Carbon dioxide phase change fracturing pressure decoupling test and data driving prediction method

Abstract

The invention discloses a carbon dioxide phase change fracturing pressure decoupling test and data driving prediction method. According to the method, during testing, a multipoint pressure testing means is effectively used, a transverse testing hole is ingeniously formed, and a pressure sensor is arranged in the hole, so that refined measurement of quasi-static gas wedge pressure in the carbon dioxide phase change fracturing process is achieved, and a carbon dioxide phase change fracturing stress wave and gas wedge pressure curve is obtained; the coupling action process of stress waves and high-pressure gas in the carbon dioxide phase change fracturing process can be effectively reflected. The two fracturing pressure curves obtained by the method have the characteristic of high fidelity, and can be directly used as load definition basis for numerical simulation of carbon dioxide phase change fracturing rock breaking. Based on a carbon dioxide phase change fracturing stress wave and gas wedge pressure curve obtained by a pressure decoupling test, high-precision data driving prediction of the carbon dioxide phase change fracturing stress wave and gas wedge pressure can be realized in combination with an encoder-decoder frame.

Description

Technical Field

[0001] This invention patent relates to the field of deep engineering geology, and in particular to a method for decoupling testing and data-driven prediction of carbon dioxide phase transformation-induced fracturing pressure. Background Technology

[0002] In recent years, carbon dioxide phase change fracturing technology has been considered an effective fracturing method for deep coalbed methane and hot dry rock reservoir stimulation due to its advantages such as no need for water injection, no pollution, low disturbance, and the ability to induce the formation of complex fracture networks. When using carbon dioxide phase change fracturing technology to stimulate a reservoir, the liquid carbon dioxide inside the fracturing tube first undergoes an endothermic phase transition to a supercritical state under the action of an activator. This process is accompanied by a continuous increase in pressure within the fracturing tube. When the pressure inside the fracturing tube exceeds the rated pressure of the rupture disc, the rupture disc ruptures and fails, and the supercritical carbon dioxide inside the tube instantly depressurizes and undergoes a phase transition. During this process, the carbon dioxide rapidly expands in volume, and the resulting high-pressure carbon dioxide gas rapidly impacts the reservoir rock mass. At this point, the reservoir rock mass fractures under the combined action of the carbon dioxide phase change gas explosion stress wave and the gas wedge pressure, forming a fracturing network within the rock mass. This process involves a complex supercritical carbon dioxide phase change expansion and impact process, where the explosive stress wave generated by the instantaneous expansion of supercritical carbon dioxide synergistically couples with the high-pressure carbon dioxide gas to break the rock.

[0003] Currently, carbon dioxide phase change fracturing load testing mainly focuses on three typical fracturing pressures within the carbon dioxide phase change fracturing explosion field: the pressure inside the fracturing borehole, the free-field overpressure, and the pressure at the fracturing borehole wall. The question of how stress waves and high-pressure carbon dioxide gas work together to break rock during carbon dioxide phase change fracturing remains unclear. Consequently, the rock-breaking mechanism of carbon dioxide phase change fracturing has not been accurately revealed. In fact, accurately revealing the rock-breaking mechanism of carbon dioxide phase change fracturing under the action of stress waves and gas pressure depends on the accurate quantitative characterization of carbon dioxide phase change fracturing stress waves and high-pressure carbon dioxide pressure, and the refined characterization of these two depends on the precise decoupling test of carbon dioxide phase change fracturing pressure. To date, a test method capable of achieving high-precision decoupling characterization of carbon dioxide phase change fracturing pressure has not been proposed, and a systematic decoupling characterization of fracturing stress waves and gas wedge pressure is still lacking, restricting a deeper analysis of their coupled rock-breaking mechanism. Summary of the Invention

[0004] To address the shortcomings, this invention aims to provide a decoupled testing and data-driven prediction method for carbon dioxide phase change fracturing pressure. This invention achieves decoupled testing of carbon dioxide phase change fracturing pressure by using a series of circumferential pressure sensors arranged around the fracturing borehole wall and a lateral pressure sensor placed in a transverse test hole perpendicular to the fracturing borehole. By using pressure curves measured at different locations, the decoupled characterization of carbon dioxide phase change fracturing stress wave and gas wedge pressure is completed, thus obtaining a refined measurement of the quasi-static gas wedge pressure during carbon dioxide phase change fracturing. The obtained carbon dioxide phase change fracturing stress wave and gas wedge pressure curves can be directly used as the load definition basis for numerical simulation of carbon dioxide phase change fracturing rock breaking. Based on this, a data-driven framework is introduced, and a high-precision prediction method applicable to carbon dioxide phase change fracturing stress wave and gas wedge pressure curves is proposed.

[0005] To address the problems in the existing technology, the technical solution adopted by this invention is as follows:

[0006] A decoupled testing and data-driven prediction method for carbon dioxide phase transformation-induced cracking pressure includes the following steps:

[0007] Step 1: Construct a carbon dioxide phase change-induced pyrolysis coupling pressure testing device;

[0008] Step 2: Take a precast rock fracture specimen and drill mutually perpendicular fracture holes and transverse test holes in the specimen. The axis of the fracture hole is perpendicular to the plane where its opening is located. A transverse test hole is also drilled on the vertical plane of the fracture hole. The axis of the transverse test hole is perpendicular to the plane where its opening is located. The fracture hole and the transverse test hole are connected and connected at the bottom sidewall of the fracture hole.

[0009] Step 3: After installing several circumferential PVDF pressure sensors and lateral PVDF pressure sensors on the inner circumferential sidewall of the bottom of the fracturing hole and in the transverse test hole, connect the data transmission lines of the circumferential PVDF pressure sensors and lateral PVDF pressure sensors to the pressure monitoring instrument; the empty part of the transverse test hole is then sealed with raw materials to ensure that no high-pressure carbon dioxide gas is ejected from the transverse test hole during the fracturing process, and each circumferential PVDF pressure sensor is installed at the same elevation on the axis of the transverse test hole;

[0010] Step 4: Insert the fracturing tube into the fracturing hole, and make sure that the line connecting the two pressure relief ports on the fracturing tube coincides with the axis of the transverse test hole.

[0011] Step 5: Inject the raw material of the rock fracturing specimen into the fracturing hole, and ensure that the sealing material fully fills the borehole gap above each circumferential PVDF pressure sensor. After the sealing material is cured, the fracturing tube and the fracturing specimen form an integral whole, and the end of the transverse test hole near the fracturing hole is in the same confined space as the bottom of the fracturing hole.

[0012] Step 6: Connect the fracturing tube to the solenoid valve and control the solenoid valve to close, so that a sealed space is formed inside the pressure vessel; connect the temperature sensor to the temperature monitor, and the piezoresistive pressure sensor to the pressure monitor, and start real-time monitoring of the temperature and pressure of the carbon dioxide phase change fracturing test inside the pressure vessel.

[0013] Step 7: Open the carbon dioxide cylinder and start the air compressor to force the liquid carbon dioxide in the cylinder into the storage tank inside the control console; open the injection valve on the pressure vessel, start the control console, and inject the liquid carbon dioxide into the pressure vessel. After the carbon dioxide is fully mixed with the air inside the pressure vessel, disconnect the connection between the control console and the injection valve to allow the liquid carbon dioxide to leak out from the injection valve. Repeat the above operation several times to thoroughly clean the pressure vessel and ensure that the pressure vessel contains only carbon dioxide gas before the fracturing pressure test; start the control console to inject liquid carbon dioxide into the pressure vessel and monitor the temperature and pressure of the carbon dioxide inside the pressure vessel in real time, and calculate the real-time filling amount of carbon dioxide in the pressure vessel based on the monitoring results.

[0014] Step 8: After the real-time carbon dioxide filling amount in the pressure vessel reaches the carbon dioxide test filling amount, close the injection port and disconnect the connection between the control console and the pressure vessel; start the heating jacket outside the pressure vessel to heat the pressure vessel, so that the internal temperature and pressure of the pressure vessel continue to rise, and continuously monitor the temperature and pressure inside the pressure vessel; after the temperature and pressure inside the pressure vessel reach the carbon dioxide phase change cracking test temperature and pressure, control the heating jacket to enter the heat preservation mode.

[0015] Step 9: Using the controller, the solenoid valve is opened, and the pressure vessel is instantly connected to the fracturing tube. At this time, the pressure inside the pressure vessel drops instantly, and the supercritical carbon dioxide is instantly depressurized and expanded, causing carbon dioxide phase change fracturing. The circumferential PVDF pressure sensors and lateral PVDF pressure sensors inside the fracturing specimen record the carbon dioxide phase change fracturing pressure curve inside the specimen. The curve collected by the circumferential PVDF pressure sensor is the carbon dioxide phase change fracturing combined pressure curve, and the curve collected by the lateral PVDF pressure sensor is the carbon dioxide phase change fracturing gas wedge pressure curve.

[0016] Step 10: Subtract the carbon dioxide phase change cracking gas wedge pressure curve from the carbon dioxide phase change cracking pressure curve measured by each circumferential PVDF pressure sensor to obtain the carbon dioxide phase change cracking impact pressure curve at each circumferential PVDF pressure sensor measuring point, which is the stress wave curve under the corresponding carbon dioxide phase change cracking test pressure.

[0017] Step 11: Adjust the amount of liquid carbon dioxide filling in the pressure vessel before carbon dioxide phase change cracking and the temperature and pressure of the carbon dioxide phase change cracking test in the pressure vessel. Repeatedly perform carbon dioxide phase change cracking pressure decoupling test to obtain carbon dioxide phase change cracking stress wave and gas wedge pressure curves under different initial parameters.

[0018] Step 12: Based on the test conditions and results of carbon dioxide phase change-induced cracking coupled pressure, a multi-input-sequence-output regression model is constructed using an encoder-decoder framework to achieve high-precision prediction of carbon dioxide phase change-induced cracking stress wave and gas wedge pressure curve.

[0019] Preferably, the carbon dioxide phase change pyrolysis coupling pressure testing equipment in step 1 includes a carbon dioxide cylinder, an air compressor, a control console, a gas explosion support, a pressure vessel, a pyrolysis tube, a solenoid valve, a controller, a temperature sensor, a piezoresistive pressure sensor, a temperature monitor, a pressure monitor, and a computer.

[0020] Preferably, the prefabricated rock-like fracture specimen in step 2 is made by mixing quartz sand, gypsum, barite powder, cement, and water.

[0021] Preferably, in step 3, five circumferential PVDF pressure sensors are installed on the inner sidewall of the bottom of the fracture hole. With the axial extension direction of the transverse test hole as the reference direction, the circumferential PVDF pressure sensors are set at angles of 0°, 30°, 45°, 60° and 90° with the reference direction, respectively. Each circumferential PVDF pressure sensor is installed at the same elevation on the axis of the transverse test hole.

[0022] Preferably, the real-time carbon dioxide filling amount in the pressure vessel in step 7 can be calculated using the Span-Wagner equation based on the internal temperature and pressure data of the pressure vessel.

[0023] Preferably, in step 8, the temperature and pressure of the carbon dioxide phase change fracturing test should both be greater than the critical temperature and critical pressure of supercritical carbon dioxide, to ensure that the carbon dioxide in the pressure vessel is in a supercritical state before it is ejected, which is consistent with the carbon dioxide phase change process in the actual engineering process of carbon dioxide phase change fracturing rock breaking.

[0024] Preferably, the sampling frequencies of each circumferential PVDF pressure sensor and the lateral PVDF pressure sensor in step 9 are the same, which facilitates the difference calculation between the carbon dioxide phase change-induced cracking pressure curve and the carbon dioxide phase change-induced cracking gas wedge pressure curve.

[0025] Preferably, the encoder-decoder framework construction process described in step 12 is as follows:

[0026] First, read the input initial fracturing parameters, including the carbon dioxide test filling amount, carbon dioxide phase change fracturing test pressure, carbon dioxide phase change fracturing test temperature, pressure calculation point and the angle between the line connecting the borehole center point and the reference direction.

[0027] Then, the four initial splitting parameters are normalized and input into a fully connected network (including several fully connected layers) to learn the complex relationships between the parameters. In the fully connected network, a batch normalization layer and a ReLU activation function are added after each fully connected layer to accelerate the training process and improve model performance. Subsequently, the number of neurons in the fully connected layers is continuously reduced to extract core features and finally output a low-dimensional vector.

[0028] Next, the low-dimensional vector is copied multiple times to match the length of the output sequence. Then, LSTM and self-attention mechanism are used to gradually generate the stress wave and gas wedge pressure values ​​at each time point. Finally, the complete prediction curves of carbon dioxide phase change-induced cracking stress wave and gas wedge pressure are output, realizing high-fidelity decoupled prediction of carbon dioxide phase change-induced cracking pressure.

[0029] Beneficial effects:

[0030] Compared with existing technologies, this invention provides a decoupled testing and data-driven prediction method for carbon dioxide phase change fracturing pressure. During testing, it effectively utilizes multi-point pressure testing and cleverly sets up transverse test holes to achieve precise measurement of the quasi-static gas wedge pressure during carbon dioxide phase change fracturing. The obtained carbon dioxide phase change fracturing stress wave and gas wedge pressure curves effectively reflect the coupling process between the stress wave and high-pressure gas during carbon dioxide phase change. The two fracturing pressure curves obtained by this method have high fidelity and can be directly used as the load definition basis for numerical simulation of carbon dioxide phase change fracturing rock breaking. Based on the carbon dioxide phase change fracturing stress wave and gas wedge pressure curves obtained from the pressure decoupling test, combined with an encoder-decoder framework, high-precision data-driven prediction of carbon dioxide phase change fracturing stress wave and gas wedge pressure can be achieved. It has the following advantages:

[0031] 1. During carbon dioxide phase change fracturing, two vents are typically symmetrically positioned at the bottom of the fracturing tube. Therefore, the explosive stress waves experienced in different directions on the borehole plane vary depending on the location of the vents. To accurately measure the carbon dioxide phase change fracturing pressure in different directions, this invention employs a series of circumferential PVDF pressure sensors directly arranged on the circumferential sidewall at the bottom of the fracturing borehole to directly monitor the carbon dioxide phase change fracturing pressure at different measuring points along the borehole circumferentially.

[0032] 2. This invention employs multiple circumferential PVDF pressure sensors combined with lateral PVDF pressure sensors arranged on the sidewall of the transverse test hole for simultaneous detection. Since the axial direction of the transverse test hole is perpendicular to the fracturing orifice, the stress wave generated by the supercritical carbon dioxide expansion and explosion during carbon dioxide phase change fracturing does not directly act on the surface of the PVDF pressure sensors. Therefore, the influence of the stress wave on the pressure curve within the transverse test hole is effectively eliminated, ensuring that the pressure curve monitored within the transverse test hole is the carbon dioxide phase change fracturing gas wedge pressure curve. Furthermore, both the fracturing orifice and the transverse test hole in this invention use raw material to fill the gaps, ensuring that unsealed material exits the hole, and that the material is identical. This ensures that the explosion stress wave does not exhibit complex transmission and reflection phenomena at the interface, thus not affecting the gas explosion fracturing effect.

[0033] 3. In this invention, the pressure curves monitored by each circumferential PVDF pressure sensor are directly subtracted from the pressure curves measured by the lateral PVDF pressure sensor to obtain the carbon dioxide phase change fracturing stress wave curves at different circumferential positions of the fracturing hole, thus realizing the decoupled characterization of carbon dioxide phase change fracturing stress wave and gas wedge pressure.

[0034] 4. This invention utilizes an encoder-decoder framework to achieve high-precision prediction of carbon dioxide phase change fracturing stress waves and gas wedge pressures. After constructing the encoder-decoder framework, multiple initial fracturing parameters (carbon dioxide test filling amount, carbon dioxide phase change fracturing test pressure, carbon dioxide phase change fracturing test temperature, the angle between the line connecting the pressure calculation point and the borehole center point and the reference direction) can be input to output complete time history curves of carbon dioxide phase change fracturing stress waves and gas wedge pressures. This enables refined data-driven modeling from multiple inputs to sequential outputs under complex conditions, thereby completing the decoupled prediction of carbon dioxide phase change fracturing pressure curves. Attached Figure Description

[0035] Figure 1 This invention relates to a carbon dioxide phase change-induced pyrolysis coupling pressure testing device.

[0036] Figure 2 Specimen for carbon dioxide phase change-induced cracking coupling pressure test;

[0037] Figure 3 A schematic diagram showing the connection between the cracked hole and the transverse test hole;

[0038] Figure 4 Cross-sectional views of carbon dioxide phase change-induced cracks and transverse test holes;

[0039] The components include: carbon dioxide cylinder 1, air compressor 2, control console 3, gas explosion support 4, pressure vessel 5, fracturing tube 6, fracturing specimen 7, solenoid valve 8, controller 9, injection port 51, temperature sensor 52, piezoresistive pressure sensor 53, temperature monitor 54, pressure monitor 55, computer 56, circumferential PVDF pressure sensor 57, lateral PVDF pressure sensor 58, fracturing hole 71, and test hole 72. Detailed Implementation

[0040] Example 1

[0041] According to such Figure 1 As shown, the carbon dioxide phase change induced cracking coupling pressure testing equipment used in this invention includes a carbon dioxide cylinder 1, an air compressor 2, a control console 3, a gas explosion support 4, a pressure vessel 5, a cracking tube 6, a cracking specimen 7, a solenoid valve 8, a controller 9, a temperature sensor 52, a piezoresistive pressure sensor 53, a temperature monitor 54, a pressure monitor 55, a computer 56, a circumferential PVDF pressure sensor 57, a lateral PVDF pressure sensor 58, and a heating jacket. The pressure vessel 5 is equipped with a liquid injection port 51 and houses the temperature sensor 52 and the piezoresistive pressure sensor 53. The temperature sensor 52 is connected to the temperature monitor 54. The piezoresistive pressure sensor 53, the circumferential PVDF pressure sensor 57, and the lateral PVDF pressure sensor 58 are all connected to the pressure monitor 55. Both the pressure monitor 55 and the temperature monitor 54 are connected to the computer 56. Furthermore, before carbon dioxide phase change induced cracking, a heating jacket is installed around the pressure vessel 5 to heat it.

[0042] Liquid carbon dioxide is stored in carbon dioxide cylinder 1. After being compressed by air compressor 2, the liquid carbon dioxide flows into a storage tank in control console 3. Control console 3 controls the liquid carbon dioxide in the internal storage tank to be injected into pressure vessel 5 through injection port 51. The lower end of pressure vessel 5 is connected to solenoid valve 8, and the lower end of solenoid valve 8 is connected to fracturing tube 6. A fracturing hole 71 is drilled vertically in the center of fracturing specimen 7. The depth of fracturing hole 71 is slightly greater than half the height of fracturing specimen 7 (it can be 8 / 15 or 9 / 17 of the height of fracturing specimen 7). A transverse test hole 72 is also drilled on the vertical plane of fracturing hole 71 to test the gas wedge pressure curve during the carbon dioxide phase change fracturing process. Figure 2 The fracturing hole 71 and the transverse test hole 72 are connected and connected at the bottom sidewall of the fracturing hole 71. During the carbon dioxide phase change fracturing pressure decoupling test, the fracturing tube 6 is fixed in the fracturing hole 71 in the fracturing specimen 7 by sealing material, and the transverse test hole 72 is partially blocked to prevent high-pressure carbon dioxide gas from being directly ejected from the transverse test hole 72.

[0043] The connection methods of the above components are all conventional connections in the field of mechanical equipment. For connection methods, please refer to CN116718490A.

[0044] Example 2

[0045] The following tests were performed using the testing apparatus built in Example 1, and the steps are as follows:

[0046] Step 1: Take a precast rock-like fracture specimen (this specimen is made by mixing quartz sand, PO 42.5R type cement, α high-strength gypsum powder, barite powder, and water in a mass ratio of 100:20:15:20:20 and then preparing it according to conventional methods. Among them, the quartz sand contains 50% 40-70 mesh quartz sand and 50% 70-140 mesh quartz sand). Drill mutually perpendicular fracture holes and transverse test holes in the specimen. The axis of the fracture hole is perpendicular to the plane where its opening is located, and the depth of the fracture hole is slightly greater than half of the height of the fracture specimen (it can be 8 / 15 or 9 / 17 of the height of the fracture specimen). A transverse test hole is also drilled on the vertical plane of the fracture hole, and the axis of the transverse test hole is perpendicular to the plane where its opening is located. The fracture hole and the transverse test hole are connected and connected at the bottom sidewall of the fracture hole.

[0047] Step 2: After installing five circumferential PVDF pressure sensors (on the inner circumferential sidewall of the bottom of the fracture-inducing hole) and five lateral PVDF pressure sensors in the fracture-inducing hole and the transverse test hole respectively, connect the data transmission lines of the circumferential and lateral PVDF pressure sensors to the pressure monitoring instrument. The remaining portion of the transverse test hole is then sealed with raw materials to ensure that no high-pressure carbon dioxide gas is ejected from the transverse test hole during the fracture-inducing process. Five circumferential PVDF pressure sensors are installed on the circumferential sidewall of the bottom of the fracture-inducing hole. With the axial extension direction of the transverse test hole as the reference direction, the circumferential PVDF pressure sensors 57 are set at angles of 0°, 30°, 45°, 60°, and 90° respectively with the reference direction (e.g., ...). Figure 4 As shown, each circumferential PVDF pressure sensor is installed at the same elevation on the axis of the transverse test hole;

[0048] Step 3: Insert the fracturing tube into the fracturing hole, ensuring that the line connecting the two pressure relief ports on the fracturing tube coincides with the axis of the transverse test hole (e.g., Figure 3 (as shown)

[0049] Step 4: Inject the raw material of the rock fracturing specimen into the fracturing hole as a sealing material, and ensure that the sealing material fully fills the drilling gap above each circumferential PVDF pressure sensor. After the sealing material is cured, the fracturing tube and the fracturing specimen form an integral whole, and the end of the transverse test hole near the fracturing hole is in the same confined space as the bottom of the fracturing hole.

[0050] Step 5: Connect the fracturing tube to the solenoid valve and control the solenoid valve to close, so that a sealed space is formed inside the pressure vessel; connect the temperature sensor to the temperature monitor, and the piezoresistive pressure sensor to the pressure monitor, and start real-time monitoring of the temperature and pressure of the carbon dioxide phase change fracturing test inside the pressure vessel.

[0051] Step 6: Open the carbon dioxide cylinder and start the air compressor to force the liquid carbon dioxide in the cylinder into the storage tank inside the control console. Open the injection valve on the pressure vessel, start the control console, and inject the liquid carbon dioxide into the pressure vessel through the injection port. After the carbon dioxide is fully mixed with the air inside the pressure vessel (after 1 minute), disconnect the connection between the control console and the injection valve to allow the liquid carbon dioxide to leak out from the injection valve. Repeat the above operation several times to thoroughly clean the pressure vessel and ensure that the pressure vessel contains only carbon dioxide gas before the fracturing pressure test. Start the control console to inject liquid carbon dioxide into the pressure vessel and monitor the temperature and pressure of the carbon dioxide inside the pressure vessel in real time. Calculate the carbon dioxide density inside the pressure vessel based on the monitoring results. The carbon dioxide density inside the pressure vessel can be calculated using the Span-Wagner equation based on the internal temperature and pressure data of the pressure vessel. The calculation process can be found in the literature Span R, Wagner W. A new equation of state for carbon dioxide covering the fluid region from the triple-point temperature to 1100 K at pressures up to 800 MPa[J]. Journal of physical and chemical reference data, 1996. 25(6): 1509-1596.;The real-time filling amount of carbon dioxide in the pressure vessel can be obtained by multiplying the density of carbon dioxide in the pressure vessel by the volume of the pressure vessel.

[0052] Step 7: After the real-time carbon dioxide filling amount in the pressure vessel reaches the carbon dioxide test filling amount, close the injection port and disconnect the connection between the control console and the pressure vessel; start the heating jacket outside the pressure vessel to heat the pressure vessel, so that the internal temperature and pressure of the pressure vessel continue to rise, and continuously monitor the temperature and pressure inside the pressure vessel; after the temperature and pressure inside the pressure vessel reach the carbon dioxide phase change cracking test temperature and pressure, control the heating jacket to enter the heat preservation mode.

[0053] Step 8: Using the controller, the solenoid valve is opened, and the pressure vessel is instantly connected to the fracturing tube. At this moment, the pressure inside the pressure vessel drops instantly, and the supercritical carbon dioxide is instantly depressurized and expanded, causing carbon dioxide phase change fracturing. The circumferential PVDF pressure sensors and lateral PVDF pressure sensors inside the fracturing specimen record the carbon dioxide phase change fracturing pressure curve inside the specimen. The curve collected by the circumferential PVDF pressure sensor is the carbon dioxide phase change fracturing combined pressure curve, and the curve collected by the lateral PVDF pressure sensor is the carbon dioxide phase change fracturing gas wedge pressure curve. It should be noted that the carbon dioxide phase change fracturing test temperature and pressure should be higher than the critical temperature and critical pressure of supercritical carbon dioxide to ensure that the carbon dioxide in the pressure vessel is in a supercritical state before being ejected, which is consistent with the carbon dioxide phase change process in the actual engineering carbon dioxide phase change fracturing rock breaking process.

[0054] Step 9: Subtract the carbon dioxide phase change cracking gas wedge pressure curve from the carbon dioxide phase change cracking pressure curve measured by each circumferential PVDF pressure sensor to obtain the carbon dioxide phase change cracking impact pressure curve at each circumferential PVDF pressure sensor measuring point, which is the stress wave curve under the corresponding carbon dioxide phase change cracking test pressure; the sampling frequency of each circumferential PVDF pressure sensor and the lateral PVDF pressure sensor is the same, which facilitates the difference calculation between the carbon dioxide phase change cracking pressure curve and the carbon dioxide phase change cracking gas wedge pressure curve.

[0055] Step 10: Adjust the carbon dioxide test filling amount in the pressure vessel before carbon dioxide phase change cracking, as well as the carbon dioxide phase change cracking test temperature and pressure, and repeatedly perform carbon dioxide phase change cracking pressure decoupling test to obtain carbon dioxide phase change cracking stress wave and gas wedge pressure curves under different initial parameters.

[0056] Step 11: Based on the carbon dioxide phase change induced fracture coupling pressure test conditions and results, a multi-input sequence-output regression model is constructed using an encoder-decoder framework to achieve high-precision prediction of the carbon dioxide phase change induced fracture stress wave and gas wedge pressure curve. Specifically, the initial fracture parameters can be read, including the carbon dioxide test filling amount, carbon dioxide phase change induced fracture test pressure, carbon dioxide phase change induced fracture test temperature, and the angle between the line connecting the pressure calculation point and the borehole center point and the reference direction. The four initial fracture parameters are then Z-score normalized. Next, the normalized four initial fracture parameters are input into a fully connected network consisting of three fully connected layers. The number of neurons in the first, second, and third fully connected layers are 128, 64, and 32, respectively. A BatchNorm (batch normalization) and ReLU activation function are added after each fully connected layer to learn the complex relationships between the input parameters. During this process, the number of neurons decreases continuously from the first to the third fully connected layer, enabling core feature extraction. The fully connected network finally outputs a 16-dimensional feature vector. Subsequently, the 16-dimensional feature vector is expanded into a 64-dimensional vector, and this 64-dimensional vector is copied N times (N being the total number of sampling points for the stress wave and gas wedge pressure curves to be predicted). The stress wave and gas wedge pressure curves are then initialized as two pressure curve vectors [0,0,0,0,…,0] and [0,0,0,0,…,0] (both vectors also have dimensions equal to N). Next, the N 64-dimensional vectors are sequentially mapped one-to-one with the N zeros in the two pressure curve vectors. Then, LSTM and a self-attention mechanism are used for iterative prediction, progressively forecasting the shock wave pressure and gas wedge pressure values ​​at each moment. Throughout this process, the predicted shock wave pressure and gas wedge pressure values ​​continuously replace the zeros at corresponding moments, ultimately outputting a complete predicted curve for carbon dioxide phase change-induced fracturing stress wave and gas wedge pressure, achieving high-fidelity decoupled prediction of carbon dioxide phase change-induced fracturing pressure.

Claims

1. A method for decoupling testing and data-driven prediction of carbon dioxide phase change-induced cracking pressure, characterized in that, Includes the following steps: Step 1: Construct a carbon dioxide phase change-induced pyrolysis coupling pressure testing device; Step 2: Take a precast rock fracture specimen and drill mutually perpendicular fracture holes and transverse test holes in the specimen. The axis of the fracture hole is perpendicular to the plane where its opening is located, and the axis of the transverse test hole is perpendicular to the plane where its opening is located. The fracture hole and the transverse test hole are connected and connected at the bottom sidewall of the fracture hole. Step 3: After installing several circumferential PVDF pressure sensors and lateral PVDF pressure sensors on the inner circumferential sidewall at the bottom of the fracturing hole and in the transverse test hole, connect the data transmission lines of the circumferential PVDF pressure sensors and lateral PVDF pressure sensors to the pressure monitoring instrument. The empty part of the transverse test hole is then sealed with raw materials to ensure that no high-pressure carbon dioxide gas is ejected from the transverse test hole during the fracturing process, and each circumferential PVDF pressure sensor is installed at the same elevation on the axis of the transverse test hole. Step 4: Insert the fracturing tube into the fracturing hole, and make sure that the line connecting the two pressure relief ports on the fracturing tube coincides with the axis of the transverse test hole. Step 5: Inject the raw material of the rock fracturing specimen into the fracturing hole, and ensure that the sealing material fully fills the borehole gap above each circumferential PVDF pressure sensor. After the sealing material is cured, the fracturing tube and the fracturing specimen form an integral whole, and the end of the transverse test hole near the fracturing hole is in the same confined space as the bottom of the fracturing hole. Step 6: Connect the fracturing tube to the solenoid valve and control the solenoid valve to close, so that a sealed space is formed inside the pressure vessel; connect the temperature sensor to the temperature monitor, and the piezoresistive pressure sensor to the pressure monitor, and start real-time monitoring of the temperature and pressure of the carbon dioxide phase change fracturing test inside the pressure vessel. Step 7: Open the carbon dioxide cylinder and start the air compressor to force the liquid carbon dioxide in the cylinder into the storage tank inside the control console; open the injection valve on the pressure vessel, start the control console, and inject the liquid carbon dioxide into the pressure vessel. After the carbon dioxide is fully mixed with the air inside the pressure vessel, disconnect the connection between the control console and the injection valve to allow the liquid carbon dioxide to leak out from the injection valve. Repeat the above operation several times to thoroughly clean the pressure vessel and ensure that the pressure vessel contains only carbon dioxide gas before the fracturing pressure test; start the control console to inject liquid carbon dioxide into the pressure vessel and monitor the temperature and pressure of the carbon dioxide inside the pressure vessel in real time, and calculate the real-time filling amount of carbon dioxide in the pressure vessel based on the monitoring results. Step 8: After the real-time carbon dioxide filling amount in the pressure vessel reaches the carbon dioxide test filling amount, close the injection port and disconnect the connection between the control console and the pressure vessel; start the heating jacket outside the pressure vessel to heat the pressure vessel, so that the internal temperature and pressure of the pressure vessel continue to rise, and continuously monitor the temperature and pressure inside the pressure vessel; after the temperature and pressure inside the pressure vessel reach the carbon dioxide phase change cracking test temperature and pressure, control the heating jacket to enter the heat preservation mode. Step 9: Using the controller, the solenoid valve is opened, and the pressure vessel is instantly connected to the fracturing tube. At this time, the pressure inside the pressure vessel drops instantly, and the supercritical carbon dioxide is instantly depressurized and expanded, causing carbon dioxide phase change fracturing. The circumferential PVDF pressure sensors and lateral PVDF pressure sensors inside the fracturing specimen record the carbon dioxide phase change fracturing pressure curve inside the specimen. The curve collected by the circumferential PVDF pressure sensor is the carbon dioxide phase change fracturing combined pressure curve, and the curve collected by the lateral PVDF pressure sensor is the carbon dioxide phase change fracturing gas wedge pressure curve. Step 10: Subtract the carbon dioxide phase change cracking gas wedge pressure curve from the carbon dioxide phase change cracking pressure curve measured by each circumferential PVDF pressure sensor to obtain the carbon dioxide phase change cracking impact pressure curve at each circumferential PVDF pressure sensor measuring point, which is the stress wave curve under the corresponding carbon dioxide phase change cracking test pressure. Step 11: Adjust the amount of liquid carbon dioxide filling in the pressure vessel before carbon dioxide phase change cracking and the temperature and pressure of the carbon dioxide phase change cracking test in the pressure vessel. Repeatedly perform carbon dioxide phase change cracking pressure decoupling test to obtain carbon dioxide phase change cracking stress wave and gas wedge pressure curves under different initial parameters. Step 12: Based on the test conditions and results of carbon dioxide phase change-induced cracking coupled pressure, a multi-input-sequence-output regression model is constructed using an encoder-decoder framework to achieve high-precision prediction of carbon dioxide phase change-induced cracking stress wave and gas wedge pressure curve.

2. The method for decoupling testing and data-driven prediction of carbon dioxide phase change-induced cracking pressure according to claim 1, characterized in that, The carbon dioxide phase change pyrolysis coupling pressure testing equipment described in step 1 includes a carbon dioxide cylinder, an air compressor, a control console, a gas explosion support, a pressure vessel, a pyrolysis tube, a solenoid valve, a controller, a temperature sensor, a piezoresistive pressure sensor, a temperature monitor, a pressure monitor, and a computer.

3. The method for decoupling testing and data-driven prediction of carbon dioxide phase change-induced cracking pressure according to claim 1, characterized in that, The precast rock-like fracture specimens described in step 2 are made by mixing quartz sand, gypsum, barite powder, cement, and water.

4. The method for decoupling testing and data-driven prediction of carbon dioxide phase change-induced cracking pressure according to claim 1, characterized in that, In step 3, five circumferential PVDF pressure sensors are installed on the inner sidewall of the bottom of the fracture hole. With the axial extension direction of the transverse test hole as the reference direction, the circumferential PVDF pressure sensors are set at angles of 0°, 30°, 45°, 60° and 90° with the reference direction, respectively. Each circumferential PVDF pressure sensor is installed at the same elevation on the axis of the transverse test hole.

5. The method for decoupling testing and data-driven prediction of carbon dioxide phase change-induced cracking pressure according to claim 1, characterized in that, In step 7, the real-time carbon dioxide filling amount in the pressure vessel can be calculated using the Span-Wagner equation based on the internal temperature and pressure data of the pressure vessel.

6. The method for decoupling testing and data-driven prediction of carbon dioxide phase change-induced cracking pressure according to claim 1, characterized in that, In step 8, the temperature and pressure of the carbon dioxide phase change fracturing test should both be greater than the critical temperature and critical pressure of supercritical carbon dioxide to ensure that the carbon dioxide in the pressure vessel is in a supercritical state before it is ejected, which is consistent with the carbon dioxide phase change process in the actual engineering process of carbon dioxide phase change fracturing rock breaking.

7. The method for decoupling testing and data-driven prediction of carbon dioxide phase change-induced cracking pressure according to claim 1, characterized in that, In step 9, the sampling frequencies of the circumferential PVDF pressure sensors and the lateral PVDF pressure sensors are the same, which facilitates the difference calculation between the carbon dioxide phase change-induced cracking pressure curve and the carbon dioxide phase change-induced cracking gas wedge pressure curve.

8. The method for decoupling testing and data-driven prediction of carbon dioxide phase change-induced cracking pressure according to claim 1, characterized in that, The encoder-decoder framework construction process described in step 12 is as follows: First, read the input initial fracturing parameters, including the carbon dioxide test filling amount, carbon dioxide phase change fracturing test pressure, carbon dioxide phase change fracturing test temperature, pressure calculation point and the angle between the line connecting the borehole center point and the reference direction. Then, the four initial splitting parameters are normalized and input into a fully connected network to learn the complex relationships between the parameters. A batch normalization layer and a ReLU activation function are then added after the fully connected layer to accelerate the training process and improve model performance. Subsequently, the number of neurons in the fully connected layer is gradually reduced to extract core features and finally output a low-dimensional vector. Next, the low-dimensional vector is copied multiple times to match the length of the output sequence. Then, LSTM and self-attention mechanism are used to gradually generate the stress wave and gas wedge pressure values ​​at each time point. Finally, the complete prediction curves of carbon dioxide phase change-induced cracking stress wave and gas wedge pressure are output, realizing high-fidelity decoupled prediction of carbon dioxide phase change-induced cracking pressure.

Images