axes and "strength retention rate" or "crack development level" as contour lines, providing an intuitive basis for parameter selection in engineering design for different corrosive environments; Output 2 (Engineering Life Prediction and Early Warning Threshold): Combining local meteorological data (rainfall-evaporation data) and groundwater chemistry data, the model can convert the number of accelerated cycles in the laboratory into natural lifespan in the field. Based on this, the service life of engineering structures (such as slopes and foundations) is predicted, and quantitative thresholds for monitoring and early warning are set (for example, when the ultrasonic wave velocity drops to 80% of the initial value, a level one early warning is triggered). Output 3 (Evaluation of the effectiveness of reinforcement measures): The model is applied to red mudstone samples treated with different modifiers. By comparing the response differences before and after modification under the same chemical-dry-wet attack, the protective efficacy of various reinforcement measures is quantitatively evaluated, providing data support for optimizing treatment plans.

[0016] The beneficial effects of this invention are as follows: 1. Highly simulated coupled environment: By combining precision atomization with programmed drying, a high-precision coupled simulation of the two major deterioration factors, chemical corrosion and physical drying and wetting, in terms of time, space and intensity is achieved, which more realistically reflects the field environment.

[0017] 2. Controllability and acceleration of the test process: The drying rate and cycle period can be automatically controlled by the program, which greatly shortens the time of a single cycle, realizes the accelerated test of the rock sample deterioration process, and improves the research efficiency.

[0018] 3. Supports multi-scale in-situ non-destructive observation: The unique in-situ monitoring interface design solves the problem of destructive sampling or inability to continuously observe the internal structure in traditional methods, and realizes the acquisition of "same sample, continuous observation, multi-scale data", which greatly improves the depth of understanding of the mechanism of crack initiation, propagation and penetration.

[0019] 4. Intelligentization and standardization: The automated control and data recording throughout the process ensure the consistency of experimental conditions and the reliability of data, which is conducive to the comparison and verification of different research results and lays the foundation for establishing standardized testing methods. Attached Figure Description

[0020] Figure 1This is a schematic diagram of the overall structure of the experimental device in this application; Figure 2 This is a schematic diagram of the overall structure of the workbench in this application; Figure 3 This is a schematic diagram of the assembly of the hull and panels in this application; Figure 4 This is a structural schematic diagram of the cabin in this application; Figure 5 This is a structural schematic diagram of the cabin from another angle in this application; Figure 6 This is a structural schematic diagram of the cabin from the third angle in this application; Figure 7 This is a structural schematic diagram of the cabin plate in this application; Figure 8 This is a structural schematic diagram of the cabin plate from another angle in this application; Figure 9 This is a structural schematic diagram of the third angle of the cabin plate in this application.

[0021] The accompanying drawings are for illustrative purposes only and should not be construed as limiting the scope of this patent. To better illustrate this embodiment, some components in the drawings may be omitted, enlarged, or reduced, and do not represent the actual dimensions of the product. It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings. Detailed Implementation

[0022] The present invention will now be further described with reference to the accompanying drawings. Example

[0023] like Figure 3 As shown, the rock mass chemical-wet-dry cycle coupled degradation test device of this embodiment includes a chamber 2, and a chamber plate 3 for sealing the chamber 2 is detachably provided at the bottom of the chamber 2. Figure 4 — Figure 6 As shown, the bottom of the cabin 2 has a connecting plate 22, and the connecting plate 22 has a first connecting hole 23, as... Figure 7 — Figure 9 As shown, the compartment 3 has a plate 31, on which a third connecting hole 32 is provided. By using bolts to pass through the first connecting hole 23 and the third connecting hole 32 in sequence, the compartment 2 and the compartment 3 can be connected or separated. A sealing strip can also be provided between the connecting plate 22 and the plate 31 to prevent foreign objects from entering and the spray solution from overflowing.

[0024] like Figure 7 — Figure 9As shown, a tray 36 is provided on the plate 31, and a test specimen 37 is placed in the tray 36. The specimen 37 can be cylindrical or cubic in shape. Depending on different testing requirements, multiple specimens 37 can be placed in the tray 36 at the same time, or only one specimen 37 can be placed.

[0025] like Figure 4 — Figure 6 As shown, the top of the chamber 2 is equipped with a nozzle 27 for spraying chemical solution onto the sample 37 and a heating device 26 for drying the sample.

[0026] In this embodiment, nozzle 27 is an atomizing nozzle, and multiple atomizing nozzles form a nozzle array, which can produce micron-sized uniform droplets to simulate natural precipitation or capillary condensation. A solution storage tank is arranged on the outside of the cabin 2. The solution storage tank is connected to the nozzle array through a liquid delivery pipeline. A precision metering pump, a metering pump, and a multi-way valve are also installed on the liquid delivery pipeline. Multiple solution storage tanks can be installed to store chemical solutions of different components, concentrations, or pH values. The aforementioned precision metering pump, metering pump, and multi-way valve on the liquid delivery pipeline are controlled by the control system, which can realize the automatic switching of different types of chemical solutions and the start, stop, and duration control of different types of chemical solution sprays.

[0027] A fan 21 for air-drying the sample 37 is installed on the side wall inside the chamber 2. In this embodiment, the heating device 26 is an infrared heating array, which can provide uniform and rapid radiant heating to simulate sunlight; the fan 21 is used to promote airflow inside the chamber to simulate the air-drying effect. Under the command and control of the control system, the heating device 26 and the fan 21 can work independently or in combination to achieve precise control of the drying rate, final temperature, and humidity decrease curve.

[0028] In this embodiment, the chamber 2 is a sealed, transparent, high-strength acrylic cavity used to house the sample 37 and create an independent and controllable testing environment. An observation window 25 is also provided on the side wall of the chamber 2 for convenient observation of the sample. The observation window 25 has high light transmittance, allowing a high-speed camera to perform point-to-point or panoramic time-series imaging of the surface of the sample 37 from the outside, observing the propagation of surface cracks.

[0029] A guide plate 33 is provided on the plate 31. The height of the guide plate 33 gradually decreases from one end to the other. A guide groove 34 is provided at the lowest point of the guide plate 33. An outlet hole 35 is provided at the lowest point of the guide groove 34. The outlet hole 35 is connected to an outlet valve 38. After the chemical solution sprayed by the nozzle 27 falls on the plate 31, it will flow along the guide plate 33 and the guide groove 34 to the outlet hole 35. After opening the outlet valve 38, it can be smoothly discharged to prevent liquid accumulation.

[0030] Inside chamber 2, a first sensor group is installed to detect chamber temperature, chamber humidity, sample surface temperature, sample surface humidity, and sample mass. On plate 31, a second sensor group is installed to detect the pH value, conductivity, and concentration of the chemical solution. The chemical solution detected by the second sensor group is the solution that is sprayed through nozzle 27, reacts with sample 37, and remains on plate 31 (the chemical solution cannot flow out when outlet valve 38 is not opened).

[0031] The control system of this embodiment includes: a first sensor group, a second sensor group, a central controller, and a human-machine interface. The central controller receives data from the sensor groups and automatically controls the start-up, shutdown, and operating intensity of the nozzle 27, fan 21, and heating device 26 according to a preset test program. The human-machine interface is used to input and edit complex test procedures, such as the time for each stage of spray wetting, constant humidity settling, and drying, the humidity of the sample 37, and the temperature of the sample 37, and displays and records all environmental parameters and equipment status in real time. This control system is a conventional technology and will not be described in detail in this embodiment.

[0032] like Figure 1 and Figure 2 As shown, to facilitate the removal of the compartment plate 3 and the sample 37 inside after removing it from the compartment 2, a workbench 1 is provided. The workbench 1 has a clearance hole 11, the size of which is slightly larger than the size of the plate 31. A fourth connecting hole 12 is also provided on the workbench 1, and a second connecting hole 24 is provided on the connecting plate 22. Bolts are passed through the second connecting hole 24 and the fourth connecting hole 12 in sequence to reliably fix the compartment 2 onto the workbench 1. In this embodiment, the first connecting hole 23 is located inside the connecting plate 22, and the second connecting hole 24 is located outside the connecting plate 22. When removing the compartment plate 3, the liquid outlet valve 38 is first opened to drain the chemical solution inside. Then, the connecting bolts of the first connecting hole 23 and the third connecting hole 32 are removed, allowing the compartment plate 3 and the sample inside to be passed through the clearance hole 11 and removed.

[0033] This embodiment also provides a rock mass chemical-wet-dry cycle coupled deterioration test method, based on the aforementioned rock mass chemical-wet-dry cycle coupled deterioration test device, including the following steps: S1: Sample preparation and pretreatment: The rock mass is processed into standard cylindrical or cubic samples, and its initial physical and mechanical parameters (such as mass, size, wave velocity, etc.) are measured. Then the sample 37 is installed on the tray 36 of the plate 31. S2: Experimental Parameter Setting: Different experimental parameters are set according to different research objectives. The experimental parameters here include: Chemical parameters: different chemical solutions and their concentrations, pH values, and spray duration and sequence of different chemical solutions; Wet-dry cycle parameters: the duration of the spray wetting phase, the constant humidity settling phase, and the drying phase in a complete cycle, as well as the sample surface temperature, sample surface humidity, and heating / drying rate; Total number of loops: N, the total number of complete loops required to complete the experiment.

[0034] S3: Perform automated wet-dry cycle test: According to the test parameters set in step S2, automatically perform spray wetting, constant humidity standing, and drying operations. After completing one cycle, automatically enter the next cycle until the set number of cycles is reached. During this process, record the monitoring data of the first sensor group and the second sensor group. S4: Intermittent in-situ monitoring: Pause the test procedure at a set critical cycle node. In this embodiment, the critical cycle node is defined as the 10th / 20th / 50th complete cycle. Those skilled in the art should understand that, depending on different engineering requirements, the cycle node can also be any other number of complete cycles.

[0035] After pausing the test procedure, image acquisition of the specimen is performed. The image acquisition of the specimen at this location includes: The surface image of the sample is acquired through the observation window 25 to obtain the crack rate, crack width and crack fractal dimension of the sample. If necessary, the chamber plate 3 and the sample 37 on it can be removed from the chamber 2, and the sample 37 and the tray 36 can be moved together to a micro-CT device via a non-destructive testing docking platform for internal scanning to obtain three-dimensional crack network data. The three-dimensional crack network data includes the three-dimensional volume, surface area, connectivity, and pore size distribution of the cracks inside the sample. After scanning, the chamber plate 3 and the sample 37 are reinstalled on the chamber, and the subsequent cycle continues. The non-destructive testing docking platform mentioned here is existing conventional equipment and is not the focus of this embodiment, so it will not be described in detail here.

[0036] S5: Data post-processing and analysis: Combining the sample images collected in step S4 and the experimental parameters in step S2, quantitatively characterize the crack evolution law of sample 37 at multiple scales (micro-meso-macro) (such as crack density, width, fractal dimension, pore volume change, etc.) and establish a response model.

[0037] This application also provides a method for constructing a rock mass chemistry-wet-dry cycle coupled deterioration response model. Based on the aforementioned rock mass chemistry-wet-dry cycle coupled deterioration test method, the response model has input X and output Y. Input X consists of the test parameters in step S2, mainly including solution type, ion species, concentration (mol / L), pH value, conductivity, spray duration in a single cycle, drying temperature curve, humidity change curve, and number of cycles. Output Y consists of the sample response parameters, including the mass change of sample 37 and the image acquisition results of the sample in step S4. This mainly refers to the surface morphology changes of sample 37 (crack rate, crack width, and crack fractal dimension extracted and analyzed through observation window 25) and the three-dimensional microstructure changes (internal crack three-dimensional volume, surface area, connectivity, and pore size distribution extracted by micro-CT scanning).

[0038] The method for constructing a response model includes the following steps: S6: Establish a quantitative functional relationship between rock mass performance parameters (P), wet-dry cycle number (N), and erosion solution concentration (C); For example, in this embodiment, the chemical solution can be set to Na2SO4, and the key indicators are unconfined compressive strength and elastic modulus, which can be expressed as negative exponential or power function: P(N) = P0 * exp(-k*N) or P(N) = P0 * N^(-b).

[0039] The attenuation coefficients k or b are further correlated with the solution concentration C, such as k = α*C + β or b = γ*C + δ. α and γ are concentration sensitivity coefficients, physically representing the increase in the attenuation coefficient for each unit increase in concentration; a value greater than zero indicates that the higher the concentration, the faster the attenuation. β and δ are baseline attenuation coefficients, representing the attenuation coefficients under theoretically "pure water" (i.e., C=0) or extremely low concentration conditions, reflecting the damage rate caused by the physical effects of the wet-dry cycle itself.

[0040] The final quantitative function is: P = f(P0, N, C). This quantitative function model is simple and intuitive, and is suitable for rapid evaluation under specific working conditions.

[0041] This quantitative function is not a specific attenuation equation, but rather a general functional expression. It indicates the mathematical relationship between the rock mass properties P and the initial properties P0, the number of wet-dry cycles N, and the solution concentration C.

[0042] By employing the two negative exponential or power function models supplemented above, a formula with a concrete mathematical form for this quantitative function is provided. The value of this application lies in the fact that it is the first time that the existence of this functional relationship has been experimentally established, and a quantifiable concrete model has been given.

[0043] S7: Establish a multivariate machine learning prediction model. Use the quantitative function from step S6 as features to construct a dataset, and train the prediction model using a supervised learning algorithm (such as random forest, gradient boosting machine, or neural network). The training process is as follows: S71: After the device completes a set of comparative experiments covering different chemical concentrations and wet / dry regimes, it obtains an "input-response" database. S72: Then, the prediction model is trained using the database in step S71, so that the prediction model learns to predict long-term key indicators that are difficult to obtain in real time or require destructive testing (such as shear strength parameters after the Nth cycle and macroscopic crack network morphology) from readily available early or real-time data (such as the mass change rate, wave velocity decrease rate, and surface crack propagation rate in the first M cycles).

[0044] The goal of the training process is to teach the model how to predict long-term results that are difficult to measure directly (such as strength or macroscopic crack state after N cycles) based on early signals (such as data from the first M cycles). Specifically, suppose we need to predict the uniaxial compressive strength (UCS) retention rate of a rock sample after 50 wet-dry cycles (N=50): First, a structured database needs to be built, and a large number of systematic experiments need to be conducted while data parameters are collected synchronously. Secondly, the dataset is constructed by transforming the raw data into a machine learning-readable format. Finally, model training and validation are performed, divided into a 70% training set (for "teaching," allowing the model to learn patterns), a 15% validation set (for adjusting model parameters during training to prevent overfitting), and a 15% test set (for the final evaluation of the model's true predictive ability on entirely new data). Random forests, gradient boosting machines, and support vector regression are chosen because they have good fitting capabilities for complex nonlinear relationships and can provide a ranking of feature importance. The training set's X and Y values ​​are input into the algorithm. The algorithm tries numerous internal rules, continuously comparing the error between its predicted Y_pred based on X and the true Y, and continuously adjusting the rules to minimize the error. Evaluation is performed using a test set (data the model has never seen before), with the core evaluation metrics being R² (coefficient of determination, closer to 1 the better) and RMSE (root mean square error, smaller the better). The scatter plot of the predicted and true values ​​on the test set should be closely distributed on both sides of the diagonal of y=x, indicating that the model training is complete, rather than merely memorizing training data.

[0045] A well-trained prediction model can ultimately achieve two functions: degradation prediction and state inversion. The degradation prediction is manifested as follows: after inputting the target environmental conditions (solution chemical parameters, expected number of wet-dry cycles), the complete degradation curve and final state of the rock mass mechanical properties are output. State inversion is manifested as follows: after inputting early deterioration data (such as wave velocity reduction value and surface crack development) of the rock mass monitored on site, the environmental erosion intensity (equivalent cycle number and erosion solution concentration range) experienced by the rock mass is estimated in reverse, so as to realize intelligent diagnosis of the engineering state.

[0046] S8: Construct a multi-scale coupled constitutive model framework. Based on real crack network data obtained from CT scans, and combined with chemical-mechanical theory, establish a discrete element numerical model to link the evolution of microcracks with macroscopic mechanical properties through the equivalent continuous medium theory.

[0047] In this embodiment, the quantitative functional relationship established in step S6 is a summary of phenomena, the multivariate machine learning prediction model established in step S7 is data-driven, and the multi-scale coupled constitutive model framework constructed in step S8 is a mechanism revelation. The goal of this constitutive model is to explain and predict crack evolution and its mechanical consequences based on physicochemical mechanisms. It does not rely on large amounts of data to "fit" or "learn" the phenomena, but rather attempts to describe basic physicochemical processes (such as intergranular corrosion, expansion stress, and crack propagation criteria) using mathematical equations. Its key parameters (such as the initial microstructure) need to be obtained from the experimental apparatus and methods of this embodiment (including CT scans and mechanical tests). The obtained experimental data (the monitoring data from the first and second sensor groups in step S3 of the aforementioned experimental methods) serves as its calibration basis. The calculated macroscopic mechanical properties can be compared with the results of empirical models or the target values ​​predicted by machine learning for mutual verification.

[0048] The experimental data in this embodiment (the monitoring data from the first and second sensor groups in step S3 of the aforementioned experimental method) is used to calibrate and verify key parameters in the theoretical model (such as the reduction factor of interparticle bond strength caused by chemical corrosion). The experimental data is used to adjust and confirm key settings in the response model so that the model can accurately simulate reality. We have a theoretical model that we want to use to simulate how chemical corrosion damages rocks from the inside, but some core "coefficients" (such as the reduction factor of interparticle bond strength) in this model control the unknown extent of the destructive force of chemical corrosion. The solution is to use this device to obtain real damage result data, and then adjust the coefficients in the calculation model in reverse until the simulated results output by the model completely correspond to the actual results measured experimentally. This process is called "calibration." Another set (or several sets) of completely new experimental data are used to test this calibrated model. If they all correspond, it means the model is reliable; this process is called "verification."

[0049] After establishing this response model, it can be integrated into an intelligent control system or independent data analysis software to form a closed loop. Its output directly serves engineering practice, and the output mainly includes the following: Output 1 (Material Property Envelope Chart): Generates an engineering chart with "solution concentration" and "number of wet-dry cycles" as coordinate axes and "strength retention rate" or "crack development level" as contour lines, providing an intuitive basis for parameter selection in engineering design for different corrosive environments.

[0050] Output 2 (Engineering Life Prediction and Early Warning Threshold): Combining local meteorological data (rainfall-evaporation data) and groundwater chemistry data, the model can convert the number of accelerated cycles in the laboratory into natural lifespan in the field. Based on this, the service life of engineering structures (such as slopes and foundations) is predicted, and quantitative thresholds for monitoring and early warning are set (for example, a level one early warning is triggered when the ultrasonic wave velocity drops to 80% of the initial value).

[0051] Output 3 (Evaluation of the effectiveness of reinforcement measures): The model is applied to red mudstone samples treated with different modifiers. By comparing the response differences before and after modification under the same chemical-dry-wet attack, the protective efficacy of various reinforcement measures is quantitatively evaluated, providing data support for optimizing treatment plans.

[0052] This embodiment constructs a response model between "chemical-wet-dry cycle parameters" and "multi-scale deterioration response of rock mass samples" using a rock mass chemical-wet-dry cycle coupled deterioration test device and its experimental method. This response model is the final goal and form of the model, answering the question of how the model quantitatively describes how "input parameters" lead to "output response." Step S7, "database training prediction model," is the specific method and process for constructing this model. It answers how the model is obtained, i.e., through machine learning algorithms that learn patterns from a database.

[0053] This response model is a data-driven prediction and evaluation system based on machine learning and physical constitutive models. It upgrades the technical solution of this embodiment from a simple simulation device into a prediction and evaluation platform, forming a complete technical closed loop from "environmental simulation → multi-scale observation → data mining → model building → engineering prediction." The response model endows the data collected by the device with engineering value, while the experimental setup and methods are the foundation for producing high-quality training data and verifying the model's reliability. The two complement each other, jointly solving the core problems of simulation distortion, incomplete observation, and lack of data-driven prediction in existing technologies.

[0054] The above embodiments are not intended to limit the shape, material, structure, etc. of the present invention in any way. Any simple modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention shall fall within the protection scope of the present invention.

[0055] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only used to facilitate the description of this invention and to simplify the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting the scope of protection of this invention.

[0056] If the terms "first" or "second" are used in this document to define components, those skilled in the art should know that the use of "first" or "second" is merely for the convenience of describing the invention and simplifying the description, and unless otherwise stated, the above terms have no special meaning.

[0057] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. However, these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A rock mass chemical-wet-dry cycle coupled degradation test device, comprising a chamber, wherein the bottom of the chamber is detachably provided with a chamber plate for sealing the chamber, characterized in that, The chamber includes a plate body, on which a guide plate is provided. A liquid outlet hole is provided at the lowest point of the guide plate. The plate body also holds a sample to be tested. The top of the chamber body is provided with a nozzle for spraying chemical solution onto the sample and a heating device for drying the sample. The side wall of the chamber body is provided with a fan for air drying the sample. The side wall of the chamber body is also provided with an observation window for easy observation of the sample. The chamber is also equipped with a first sensor group for detecting the chamber temperature, chamber humidity, sample surface temperature, sample surface humidity, and sample mass, as well as a second sensor group for detecting the pH value, conductivity, and concentration of the chemical solution.

2. A rock mass chemical-wet-dry cycle coupled deterioration test method, based on the rock mass chemical-wet-dry cycle coupled deterioration test device according to claim 1, characterized in that, Includes the following steps: S1: Sample preparation and pretreatment: The rock mass is processed into standard cylindrical or cubic samples, and its initial physical and mechanical parameters are measured. Then the samples are installed on the plate. S2: Experimental parameter setting: Different experimental parameters are set according to different research objectives; S3: Perform automated wet-dry cycle test: According to the test parameters set in step S2, automatically perform spray wetting, constant humidity standing, and drying operations. After completing one cycle, automatically enter the next cycle until the set number of cycles is reached. During this process, record the monitoring data of the first sensor group and the second sensor group. S4: Intermittent in-situ monitoring: Pause the test procedure at the set critical cycle points to acquire images of the sample; S5: Data post-processing and analysis: Combine the sample images collected in step S4 with the test parameters in step S2 to establish a response model.

3. The rock mass chemical-wet-dry cycle coupled degradation test method according to claim 2, characterized in that, The test parameters in step S2 include, Chemical parameters: different chemical solutions and their concentrations, pH values, and spray duration and sequence of different chemical solutions; Dry-wet cycle parameters: duration of spray wetting phase, constant humidity settling phase, and drying phase in a complete cycle, sample surface temperature, sample surface humidity, and heating / drying rate; Total number of loops: N, the total number of complete loops required to complete the experiment.

4. The rock mass chemical-wet-dry cycle coupled degradation test method according to claim 2, characterized in that, The key loop nodes in step S4 refer to the 10th / 20th / 50th complete loop.

5. The rock mass chemical-wet-dry cycle coupled degradation test method according to claim 2, characterized in that, The image acquisition of the sample in step S4 includes: The surface image of the sample is acquired through the observation window to obtain the crack rate, crack width, and crack fractal dimension of the sample. The chamber plate and the sample on it were removed from the chamber and moved to a micro-CT device for internal scanning to obtain three-dimensional crack network data. The three-dimensional crack network data included the three-dimensional volume, surface area, connectivity and pore size distribution of the cracks inside the sample. After the scanning was completed, the chamber plate and the sample were reinstalled on the chamber.

6. A rock mass chemical-wet-dry cycle coupled degradation response model, based on the rock mass chemical-wet-dry cycle coupled degradation test method according to any one of claims 2-5, characterized in that, The response model has an input X and an output Y, where the input X is the experimental parameter in step S2; The output Y is the sample response parameter, including the sample mass change and the sample image acquisition result in step S4.

7. The rock mass chemical-wet-dry cycle coupled degradation response model according to claim 6, characterized in that, The method for constructing a response model includes the following steps: S6: Establish a quantitative functional relationship between rock mass performance parameters and the number of wet-dry cycles and the concentration of erosion solution; S7: Establish a multivariate machine learning prediction model, use the quantitative function in step S6 as features, construct a dataset, and use a supervised learning algorithm to train the prediction model so that the prediction model can achieve two functions: degradation prediction and state inversion. S8: Construct a multi-scale coupled constitutive model framework. Based on real crack network data obtained from CT scans, and combined with chemical-mechanical theory, establish a discrete element numerical model to link the evolution of microcracks with macroscopic mechanical properties through the equivalent continuous medium theory.

8. The rock mass chemical-wet-dry cycle coupled degradation response model according to claim 7, characterized in that, The process of training the prediction model in step S7 is as follows: S71: After the device completes a set of comparative experiments covering different chemical concentrations and wet / dry regimes, it obtains an "input-response" database; S72: Use the database from step S71 to train the prediction model so that the prediction model learns to predict long-term key indicators that are difficult to obtain in real time or require destructive testing from readily available early or real-time data.

9. The rock mass chemical-wet-dry cycle coupled degradation response model according to claim 7, characterized in that, The degradation prediction of the prediction model in step S7 is manifested as follows: after inputting the target environmental conditions, the complete decay curve and final state of the rock mass mechanical properties are output. In step S7, the state inversion of the prediction model is manifested as follows: after inputting the early deterioration data of the rock mass monitored on site, the environmental erosion intensity it has experienced is estimated in reverse, so as to realize intelligent diagnosis of the engineering state.

10. The rock mass chemical-wet-dry cycle coupled degradation response model according to claim 7, characterized in that, It has the following output: Output 1 (Material Property Envelope Chart): Generates an engineering chart with "solution concentration" and "number of wet-dry cycles" as coordinate axes and "strength retention rate" or "crack development level" as contour lines, providing an intuitive basis for parameter selection in engineering design for different corrosive environments; Output 2 (Engineering Life Prediction and Early Warning Threshold): Combining local meteorological data (rainfall-evaporation data) and groundwater chemistry data, the model can convert the number of accelerated cycles in the laboratory into natural lifespan in the field. Based on this, the service life of engineering structures (such as slopes and foundations) is predicted, and quantitative thresholds for monitoring and early warning are set (for example, when the ultrasonic wave velocity drops to 80% of the initial value, a level one early warning is triggered). Output 3 (Evaluation of the effectiveness of reinforcement measures): The model is applied to red mudstone samples treated with different modifiers. By comparing the response differences before and after modification under the same chemical-dry-wet attack, the protective efficacy of various reinforcement measures is quantitatively evaluated, providing data support for optimizing treatment plans.

Images