Method and test device for simulating penetration deformation of deepwater rock

Abstract

The invention discloses a method and a test device for simulating deepwater rock seepage deformation, the test device comprises a pressurization mechanism, a pressure container and a data acquisition system which are matched with one another, and the test device is used for simulating deepwater rock seepage deformation so as to obtain water pressure, temperature and strain data with continuity and time series; the use method comprises the following steps: fitting the obtained water pressure, temperature and strain data with continuity and time series by adopting a plurality of time series models respectively, and analyzing the deformation rule of the rock mass by the change of water pressure and water temperature under the long-term action through the fitted data through a constructed test data prediction model.

Description

Technical Field

[0001] The invention relates to the technical field of water conservancy engineering, and in particular to a method and a test device for simulating deep-water rock seepage deformation. Background Art

[0002] Arch dam water storage will cause the fractured rock mass on both sides to be saturated and pressured. The fractured rock mass will deform under the high water pressure cycle and cause the arch dam valley to shrink. Excessive valley shrinkage will seriously threaten the safe operation of the arch dam. Studies have shown that most arch dams in my country have a continuous valley shrinkage problem. The expansion of rocks on both sides of the valley after water storage is one of the main reasons for the valley shrinkage. Therefore, revealing the deformation mechanism of rock under high water pressure has become an important issue that needs to be solved urgently.

[0003] In the current research field of rock deformation under high water pressure, the rock failure process under different confining pressure conditions is usually studied. The confining pressure is high, but in actual engineering, the water pressure changes caused by the water level changes of high dam reservoirs are not large but the number of times is large. Large rock servo testers can provide high confining pressure, which is easy to cause large deviations in data measurement. Using other types of data acquisition systems will cause asynchronous data collection, which will lead to complex data processing and other problems. In addition, large rock servo testers are expensive, and general researchers do not have such research conditions. Finally, it is difficult to simulate long-term (monthly, annual) water pressure changes in indoor tests, and it is difficult to obtain the corresponding rock deformation laws.

[0004] Based on this, the existing test equipment and methods still have great limitations in studying deep-water rock seepage and deformation problems. The core problems are measurement errors of large equipment, high research costs and difficulty in conducting annual long-term tests. Summary of the invention

[0005] In order to solve the current technical problems, the main purpose of the present invention is to provide a method and test device for simulating deep-water rock seepage deformation. By establishing a prediction model and adopting a combination of experiments and models to analyze the deformation law of rock mass under long-term effects, the data measurement error is reduced while reducing the cost of such research.

[0006] The technical solution adopted by the present invention is: a method for simulating deep-water rock seepage deformation, comprising: S10, obtaining water pressure, temperature and strain data of the test core to obtain a time series data set of various data; S20, fitting the data set with multiple time series models respectively; S30, the fitted data is processed by the test data prediction model and then the test results are output. The test data prediction model is defined as follows: ; in: The experimental data after weighted calculation for model fitting prediction, For the timing model The fitting results are The test data at the time, For the timing model The coefficient of determination of the fitted data is is the corresponding timing model; S40. The more accurate test data obtained by fitting is brought into the established water pressure-strain relationship for inversion to obtain the pore water pressure value in the test. The mathematical model of the water pressure-strain relationship is as follows: ; in: is the rock strain data, is the temperature strain, is the water pressure strain, is the coefficient of thermal expansion, is the original length, is the current temperature, is the initial temperature, is the total stress, is the interaction coefficient with water, is the pore water pressure, represents the rock bulk elastic modulus, represents the bulk elastic modulus of rock skeleton, represents the initial porosity, represents the pore pressure difference, Indicates the current porosity.

[0007] In S1, it also includes: S11: homogenize the test cores of the same group to make them consistent in size, diameter and height; S12: After cleaning the surface of the test core, strain gauges are attached to the top, bottom and sides; S13: placing the temperature sensor and the pressure sensor, as well as the test core with the strain gauge attached thereto, into the pressure vessel, and setting the data acquisition frequency of the dynamic data acquisition system at the same time; S14: injecting water of corresponding temperature into the booster mechanism according to the established working conditions of the test, then starting the booster mechanism to adjust the internal pressure of the pressure vessel, and starting the data acquisition system to start recording the internal pressure and temperature of the pressure vessel, as well as the strain data of the test core, taking out the test core after the test, and draining the water in the booster mechanism and the pressure vessel; S15: Repeat S12-S14 to analyze the data and mechanism presented by the deformation of the test core under deep water pressure, and obtain a time series data set of water pressure, temperature and strain data.

[0008] In S13, the strain gauge adopts a full-bridge strain gauge.

[0009] A test device for simulating deep-water rock seepage deformation, used to implement the method for simulating deep-water rock seepage deformation, the test device comprising: A pressure vessel, the pressure vessel comprising a test tank for placing a test core, a detachable top cover being provided on the top of the test tank, and a pressure boosting valve and a pressure relief valve being mounted on the top cover; A data acquisition system, the data acquisition system comprising a dynamic data acquisition system, and a temperature sensor, a pressure sensor and a plurality of strain gauges disposed on the test core, which are respectively connected to the dynamic data acquisition system and disposed in the test tank; The boosting mechanism is connected to the boosting valve of the pressure vessel and is used to inject pressurized water of different temperatures into the test tank.

[0010] The boosting mechanism comprises a water tank and a boosting pump which are connected to each other. The boosting pump has a pressure regulating valve block. The water outlet of the boosting pump is connected to the boosting valve through a high-pressure water pipe.

[0011] The pressure relief valve is connected to the water tank through a pressure relief pipe.

[0012] The thickness of the pressure vessel is determined by the following formula: ; In the formula, is the thickness of the pressure vessel, mm; is the maximum pressure of the pressure vessel, MPa; is the internal diameter of the pressure vessel; Allowable stress of materials used to manufacture pressure vessels; is the welding coefficient, which is taken as 1.0.

[0013] The top cover is also equipped with a pressure measuring device.

[0014] The top cover is connected to the test tank by bolts, and a silicone rubber plate is arranged between the test tank and the top cover; three assembly holes with two ends through are opened on the top cover, and the boost valve, the pressure relief valve and the pressure measuring device are assembled on the assembly holes one by one.

[0015] The temperature sensor, the pressure sensor and the plurality of strain gauges are all connected to the dynamic data acquisition system via transmission lines. The top cover is also provided with threading holes for the transmission lines to pass through, and anti-seepage treatment is performed between the transmission lines and the threading holes.

[0016] The present invention has the following beneficial effects: The present invention utilizes the set test device to simulate the seepage deformation of deep-water rocks, and then obtains water pressure, temperature and strain data with continuity and time series. Then, through the obtained water pressure, temperature and strain data with continuity and time series, a test data prediction model is constructed to realize the deformation law analysis of the rock mass under the long-term action of water pressure and water temperature changes. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 It is a schematic diagram of the overall structure of the test device of the present invention.

[0019] Figure 2 It is a schematic diagram of the structure of the top cover in the test device of the present invention.

[0020] Figure 3 The figure shows the result of fitting the experimental data through LSTM, BP and BiLSTM time series models.

[0021] Figure 4 This is a graph showing the results of fitting and predicting the test data using the test data prediction model.

[0022] Figure identification: 11. Water tank, 12. Booster pump, 13. Pressure regulating valve block, 21. Test tank, 22. Top cover, 23. Booster valve, 24. Pressure relief valve, 25. Pressure measuring device, 26. Assembly hole, 27. Threading hole, 31. Dynamic data acquisition system, 32. Temperature sensor, 33. Pressure sensor, 34. Strain gauge, 35. Transmission line, 40. Test core, 50. High-pressure water pipe, 60. Pressure relief pipe. DETAILED DESCRIPTION

[0023] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0024] This embodiment discloses a test device and a method for using the device to simulate deep-water rock seepage deformation, and adopts a method combining test and model to analyze the deformation law of rock mass under long-term action.

[0025] Under the action of water pressure at the hundred-meter level, the elastic modulus of rock specimens is relatively high and they are still in an elastic state. However, due to the presence of pores or cracks inside the rock, and the water level of the reservoir will change continuously according to the operation status of the reservoir, under the action of long-term changing deep water pressure, high-pressure water will continue to penetrate into the pores from the cracks on the rock surface, damaging the rock pore channels and causing deformation of the rock surface. At present, the sensors used to measure rock deformation are strain gauges and strain gauges. Concrete tests can be measured with strain gauges, while core experiments can be measured by sticking strain gauges on the surface of the core. Strain gauges are divided into full-bridge connection, half-bridge connection and 1 / 4-bridge connection according to different connection methods, and the corresponding accuracy and anti-interference performance are from high to low. Since the test is measured underwater throughout the whole process, full-bridge strain gauges are used. Rock deformation is also affected by temperature and environmental parameters. Therefore, the stress and temperature parameters at the location of the core need to be measured while measuring the core strain. Therefore, the adjustment of stress and temperature parameters needs to be fully considered when designing the test device and the test.

[0026] like Figure 1-2 The test device for simulating deep-water rock seepage deformation shown in the figure can measure water pressure and temperature data with time series and rock strain data in the test of simulating deep-water rock seepage deformation, and the obtained test data are all continuous. The specific structure includes: a boosting mechanism, the boosting mechanism includes a water tank 11 and a boosting pump 12 connected to each other, and the boosting pump 12 has a pressure regulating valve block 13; a pressure vessel, the pressure vessel includes a test tank 21 for placing a test rock core 40, a detachable top cover 22 is provided on the top of the test tank 21, and a boosting valve 23 and a pressure relief valve 24 are assembled on the top cover 22, the water outlet end of the boosting pump 12 is connected to the boosting valve 23 through a high-pressure water pipe 50, and the pressure relief valve 24 is connected to the boosting valve 23 through a high-pressure water pipe 50. The door 24 is connected to the water tank 11 through the pressure relief pipe 60; the data acquisition system includes a dynamic data acquisition system 31, and a temperature sensor 32, a pressure sensor 33 and a plurality of strain gauges 34 arranged on the test core 40, which are respectively connected to the dynamic data acquisition system 31 and placed in the test tank 21, and a booster mechanism is used to adjust and control the pressure inside the pressure vessel during the experiment; the data acquisition system is used to collect temperature and pressure data of the environment in which the test core 40 is located during the experiment, as well as deformation data of the test core 40 under temperature and pressure data; during the experiment, the pressure vessel is used to test the test core 40, and it cooperates with the booster mechanism to form different test environments.

[0027] During the specific implementation process, the booster mechanism provides a stable high water pressure to the pressure vessel. The pressure vessel is pressurized by the internal water pressure of the pressure relief valve 24 through the pressure relief pipe 60. The high-pressure water inside the pressure vessel penetrates into the test core 40 through the surface cracks of the test core 40, inducing a slight deformation of the surface of the test core 40, which is monitored by the strain gauge 34 pasted on the surface of the test core 40 and reflected in the form of strain data. The environmental water pressure data and temperature data of the test core 40 are detected by the pressure sensor 33 and the temperature sensor 32 respectively, and the surface strain data of the test core 40, the environmental water pressure data and the temperature data of the test core 40 are collected simultaneously by the dynamic data acquisition system 31. Subsequently, the LSTM, BP and BiLSTM time series models are used to fit and predict the test data respectively. Finally, the deformation mode and deformation mechanism of the test core 40 under the condition of long-term deep-water pressure change are analyzed through the stress, strain and temperature data processed by the model.

[0028] The thickness of the above pressure vessel is determined by the following formula: ; In the formula, is the thickness of the pressure vessel, mm; is the maximum pressure of the pressure vessel, MPa; is the internal diameter of the pressure vessel; Allowable stress of materials used to manufacture pressure vessels; is the welding coefficient, which is taken as 1.0.

[0029] As a preferred solution of the above embodiment, the top cover 22 is also equipped with a pressure measuring device 25. During the experiment, the pressure measuring device 25 can read the internal pressure of the pressure vessel in real time so that the boosting mechanism can maintain the internal pressure of the pressure vessel stable at a predetermined working condition.

[0030] As a preferred solution of the above embodiment, the top cover 22 is connected to the test tank 21 by bolts, and a silicone rubber plate is provided between the test tank 21 and the top cover 22 to ensure the air tightness between the top cover 22 and the test tank 21. A plurality of assembly holes 26 with through ends are provided on the top cover 22, and the boost valve 23, the pressure relief valve 24 and the pressure measuring device 25 are assembled one by one on the plurality of assembly holes 26.

[0031] As a preferred solution of the above embodiment, the temperature sensor 32, the pressure sensor 33 and the plurality of strain gauges 34 are connected to the dynamic data acquisition system 31 via a transmission line 35 to ensure the stability of data transmission. Specifically, a threading hole 27 for the transmission line 35 to pass through is provided on the top cover 22, and an anti-seepage treatment is performed between the transmission line 35 and the threading hole 27.

[0032] Based on the above-mentioned test device for simulating deep-water rock seepage deformation, the test method includes the following steps: S1: homogenizing the test cores 40 of the same group to make them consistent in size, diameter and height so as to reduce external factors that affect the deformation of the test cores 40. Specifically, the test cores 40 may be natural cores. Since the shapes and sizes of natural cores vary during the mining process, the sizes of the tests of the same group need to be consistent during the test. A rock cutter and a grinder may be used to homogenize the natural cores to ensure that the diameters of the tests of the same group are highly consistent. The test cores 40 may also be concrete test blocks, but a unified mold needs to be used when preparing the concrete. S2: After cleaning the surface of the test core 40, strain gauges 34 are attached to the top, bottom and circumferential sides to measure the axial, radial and circumferential deformations of the test core 40 under high water pressure changes. The strain gauges can be divided into full-bridge connection, half-bridge connection and 1 / 4-bridge connection according to different connection methods, and the corresponding accuracy and anti-interference performance are from high to low. Considering that the whole test is underwater measurement, the accuracy requirement is high, so the strain gauge 34 uses a full-bridge strain gauge; S3: Place the temperature sensor 32 and the pressure sensor 33, as well as the test core 40 with the strain gauge 34 attached thereto, into the pressure vessel, and simultaneously set the data acquisition frequency of the dynamic data acquisition system 31, and check the connection status of the test equipment; S4: according to the established working conditions of the test, water of corresponding temperature is injected into the water tank 11, and then the booster mechanism is started to adjust the internal pressure of the pressure vessel, and the data acquisition system is started to record the internal pressure and temperature of the pressure vessel, as well as the strain data of the test core 40. After the test, the test core 40 is taken out, and the water in the booster mechanism and the pressure vessel is discharged. During this experiment, the dynamic data acquisition system 31 obtains water pressure and temperature data with continuity and time series, as well as the strain data of the test core 40; S5: Repeat S2-4 to analyze the data and mechanism presented by the deformation of the core under deep water pressure. The purpose of repeatedly obtaining data is to control the test. Generally speaking, the test is full of errors. Multiple tests can reduce the errors to achieve more accurate test conclusions. Specifically, since the above test processes are the same, but the loading conditions are inconsistent, it is necessary to repeat the experiment to obtain data to reduce the test error. S6: The obtained series test data of the test core 40 are respectively subjected to regression fitting using the time series models LSTM, BP and BiLSTM. The fitted data are processed by the test data prediction model and then the test results are output, thereby analyzing the deformation law of the test core 40 under the change of water pressure.

[0033] Specifically, the above LSTM is a special recurrent neural network (RNN) that controls the flow of information through three main gates (input gate, forget gate, and output gate). It performs well in processing time series data and sequence data, especially in the context modeling of long sequences. LSTM can effectively avoid the gradient vanishing or gradient exploding problems encountered by traditional RNN models during training. The model function at any time is expressed as follows: ; ; ; ; ; ; (1) In formula (1), express Controls whether information flows into the cell unit at all times; The flag controls whether the information in the cell unit at the previous moment is accumulated into the cell unit at the current moment; Control the current Whether the information in the cell unit at this moment flows into the current hidden state middle; Represents a long-term memory unit, which represents the memory of the neuron state, so that the LSTM unit has the ability to save, read, reset and update long-distance historical information; Indicates the current memory unit; , , , The weight matrix of the forget gate, input gate, output gate, and computational unit state; , , , Represent the bias terms of the forget gate, input gate, output gate and computing unit state respectively; and There are two activation functions respectively.

[0034] LSTM, through its unique gating structure, can maintain good stability when processing time series data and can capture long-term dependency features. LSTM is widely used in various time series prediction tasks.

[0035] The above-mentioned BP neural network is a feedforward artificial neural network. Its basic structure consists of an input layer, a hidden layer, and an output layer. It optimizes the weights of the network through a back-propagation algorithm to minimize the error between the output and the target. It focuses on the forward and back-propagation processes of sequence data, which respectively calculate the predicted value and use the error to correct the prediction model parameters.

[0036] Forward pass: Assume a single-layer RNN network model, where the input is time series data , the output is the predicted value , and use the activation function ,include or .

[0037] Weights from input layer to hidden layer: ; (1) The weights from the hidden layer to the output layer are: ; (2) In the formula, is the current time step The hidden layer state of is the weight matrix input to the hidden layer; is the input of the current time step; is the bias term of the hidden layer; is the output of the model (i.e., the predicted value); is the weight matrix from the hidden layer to the output layer; is the bias term of the output layer.

[0038] Backward Propagation: Output layer gradient: ; (3) Hidden layer gradient: ; (4) Gradient of the input layer: ; (5) Parameter update: ; ; ; ; (7) In the formula, is the output layer error; is the error of the hidden layer; is the activation function in the current hidden state The derivative at ; Represents the current time step input; Represents the learning rate.

[0039] The BP time series model optimizes the weights of the neural network through the back propagation algorithm, which can better fit the nonlinear patterns in the time series and is suitable for various types of prediction tasks. However, due to its shortcomings in modeling long-term dependencies, it usually needs to be combined with other network structures (such as LSTM) to give full play to its advantages in complex time series problems. In practical applications, the BP network is usually used as a basic prediction tool in combination with other models.

[0040] BiLSTM is a recurrent neural network (RNN) based on the LSTM structure, which is used to process and predict time series data. Unlike the traditional unidirectional LSTM, BiLSTM can capture more contextual information by learning information in both directions (forward and backward) of time series data at the same time, especially for data with long-term dependencies.

[0041] BiLSTM data passes through two LSTM networks at the same time, one is the forward (from front to back) LSTM, and the other is the reverse (from back to front) LSTM, and finally their outputs are combined. It is an extension of the standard LSTM, which outputs the states of the forward LSTM and the reverse LSTM at the same time, and finally concatenates or weighted merges their results. Assume that the forward output is , the reverse output is , the final hidden state is usually concatenated from the outputs of the forward and backward LSTMs: ; (6) In the formula, It indicates that the forward LSTM is at time step Output: Represents the reverse LSTM at time step Output.

[0042] The BiLSTM time series model combines forward and backward learning capabilities, showing significant advantages in many practical time series prediction and classification tasks. By considering both past and future information, BiLSTM can provide more accurate predictions, especially in scenarios with complex time dependencies.

[0043] Three models were selected for simulation prediction of the above series of data obtained from the experiment, and the model determination coefficient R2 determination coefficient was used as the weight to perform weighted average on the fitted prediction data to obtain the experimental data prediction model: ; (7) In the formula, It represents the experimental data after the model fitting prediction weighted calculation; Representation Model Fitting results Test data at the moment; Representation Model The coefficient of determination of the fitted data, It is a time series model LSTM, BP or BiLSTM.

[0044] Calculated by formula (7) That is, the test data prediction model outputs the test results. The method is also applicable to model processing of water pressure, temperature, and strain data. According to the corresponding relationship between the three, the deformation law of rock under water pressure changes can be analyzed.

[0045] Finally, the accuracy of the test data was further improved according to the determination coefficient in the model evaluation index. The specific method is as follows: the stress data obtained from the test was selected as a sample, and the LSTM, BP and BiLSTM time series models were selected for fitting and prediction. The fitting results Figure 3 As shown in Table 1: Table 1 is the model evaluation table

[0046] Depend on Figure 3 It can be seen that the gap between the model prediction data and the true value is small, and the prediction result is good. It can be seen from the model evaluation table 1 that the overall prediction accuracy of the model is high and the feasibility is strong.

[0047] Depend on Figure 3 The test data calculated in Table 1 and formula (7) are the results of fitting and prediction by the test data prediction model. Figure 4 As shown, the corresponding temperature and strain data can be calculated by the same logic, and long-term predictions can be made to obtain the deformation mode and deformation mechanism of the core under water pressure.

[0048] The effects of temperature and water pressure on rock deformation are particularly important for fitting and predicting the experimental results.

[0049] 1. Effect of temperature on rock deformation: Temperature changes have a significant impact on the mechanical properties of rocks. Rocks exhibit different deformation behaviors at different temperatures, mainly including the following: ① Thermal expansion: When the temperature rises, the minerals in the rock will expand. This expansion may cause the rock to increase in volume, which in turn causes internal stress and may even lead to the formation of cracks. Generally, an increase in temperature will cause the elastic modulus of the rock to decrease and its deformation capacity to increase.

[0050] ②Degradation and softening: When the temperature reaches a certain level, the mineral composition of the rock may degrade or soften. For example, at high temperatures, some minerals (such as quartz) may undergo changes in their crystal structure, causing the rock to become more deformable.

[0051] Common rocks will only degrade and soften when the temperature reaches above 800℃. For this test and the corresponding actual project, the ambient temperature of the rock is 30-50℃. At this temperature, the rock will not degrade or soften. Therefore, the effect of temperature on rock deformation only includes thermal expansion. The thermal expansion formula is: ; (8) In the formula, is the change in rock length, is the coefficient of thermal expansion, is the original length, is the current temperature, is the initial temperature.

[0052] The change in rock length can be converted into strain change according to the principle of strain formula: ; (9) In the formula, is the temperature strain component.

[0053] 2. Effect of water pressure on rock deformation: Water pressure also has a significant effect on rock deformation, especially when the rock is in a groundwater-saturated environment. The effect of water pressure on rock deformation is mainly reflected in the following aspects: ① Pore water pressure: When the pores of a rock are filled with water, the pressure of the water exerts additional stress on the interior of the rock, especially when the rock is in compression. The water pressure pushes the water in the pores of the rock, causing the rock to expand and deform.

[0054] ② The influence of water pressure on pores: The presence of water in rock pores may change the microstructure of the rock. For example, water can penetrate into the microcracks of the rock, causing the cracks to expand further, thereby reducing the strength of the rock and making it more susceptible to deformation.

[0055] ③ Lubricating effect of water: Water acts as a lubricant in the cracks of rocks, making the cracks easier to expand and thus promoting plastic deformation of rocks.

[0056] The effect of pore water pressure on rock strength and deformation can be described by the effective stress principle of water pressure: ;(10) In the formula, is the effective stress, is the total stress, is the interaction coefficient with water, is the pore water pressure.

[0057] The effect of water pressure on pores and the lubrication of water will change the physical parameters of the rock, directly affecting the bulk modulus of the rock: ;(11) In the formula, represents the rock bulk elastic modulus, represents the bulk elastic modulus of rock skeleton, represents the initial porosity, represents the pore pressure difference, Indicates the current porosity.

[0058] The mathematical formula for the effect of water pressure change on rock strain is: ;(12) In the formula, is the water pressure strain component.

[0059] The total strain formula of rock under water pressure and temperature changes is: ;(13) In this formula, , , , It is the essential parameter of rock material and can be obtained by experience. is the rock porosity, which can be measured by instruments. , , , Can be measured in an experiment, i.e., the independent variable.

[0060] Since the actual pore water pressure of rock cannot be measured in the test, the test data can be inverted first to calculate the actual pore water pressure change curve of rock during the test, and the curve can be extended according to the law. Finally, the change law of rock under the action of long-term water pressure cycle can be calculated based on pore water pressure, temperature and external water pressure.

[0061] Of course, the present invention may have many other embodiments. Without departing from the spirit and essence of the present invention, technicians familiar with the field may make various corresponding changes and modifications based on the present invention, but these corresponding changes and modifications should all fall within the scope of protection of the claims attached to the present invention.

Claims

1. A method for simulating deep-water rock seepage deformation, characterized in that: include: S10, obtaining water pressure, temperature and strain data of the test core (40) to obtain a time series data set of each data; S20, fitting the data set with multiple time series models respectively; S30, the fitted data is processed by the test data prediction model and then the experimental results are output. The test data fitting model is defined as follows: ； in: The experimental data after weighted calculation for model fitting prediction, For the timing model The fitting results are The test data at the time, For the timing model The coefficient of determination of the fitted data is is the corresponding timing model; S40, the test data obtained by fitting is brought into the established water pressure-strain relationship for inversion to obtain the pore water pressure value in the test. The mathematical model of the water pressure-strain relationship is as follows: ； in: is the rock strain data, is the temperature strain, is the water pressure strain, is the coefficient of thermal expansion, is the original length, is the current temperature, is the initial temperature, is the total stress, is the interaction coefficient with water, is the pore water pressure, represents the rock bulk elastic modulus, represents the bulk elastic modulus of rock skeleton, represents the initial porosity, represents the pore pressure difference, Indicates the current porosity.

2. A method for simulating deep-water rock seepage deformation according to claim 1, characterized in that: The time series models include LSTM models, BP models and BiLSTM models.

3. The method for simulating deep-water rock seepage deformation according to claim 1, characterized in that: In S1, it also includes: S11: homogenizing the test cores (40) of the same group to make their sizes, diameters and heights consistent; S12: After cleaning the surface of the test core (40), strain gauges (34) are attached to the top, bottom and circumference; S13: placing the temperature sensor (32) and the pressure sensor (33), as well as the test rock core (40) with the strain gauge (34) attached thereto, into a pressure vessel, and simultaneously setting the data acquisition frequency of the dynamic data acquisition system (31); S14: according to the established working conditions of the test, water having a temperature corresponding to the initial working conditions is injected into the pressurizing mechanism, and then the pressurizing mechanism is started to adjust the internal pressure of the pressure vessel, and the data acquisition system is started to record the changes in the internal pressure and temperature of the pressure vessel, as well as the strain data of the test core (40). After the test is completed, the test core (40) is taken out, and the water in the pressurizing mechanism and the pressure vessel is discharged; S15: Repeat S12-S14 to analyze the data and mechanism presented by the deformation of the test core (40) under deep water pressure, and obtain a time series data set of water pressure, temperature and strain data.

4. A method for simulating deep-water rock seepage deformation according to claim 3, characterized in that: In S13, the strain gauge (34) is a full-bridge strain gauge.

5. A test device for simulating deep-water rock seepage deformation, characterized in that: For implementing the method for simulating deep-water rock seepage deformation as described in any one of claims 1 to 3, the test device comprises: A pressure vessel, the pressure vessel comprising a test tank (21) for placing a test core (40), the top of the test tank (21) being provided with a detachable top cover (22), the top cover (22) being equipped with a pressure boosting valve (23) and a pressure relief valve (24); A data acquisition system, the data acquisition system comprising a dynamic data acquisition system (31), a temperature sensor (32), a pressure sensor (33) respectively connected to the dynamic data acquisition system (31) and placed in the test tank (21), and a plurality of strain gauges (34) arranged on the test core (40); A pressure boosting mechanism is connected to a pressure boosting valve (23) of the pressure container and is used to inject pressurized water into the test tank (21).

6. The test device for simulating deep-water rock seepage deformation according to claim 5, characterized in that: The boosting mechanism comprises a water tank (11) and a boosting pump (12) which are connected to each other. The boosting pump (12) has a pressure regulating valve block (13). The water outlet of the boosting pump (12) is connected to the boosting valve (23) via a high-pressure water pipe (50).

7. The test device for simulating deep-water rock seepage deformation according to claim 6, characterized in that: The pressure relief valve (24) is connected to the water tank (11) via a pressure relief pipe (60).

8. The test device for simulating deep-water rock seepage deformation according to claim 5, characterized in that: The thickness of the pressure vessel is determined by the following formula: ； In the formula, is the thickness of the pressure vessel, mm; is the maximum pressure of the pressure vessel, MPa; is the internal diameter of the pressure vessel; Allowable stress of materials used to manufacture pressure vessels; is the welding coefficient, which is taken as 1.

0.

9. The test device for simulating deep-water rock seepage deformation according to claim 5, characterized in that: The top cover (22) is also equipped with a pressure measuring device (25); the top cover (22) is connected to the test tank (21) by bolts, and a silicone rubber plate is provided between the test tank (21) and the top cover (22); the top cover (22) is provided with three assembly holes (26) with two ends passing through, and the boost valve (23), the pressure relief valve (24) and the pressure measuring device (25) are assembled one by one on the assembly holes (26).

10. The test device for simulating deep-water rock seepage deformation according to claim 5, characterized in that: The temperature sensor (32), the pressure sensor (33), and the plurality of strain gauges (34) are connected to the dynamic data acquisition system (31) via transmission lines (35); a threading hole (27) for the transmission line (35) to pass through is also provided on the top cover (22); and an anti-seepage treatment is performed between the transmission line (35) and the threading hole (27).

Images