Real-time prediction method, device, equipment and medium for surrounding rock rupture of gas storage reservoir

Abstract

The present application provides a real-time prediction method, device, equipment and medium for the fracture of the surrounding rock of a gas storage reservoir. The method includes: acquiring acoustic emission signals and stress-strain parameters generated by a specified type of rock in the surrounding rock of the gas storage reservoir during the fracture process; screening out target acoustic emission signals meeting the training conditions based on the acoustic emission characteristic parameters corresponding to the acoustic emission signals; determining the rock state stage label of the target acoustic emission signals based on the stress-strain curves of the stress-strain parameters corresponding to the target acoustic emission signals; constructing a sample data set based on the acoustic emission characteristic parameters of the target acoustic emission signals and the rock state stage labels of the target acoustic emission signals; training a rock fracture prediction model based on the sample data set; and processing the to-be-detected acoustic emission signals of the target rock collected in real time by using the rock fracture prediction model to obtain the rock state stage, where the to-be-detected acoustic emission signals are the signals of the target rock around the reinforced structure of the gas storage reservoir collected by a piezoelectric ceramic sensor.

Description

Technical Field

[0001] This application relates to the technical field of rock fracture prediction, and particularly to a real-time prediction method, device, equipment and medium for surrounding rock fracture of a gas storage reservoir. Background Art

[0002] Compressed air energy storage power station is a new type of energy storage power station, which uses the excess electric energy during the low-capacity load of the power system to compress air and store it in underground caves, and releases it when needed, generating electricity through a generator set to meet the needs of peak loads. Underground caves usually include natural salt caverns and artificial underground energy storage chambers, etc. However, the siting of natural salt cavern gas storage reservoirs is not easy due to the limitation of salt cavern resource distribution. The gas storage devices of large-scale compressed air energy storage power stations generally require a large gas storage capacity. Artificial excavation of underground gas storage reservoirs is considered a promotable gas storage reservoir type. For the artificial excavation and construction of underground gas storage reservoirs in hard rock, the current mainstream design idea is that the sealing layer only ensures airtightness, and the high-pressure gas acts on the inner wall of the gas storage reservoir and is transmitted through the lining, and the surrounding rock mainly bears the high internal pressure. In the scenario of underground gas storage reservoirs, rock fracture prediction technology is crucial.

[0003] Rock fracture prediction technology is not only related to the safety and economy of gas storage reservoirs, but also can prevent geological disasters caused by rock fractures. Traditional rock fracture monitoring methods mainly rely on direct observation and deformation monitoring, but these methods have problems such as poor real-time performance, insufficient precursors, and low success rate. Summary of the Invention

[0004] Embodiments of this application provide a real-time prediction method, device, equipment and medium for surrounding rock fracture of a gas storage reservoir to solve the problems of poor real-time performance, insufficient precursors and low success rate existing in traditional rock fracture monitoring methods.

[0005] To solve the above technical problems, the embodiments of this application are implemented as follows:

[0006] In a first aspect, embodiments of this application provide a real-time prediction method for surrounding rock fracture of a gas storage reservoir, and the method includes:

[0007] Obtain the acoustic emission signals generated by a specified type of rock in the surrounding rock of the gas storage reservoir during the fracture process and the stress-strain parameters of the rock. The acoustic emission signals are signals collected by piezoelectric ceramic sensors arranged on the surface of the rock during the process of continuously applying external force to the rock, and the stress-strain parameters are parameters monitored by displacement monitoring sensors arranged on the side of the rock during the process of continuously applying external force to the rock;

[0008] Based on the acoustic emission characteristic parameters corresponding to the acoustic emission signals, screen out the target acoustic emission signals that meet the training conditions from the acoustic emission signals;

[0009] Determine the rock state stage label corresponding to the target acoustic emission signal based on the stress-strain curve of the stress-strain parameters corresponding to the target acoustic emission signal;

[0010] Construct a sample data set based on the acoustic emission characteristic parameters of the target acoustic emission signal and the rock state stage label corresponding to the target acoustic emission signal;

[0011] Train a rock fracture prediction model for predicting the rock state stage based on the sample data set;

[0012] Process the acoustic emission signal to be measured of the target rock of the specified type in the surrounding rock of the gas storage reservoir collected in real time by using the rock fracture prediction model to obtain the rock state stage corresponding to the target rock, and the acoustic emission signal to be measured is the signal of the target rock around the reinforced structure of the gas storage reservoir collected by a piezoelectric ceramic sensor.

[0013] Optionally, the acoustic emission characteristic parameters include: acoustic emission hit number, acoustic emission energy, acoustic emission RA value, acoustic emission b value, signal duration, ring count and acoustic emission rate.

[0014] Optionally, the screening of the target acoustic emission signal that meets the training conditions from the acoustic emission signal based on the acoustic emission characteristic parameters corresponding to the acoustic emission signal includes:

[0015] Eliminate the acoustic emission signals in the acoustic emission signal with a signal duration less than the time threshold and / or a ring count less than the count threshold to obtain the target acoustic emission signal.

[0016] Optionally, the determination of the rock state stage label corresponding to the target acoustic emission signal based on the stress-strain curve of the stress-strain parameters corresponding to the target acoustic emission signal includes:

[0017] Determine the stress-strain curve corresponding to the target acoustic emission signal based on the stress-strain parameters corresponding to the target acoustic emission signal;

[0018] Determine the rock elastic modulus of the rock based on the slope of the stress-strain curve;

[0019] Determine the rock state stage label corresponding to the target acoustic emission signal based on the rock elastic modulus, the instantaneous secant modulus corresponding to the stress-strain curve, the strain parameter of the rock and the peak strain parameter.

[0020] Optionally, the determination of the rock elastic modulus of the rock based on the slope of the stress-strain curve includes:

[0021] Obtain the target stress-strain curve with a linear relationship on the stress-strain curve;

[0022] Based on the stress parameters and strain parameters corresponding to the starting point and the ending point of the target stress-strain curve respectively, the elastic modulus of the rock is calculated.

[0023] Optionally, determining the rock state stage label corresponding to the target acoustic emission signal based on the elastic modulus of the rock, the instantaneous secant modulus corresponding to the stress-strain curve, the strain parameter of the rock, and the peak strain parameter includes:

[0024] Based on the stress parameters and strain parameters of each point on the stress-strain curve, the instantaneous secant modulus corresponding to each point is calculated;

[0025] When the ratio of the instantaneous secant modulus corresponding to the first point on the stress-strain curve to the elastic modulus of the rock is greater than a first threshold, and the strain parameter of the first point is less than the peak strain parameter, it is determined that the rock state stage label of the target acoustic emission signal corresponding to the first point is the elastic deformation stage label, and the first threshold is a value greater than 0 and less than 1;

[0026] When the ratio of the instantaneous secant modulus corresponding to the second point on the stress-strain curve to the elastic modulus of the rock is greater than 0, and the strain parameter of the second point is less than a set multiple of the peak strain parameter, it is determined that the rock state stage label of the target acoustic emission signal corresponding to the second point is the compaction stage label, and the set multiple is a value greater than 0 and less than 1;

[0027] When the ratio of the instantaneous secant modulus corresponding to the third point on the stress-strain curve to the elastic modulus of the rock is less than 0, it is determined that the rock state stage label of the target acoustic emission signal corresponding to the third point is the post-peak failure stage label;

[0028] The rock state stage labels of the remaining acoustic emission signals in the target acoustic emission signal are determined as the plastic stage labels, and the remaining acoustic emission signals refer to the acoustic emission signals in the target acoustic emission signal that are not the elastic deformation stage label, the compaction stage label, and the post-peak failure stage label.

[0029] Optionally, training the rock fracture prediction model for predicting the rock state stage based on the sample data set includes:

[0030] Input the acoustic emission characteristic parameters of the target acoustic emission signal in the sample data set into the rock fracture prediction model to be trained, and the rock fracture prediction model to be trained processes the acoustic emission characteristic parameters of the target acoustic emission signal;

[0031] Obtain the predicted rock state stage corresponding to the target acoustic emission signal output by the rock fracture prediction model to be trained;

[0032] Based on the rock state stage label and the predicted rock state stage, the loss value of the to-be-trained rock fracture prediction model is calculated.

[0033] When the loss value is within a preset range, the trained to-be-trained rock fracture prediction model is used as the final rock fracture prediction model.

[0034] In a second aspect, an embodiment of the present application provides a real-time prediction device for surrounding rock fracture of a gas storage reservoir. The device includes:

[0035] An acoustic emission signal acquisition module, configured to acquire acoustic emission signals generated during the fracture of a specified type of rock in the surrounding rock of the gas storage reservoir and stress-strain parameters of the rock. The acoustic emission signals are signals collected by piezoelectric ceramic sensors arranged on the surface of the rock during the continuous application of external force to the rock, and the stress-strain parameters are parameters monitored by displacement monitoring sensors arranged on the side of the rock during the continuous application of external force to the rock.

[0036] A target signal screening module, configured to screen out target acoustic emission signals meeting the training conditions from the acoustic emission signals based on the acoustic emission characteristic parameters corresponding to the acoustic emission signals.

[0037] A state stage label determination module, configured to determine the rock state stage label corresponding to the target acoustic emission signal based on the stress-strain curve of the stress-strain parameters corresponding to the target acoustic emission signal.

[0038] A sample data set construction module, configured to construct a sample data set based on the acoustic emission characteristic parameters of the target acoustic emission signal and the rock state stage label corresponding to the target acoustic emission signal.

[0039] A rock fracture prediction model training module, configured to train a rock fracture prediction model for predicting the rock state stage based on the sample data set.

[0040] A rock state real-time prediction module, configured to process the to-be-tested acoustic emission signals of the target rock in the surrounding rock of the gas storage reservoir collected in real time by using the rock fracture prediction model to obtain the rock state stage corresponding to the target rock. The to-be-tested acoustic emission signals are signals of the target rock around the reinforced structure of the gas storage reservoir collected by piezoelectric ceramic sensors.

[0041] Optionally, the acoustic emission characteristic parameters include: acoustic emission hit number, acoustic emission energy, acoustic emission RA value, acoustic emission b value, signal duration, ring count, and acoustic emission rate.

[0042] Optionally, the target signal screening module includes:

[0043] A target signal acquisition unit, configured to eliminate acoustic emission signals in the acoustic emission signals whose signal duration is less than a time threshold and / or whose ring count is less than a count threshold, so as to obtain the target acoustic emission signals.

[0044] Optionally, the state stage label determination module includes:

[0045] A stress-strain curve determination unit, configured to determine a stress-strain curve corresponding to the target acoustic emission signal based on the stress-strain parameters corresponding to the target acoustic emission signal;

[0046] A rock elastic modulus determination unit, configured to determine the rock elastic modulus of the rock based on the slope of the stress-strain curve;

[0047] A state stage label determination unit, configured to determine a rock state stage label corresponding to the target acoustic emission signal based on the rock elastic modulus, the instantaneous secant modulus corresponding to the stress-strain curve, the strain parameter of the rock, and the peak strain parameter.

[0048] Optionally, the rock elastic modulus determination unit includes:

[0049] A target curve acquisition subunit, configured to acquire a target stress-strain curve having a linear relationship on the stress-strain curve;

[0050] A rock elastic modulus calculation subunit, configured to calculate the rock elastic modulus of the rock based on the stress parameters and strain parameters corresponding to the starting point and the ending point of the target stress-strain curve respectively.

[0051] Optionally, the state stage label determination unit includes:

[0052] An instantaneous secant modulus calculation subunit, configured to calculate an instantaneous secant modulus corresponding to each point based on the stress parameter and strain parameter of each point on the stress-strain curve;

[0053] An elastic stage label determination subunit, configured to determine that the rock state stage label of the target acoustic emission signal corresponding to the first point is an elastic deformation stage label when the ratio of the instantaneous secant modulus corresponding to the first point on the stress-strain curve to the rock elastic modulus is greater than a first threshold and the strain parameter of the first point is less than the peak strain parameter, where the first threshold is a value greater than 0 and less than 1;

[0054] A compaction stage label determination subunit, configured to determine that the rock state stage label of the target acoustic emission signal corresponding to the second point on the stress-strain curve is a compaction stage label when the ratio of the instantaneous secant modulus corresponding to the second point to the rock elastic modulus is greater than 0, and the strain parameter of the second point is less than a set multiple of the peak strain parameter, where the set multiple is a value greater than 0 and less than 1;

[0055] A post-peak stage label determination subunit, configured to determine that the rock state stage label of the target acoustic emission signal corresponding to the third point on the stress-strain curve is a post-peak failure stage label when the ratio of the instantaneous secant modulus corresponding to the third point to the rock elastic modulus is less than 0;

[0056] A plastic stage label determination subunit, configured to determine that the rock state stage labels of the remaining acoustic emission signals in the target acoustic emission signal are plastic stage labels, where the remaining acoustic emission signals refer to the acoustic emission signals in the target acoustic emission signal that are not the elastic deformation stage label, the compaction stage label, and the post-peak failure stage label.

[0057] Optionally, the rock fracture prediction model training module includes:

[0058] An acoustic emission feature parameter input unit, configured to input the acoustic emission feature parameters of the target acoustic emission signal in the sample data set into the rock fracture prediction model to be trained, and the rock fracture prediction model to be trained processes the acoustic emission feature parameters of the target acoustic emission signal;

[0059] A rock state stage prediction unit, configured to obtain the predicted rock state stage corresponding to the target acoustic emission signal output by the rock fracture prediction model to be trained;

[0060] A loss value calculation unit, configured to calculate the loss value of the rock fracture prediction model to be trained based on the rock state stage label and the predicted rock state stage;

[0061] A rock fracture prediction model acquisition unit, configured to use the trained rock fracture prediction model to be trained as the final rock fracture prediction model when the loss value is within a preset range.

[0062] In a third aspect, an embodiment of the present application provides an electronic device, including:

[0063] A memory, a processor, and a computer program stored on the memory and executable on the processor, where when the computer program is executed by the processor, it implements the real-time prediction method for the rupture of the surrounding rock of the gas storage reservoir described in any one of the above.

[0064] Fourthly, an embodiment of the present application provides a readable storage medium. When the instructions in the storage medium are executed by a processor of an electronic device, the electronic device can execute the real-time prediction method for the rupture of the surrounding rock of the gas storage reservoir described in any one of the above.

[0065] In the embodiment of the present application, by acquiring the acoustic emission signals generated by a specified type of rock in the process of rupture in the surrounding rock of the gas storage reservoir and the stress-strain parameters of the rock, the acoustic emission signals are the signals collected by piezoelectric ceramic sensors arranged on the surface of the rock during the continuous application of external force to the rock, and the stress-strain parameters are the parameters monitored by displacement monitoring sensors arranged on the side of the rock during the continuous application of external force to the rock. Based on the acoustic emission characteristic parameters corresponding to the acoustic emission signals, the target acoustic emission signals meeting the training conditions are screened out from the acoustic emission signals. Based on the stress-strain curve of the stress-strain parameters corresponding to the target acoustic emission signals, the rock state stage label corresponding to the target acoustic emission signals is determined. Based on the acoustic emission characteristic parameters of the target acoustic emission signals and the rock state stage label corresponding to the target acoustic emission signals, a sample data set is constructed. Based on the sample data set, a rock rupture prediction model for predicting the rock state stage is trained. The rock rupture prediction model is used to process the to-be-detected acoustic emission signals of the target rock of the specified type in the surrounding rock of the gas storage reservoir collected in real time, and the rock state stage corresponding to the target rock is obtained. The to-be-detected acoustic emission signals are the signals of the target rock around the stiffening structure of the gas storage reservoir collected by the piezoelectric ceramic sensors. In the embodiment of the present application, the acoustic emission characteristic parameters at different stages of rock rupture are monitored by acoustic emission technology to train the rock rupture prediction model, so that the trained model can predict the stage of rock rupture, improve the real-time monitoring ability of the rock rupture process, can quickly capture and process micro-rupture signals, realize timely warning, and improve the timeliness of rock rupture monitoring. At the same time, the use of the model prediction method can greatly improve the accuracy of acoustic emission signal feature extraction and state stage classification prediction, and can more accurately identify the precursor signals and rupture modes of rock rupture.

[0066] The above description is only an overview of the technical solution of the present application. In order to be able to understand the technical means of the present application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features and advantages of the present application more obvious and understandable, the specific embodiments of the present application are given below. Description of the Drawings

[0067] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required to be used in the description of the embodiments of the present application will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0068] Figure 1 It is a flowchart of the steps of a real-time prediction method for the rupture of the surrounding rock of a gas storage reservoir provided by an embodiment of the present application;

[0069] Figure 2 It is a schematic diagram of a label generation process provided by an embodiment of the present application;

[0070] Figure 3 It is a schematic diagram of a model training process provided by an embodiment of the present application;

[0071] Figure 4 It is a schematic structural diagram of a real-time prediction device for the rupture of the surrounding rock of a gas storage reservoir provided by an embodiment of the present application;

[0072] Figure 5 It is a schematic structural diagram of an electronic device provided by an embodiment of the present application. Specific embodiments

[0073] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0074] In this embodiment, a compressed air energy storage power station (CAES, Compressed Air Energy Storage) is a power storage technology that uses compressed air to store energy. It mainly compresses air to a high-pressure state by using an electric motor and stores it in an underground cave during the low-demand period of the power grid. When the power grid demand is high, the stored compressed air is released, and a generator is driven by a turbine or expander to generate electric energy, thereby realizing the storage and release of electric energy. Underground caves usually include natural salt caves and artificial underground energy storage chambers, etc. However, the siting of natural salt cave gas storage reservoirs is not easy due to the limitation of salt cave resource distribution. The gas storage devices of large-scale compressed air energy storage power stations generally require a large gas storage capacity, and artificial excavation of underground gas storage reservoirs is considered a promotable gas storage reservoir type. For the artificial excavation and construction of underground gas storage reservoirs in hard rock, the current mainstream design idea is that the sealing layer only ensures airtightness, and the high-pressure gas acts on the inner wall of the gas storage reservoir and is transmitted through the lining, and the surrounding rock mainly bears the high internal pressure.

[0075] However, underground rock is a material containing defects such as joints, microcracks, and pores. Under the action of external loads and environmental factors, the expansion of internal cracks in the rock will be induced, ultimately leading to rock failure. Rock failure may trigger geological disasters such as karst collapse and land subsidence. Rock failure may cause impacts or squeezes on the gas storage cavern, resulting in damage or failure of the gas storage cavern, greatly reducing the operational stability of the gas storage cavern.

[0076] Based on the above technical problems, the embodiments of the present application are proposed. For details, refer to the description of the corresponding method steps in the embodiments of the present application below.

[0077] Refer to Figure 1 , which shows the step flow chart of a real-time prediction method for the rupture of the surrounding rock of a gas storage cavern provided by the embodiment of the present application. As Figure 1 shown, the real-time prediction method for the rupture of the surrounding rock of the gas storage cavern may include: Step 101, Step 102, Step 103, Step 104, Step 105, and Step 106.

[0078] Step 101: Obtain the acoustic emission signals generated during the rupture of the rock of a specified type in the surrounding rock of the gas storage cavern and the stress-strain parameters of the rock. The acoustic emission signals are the signals collected by piezoelectric ceramic sensors arranged on the surface of the rock during the continuous application of external forces to the rock, and the stress-strain parameters are the parameters monitored by displacement monitoring sensors arranged on the side of the rock during the continuous application of external forces to the rock.

[0079] The embodiments of the present application can be applied to the scenario of training a rock rupture prediction model through the acoustic emission signal parameters in different rock state stages.

[0080] The surrounding rock of a gas storage cavern refers to the rock mass surrounding the gas storage cavern in an underground gas storage project. These rock masses will be affected by various factors such as excavation, gas pressure, and groundwater during the construction and operation of the gas storage cavern, resulting in changes in the stress state.

[0081] The specified type refers to the type of rock selected when obtaining the characteristic parameters of the acoustic emission signals during rock rupture. In this example, the rock of the specified type may be, but is not limited to, granite, etc.

[0082] In this example, the rock of the specified type may be one type of rock or multiple types of rocks, and this embodiment does not limit this. It can be understood that the types of rocks forming the surrounding rock of the gas storage cavern may be one or multiple, and specifically, it can be determined according to the actual situation, and this embodiment does not limit this.

[0083] When training a model for predicting the rock state of a specified type of rock during the training phase, it is possible to obtain the acoustic emission signals generated by the specified type of rock in the surrounding rock of the gas storage reservoir during the fracture process and the stress-strain parameters of the rock. Among them, the acoustic emission signal refers to the transient elastic wave (i.e., the acoustic emission signal) generated by the rock during the fracture process. The stress-strain parameters include stress parameters and strain parameters. Among them, the stress parameter refers to the force per unit area, which is used to describe the ability of the rock to resist deformation inside. The stress parameter refers to the ratio of the amount of deformation generated by the rock after being stressed to the original size, which is used to describe the degree of rock deformation.

[0084] In this embodiment, the acoustic emission signal can be the signal collected by a piezoelectric ceramic sensor arranged on the rock surface during the process of continuously applying an external force to the rock, and the stress-strain parameter is the parameter monitored by a displacement monitoring sensor arranged on the side of the rock during the process of continuously applying an external force to the rock.

[0085] In this embodiment, the acquisition of acoustic emission signals and stress-strain parameters in different rock state stages is combined with a specific experiment. During the experiment, granite can be selected as the sample. During the experiment, 2 piezoelectric ceramic sensors are arranged on the surface of the sample (the size and resonance frequency of the sensor can be determined according to the experimental requirements). The acoustic emission characteristic parameters during the damage and failure process of the rock can be received through the piezoelectric ceramic sensor. In addition, a linear variable differential transformer and a circumferential displacement sensor are arranged on both sides of the rock to monitor the axial strain and radial strain of the rock during the deformation process.

[0086] The experiment can be carried out under a uniaxial compression device and a multi-channel acoustic emission waveform acquisition system. This system mainly includes three parts: a loading system, a stress-strain monitoring system, and an acoustic emission monitoring system. The loading system uses a servo hydraulic controller for uniaxial loading. The waveform signal received by the piezoelectric ceramic sensor is amplified by an amplifier with 40 dB and then transmitted to the monitoring device, and all microcrack signals in the entire loading process can be recorded. During the loading process, acoustic emission signals and stress-strain monitoring are carried out simultaneously to study the mesoscopic fracture law of the rock.

[0087] After obtaining the acoustic emission signals generated by the specified type of rock in the surrounding rock of the gas storage reservoir during the fracture process and the stress-strain parameters of the rock, step 102 is executed.

[0088] Step 102: Based on the acoustic emission characteristic parameters corresponding to the acoustic emission signals, screen out the target acoustic emission signals that meet the training conditions from the acoustic emission signals.

[0089] Acoustic emission characteristic parameters refer to the characteristic parameters that are extracted from the acoustic emission waveform and described in a series of numerical forms. In this example, the acoustic emission characteristic parameters may include: the number of acoustic emission hits, acoustic emission energy, acoustic emission RA value, acoustic emission b value, signal duration, ring count, acoustic emission rate, etc.

[0090] Among them, the number of acoustic emission hits refers to the number of signals that pass through the threshold value and cause a certain channel to acquire data, and can be used to reflect the number of microcracks generated during the rock failure process.

[0091] Acoustic emission energy refers to the energy of an acoustic emission event, which is proportional to the square of the amplitude of the waveform where the observed acoustic emission event is located, and can be used to reflect the strength of the acoustic emission event. In the specific implementation, a single local change in the rock that generates acoustic emission is called an acoustic emission event, which can be divided into total count and count rate.

[0092] The acoustic emission RA value refers to the ratio of the rise time to the amplitude of an acoustic emission hit. The magnitude of the acoustic emission RA value reflects the nature of the crack. A higher RA value corresponds to a tensile crack, and a lower RA value corresponds to a shear crack. The formula for the acoustic emission RA value is shown in formula (1) below:

[0093] (1)

[0094] In the above formula (1), is the rise time of the acoustic emission event, is the amplitude of the acoustic emission event.

[0095] In acoustic emission monitoring, the number of acoustic emission hits is proportional to the number of internal crack propagations in the rock, while the AE amplitude (the maximum voltage peak of the acoustic emission signal waveform) is proportional to the incremental length of crack propagation in the rock. Therefore, based on the Gutenberg-Richter relationship, the brittle deformation and crack propagation mechanism of the rock can be inferred:

[0096] (2)

[0097] In the above formula (2), represents the number of acoustic emission hits with a magnitude ≥ , is a constant, represents the acoustic emission b value. And in acoustic emission monitoring, the magnitude of the seismic source can be represented by the AE amplitude, , is the monitored acoustic emission amplitude during the experiment, is the seismic magnitude.

[0098] As a parameter introduced from seismology, the acoustic emission b-value can quantitatively analyze the relationship between the number and size of cracks inside rocks. Generally speaking, the larger the b-value, the more small fractures there are at this moment. On the contrary, it indicates that there are mainly large fractures at this moment.

[0099] That is, the acoustic emission b-value during the experiment can be obtained by using the least squares method in the above formula (2):

[0100] (3)

[0101] In the above formula (2), is the acoustic emission b-value, is the total number of acoustic emission impacts in the whole experiment, is the median of the th magnitude, is the number of acoustic emission impacts of the

[0102] In this embodiment, in order to analyze the change of the acoustic emission b-value at different times, the sliding sampling method (also called the sliding window method) can be adopted. The calculation window can be set to 1000, and the sliding step can be set to 1 / 4 of the calculation window, so as to slide a fixed-size window on the data sequence and perform calculations or analyses at each window position.

[0103] The signal duration refers to the time interval from when the signal first crosses the threshold to when it finally drops below the threshold.

[0104] The ring count refers to the number of times the waveform crosses the threshold (a value set artificially). In this embodiment, the ring count is used to indicate the number of times the waveform corresponding to the acoustic emission signal is greater than the set threshold.

[0105] The acoustic emission rate refers to the frequency or number of acoustic emission events of the target acoustic emission signal per unit time. The calculation of the acoustic emission rate is usually based on the number of acoustic emission events accumulated within a certain time interval. For example, within one hour of measurement time, a total of 1000 acoustic emission events are detected, then the acoustic emission rate during this time period is 1000 times / hour.

[0106] The calculation formula of the acoustic emission rate is as follows:

[0107] Introduce the event interval time function , the function reflects the average frequency of acoustic emission impacts occurring in the time window of N acoustic emission impacts. In order to derive the time function , the time interval between events can be defined as:

[0108] (4)

[0109] In the formula, represents the time interval between two adjacent events, represents the moment when the -th impact event occurs, while represents the moment when the -th impact event occurs. Then, the average value of the time taken for N consecutive impacts is defined as the event interval time

[0110] at this moment, as shown in the following formula:

[0111] For the first time interval , the above formula can be modified as follows:

[0112] (6)

[0113] Within the given time window between and , the definition of the acoustic emission rate can be the reciprocal of :

[0114] (7)

[0115] After acquiring the acoustic emission signals collected during the fracture process of a specified type of rock, based on the acoustic emission characteristic parameters corresponding to the acoustic emission signals, the target acoustic emission signals meeting the training conditions can be screened out from the acoustic emission signals.

[0116] In this embodiment, the signal duration and ring count can be combined for screening the target acoustic emission signals. Specifically, the acoustic emission signals with a signal duration less than the time threshold and / or a ring count less than the count threshold in the acoustic emission signals can be excluded, and the remaining acoustic emission signals are used as the target acoustic emission signals. It can be understood that the specific values of the time threshold and the count threshold can be determined according to service requirements, and this embodiment does not limit them.

[0117] After screening out the target acoustic emission signals meeting the training conditions from the acoustic emission signals based on the acoustic emission characteristic parameters corresponding to the acoustic emission signals, step 103 is executed.

[0118] Step 103: Determine the rock state stage label corresponding to the target acoustic emission signal based on the stress-strain curve of the stress-strain parameters corresponding to the target acoustic emission signal.

[0119] Through experiments, the rock state stage can be divided into four stages, namely: elastic deformation stage, compaction stage, post-peak failure stage, and plastic stage. Among them, the elastic deformation stage refers to the stage when the rock undergoes elastic deformation. The compaction stage refers to the stage when the original cracks inside the rock gradually close. The post-peak failure stage refers to the stage when the internal structure of the rock is damaged. The plastic stage refers to the stage when the deformation characteristics of the rock change significantly.

[0120] After screening out the target acoustic emission signals, based on the stress-strain curve of the stress-strain parameters corresponding to the target acoustic emission signals, the rock state stage label corresponding to the target acoustic emission signals can be determined. The implementation process can be described in detail in combination with the following specific implementation manners.

[0121] In a specific implementation of the present application, the above step 103 may include:

[0122] Sub-step A1: Based on the stress-strain parameters corresponding to the target acoustic emission signals, determine the stress-strain curve corresponding to the target acoustic emission signals.

[0123] In this embodiment, based on the stress-strain parameters corresponding to the target acoustic emission signals, the stress-strain curve corresponding to the target acoustic emission signals can be determined.

[0124] After determining the stress-strain curve corresponding to the target acoustic emission signals based on the stress-strain parameters corresponding to the target acoustic emission signals, execute sub-step A2.

[0125] Sub-step A2: Based on the slope of the stress-strain curve, determine the rock elastic modulus of the rock.

[0126] After determining the stress-strain curve corresponding to the target acoustic emission signals based on the stress-strain parameters corresponding to the target acoustic emission signals, the rock elastic modulus of the rock can be determined based on the slope of the stress-strain curve. In a specific implementation, the target stress-strain curve showing a linear relationship on the stress-strain curve can be obtained. And based on the stress parameters and strain parameters corresponding to the starting point and the ending point of the target stress-strain curve respectively, the rock elastic modulus of the rock can be calculated.

[0127] Sub-step A3: Based on the rock elastic modulus, the instantaneous secant modulus corresponding to the stress-strain curve, the strain parameter of the rock, and the peak strain parameter, determine the rock state stage label corresponding to the target acoustic emission signals.

[0128] After calculating the rock elastic modulus of the rock, based on the rock elastic modulus, the instantaneous secant modulus corresponding to the stress-strain curve, the strain parameter of the rock, and the peak strain parameter, the rock state stage label corresponding to the target acoustic emission signals can be determined. The implementation process can be described in detail in combination with the following specific implementation manners.

[0129] In a specific implementation of the present application, the above sub-step A3 may include:

[0130] Sub-step B1: Based on the stress parameters and strain parameters of each point on the stress-strain curve, calculate the instantaneous secant modulus corresponding to each point.

[0131] In this embodiment, after obtaining the stress-strain curve corresponding to the target acoustic emission signal, the instantaneous secant modulus corresponding to each point can be calculated based on the stress parameters and strain parameters of each point on the stress-strain curve.

[0132] In a specific implementation, if you want to calculate the instantaneous secant modulus corresponding to each point on the stress-strain curve, it is necessary to calculate the differences in stress and strain between each point (except the starting point) on the curve and its previous point. Since it is calculating the instantaneous secant modulus of each point, it is actually calculating the secant slope between adjacent two points.

[0133] Sub-step B2: When the ratio of the instantaneous secant modulus corresponding to the first point on the stress-strain curve to the rock elastic modulus is greater than the first threshold, and the strain parameter of the first point is less than the peak strain parameter, determine that the rock state stage label of the target acoustic emission signal corresponding to the first point is the elastic deformation stage label, where the first threshold is a value greater than 0 and less than 1.

[0134] When the ratio of the instantaneous secant modulus corresponding to the first point on the stress-strain curve to the rock elastic modulus is greater than the first threshold, and the strain parameter of the first point is less than the peak strain parameter, then it can be determined that the rock state stage label of the target acoustic emission signal corresponding to the first point is the elastic deformation stage label, where the first threshold is a value greater than 0 and less than 1.

[0135] Sub-step B3: When the ratio of the instantaneous secant modulus corresponding to the second point on the stress-strain curve to the rock elastic modulus is greater than 0, and the strain parameter of the second point is less than a set multiple of the peak strain parameter, determine that the rock state stage label of the target acoustic emission signal corresponding to the second point is the compaction stage label, where the set multiple is a value greater than 0 and less than 1.

[0136] When the ratio of the instantaneous secant modulus corresponding to the second point on the stress-strain curve to the rock elastic modulus is greater than 0, and the strain parameter of the second point is less than a set multiple of the peak strain parameter, then it can be determined that the rock state stage label of the target acoustic emission signal corresponding to the second point is the compaction stage label, where the set multiple is a value greater than 0 and less than 1.

[0137] Sub-step B4: When the ratio of the instantaneous secant modulus corresponding to the third point on the stress-strain curve to the elastic modulus of the rock is less than 0, determine that the rock state stage label of the target acoustic emission signal corresponding to the third point is the post-peak failure stage label.

[0138] When the ratio of the instantaneous secant modulus corresponding to the third point on the stress-strain curve to the elastic modulus of the rock is less than 0, it can be determined that the rock state stage label of the target acoustic emission signal corresponding to the third point is the post-peak failure stage label.

[0139] Sub-step B5: Determine that the rock state stage label of the remaining acoustic emission signals in the target acoustic emission signal is the plastic stage label, where the remaining acoustic emission signals refer to the acoustic emission signals in the target acoustic emission signal that are not the elastic deformation stage label, the compaction stage label, and the post-peak failure stage label.

[0140] At the same time, it can be determined that the rock state stage label of the remaining acoustic emission signals in the target acoustic emission signal is the plastic stage label, where the remaining acoustic emission signals refer to the acoustic emission signals in the target acoustic emission signal that are not the elastic deformation stage label, the compaction stage label, and the post-peak failure stage label. That is, the labels of the other acoustic emission signals in the target acoustic emission signal except the above three labels are set as the plastic stage labels.

[0141] For the above implementation process, it can be combined with Figure 2 for the following illustrative examples.

[0142] The classification label is the target of supervised machine learning. The label is usually determined by the measured data or the classification criteria defined by the data. In the embodiment, the uniaxial compression process can be divided into four stages by the instantaneous secant modulus and the axial strain value. As Figure 2 shown is the method of generating four types of labels. Figure 2 in is the elastic modulus of the rock while is the peak strain.

[0143] When setting the label for the target acoustic emission signal, the instantaneous secant modulus and the axial strain can be input.

[0144] In and , then set the label of the corresponding target acoustic emission signal as label 2, that is, the elastic stage label. Among them, 0.95 is the first threshold in this embodiment.

[0145] In and When it is, the label of the corresponding target acoustic emission signal is set to label 1, that is, the label in the compaction stage. Among them, 0.6 is the set multiple in this embodiment.

[0146] At When it is, the label of the corresponding target acoustic emission signal is set to label 4, that is, the label in the post-peak stage.

[0147] At the same time, the label of the target acoustic emission signal that does not meet the above three conditions is set to label 3, that is, the label in the plastic stage.

[0148] It can be understood that Figure 2 only describes the label setting process in combination with the axial strain. In practical applications, the label can also be set in combination with the radial strain. Specifically, the label setting conditions can be set according to the actual situation, and this embodiment does not limit this.

[0149] After determining the rock state stage label corresponding to the target acoustic emission signal based on the stress-strain curve of the stress-strain parameters corresponding to the target acoustic emission signal, step 104 is executed.

[0150] Step 104: Construct a sample data set based on the acoustic emission characteristic parameters of the target acoustic emission signal and the rock state stage label corresponding to the target acoustic emission signal.

[0151] Furthermore, a sample data set can be constructed based on the acoustic emission characteristic parameters of the target acoustic emission signal and the rock state stage label corresponding to the target acoustic emission signal, that is, the acoustic emission characteristic parameters of the target acoustic emission signal and its corresponding rock state stage label are used as the sample data for model training, and numerous sample data constitute the sample data set.

[0152] After constructing the sample data set based on the acoustic emission characteristic parameters of the target acoustic emission signal and the rock state stage label corresponding to the target acoustic emission signal, step 105 is executed.

[0153] Step 105: Train a rock fracture prediction model for predicting the rock state stage based on the sample data set.

[0154] After constructing the sample data set based on the acoustic emission characteristic parameters of the target acoustic emission signal and the rock state stage label corresponding to the target acoustic emission signal, a rock fracture prediction model for predicting the rock state stage can be trained based on the sample data set, so as to perform real-time prediction on the rock state stage of the surrounding rock of the gas storage reservoir according to the rock fracture prediction model.

[0155] In this embodiment, the rock fracture prediction model to be trained can be selected as XGBoost (eXtreme Gradient Boosting). XGBoost is a boosting method that uses a set of decision trees to estimate a logical function to determine the expected class of the input data. The boosting model uses an iterative method to construct multiple estimators, using the residuals of the previous estimators as data to generate new predictions. The output of the logical classifier of XGBoost is a linear combination of the weighted average of the output probabilities of all trees. Similar to random forests, XGBoost uses random subsamples of columns and rows for each tree to reduce bias, and the maximum depth of the tree is specified by the user.

[0156] In classification problems, the imbalance of data samples is a very common problem, and imbalanced data samples have a great impact on the training and testing of the model. During the experiment, the acoustic emission signal value generated by the rock in the plastic stage is 44070, while in the elastic stage it is only 11364, which means that in order to maintain high accuracy, the model will only predict the plastic stage. To balance this, in this embodiment, the SMOTE (Synthetic Minority Over-sampling Technique) method is combined with a machine learning algorithm to optimize the generalization of the model. The SMOTE method is an improved scheme for the random over-sampling algorithm. Since the random over-sampling adopts the strategy of simply replicating samples to increase the minority class samples, this is likely to cause the problem of model overfitting, that is, the information learned by the model is too specific and not general enough. The basic idea of the SMOTE algorithm is to analyze the minority class samples and artificially synthesize new samples according to the minority class samples and add them to the data set. Its algorithm process is as follows

[0157] (1) For each sample in the class with a smaller number , calculate its distance to all samples in the minority class sample set using the Euclidean distance as the standard to obtain its k nearest neighbors.

[0158] (2) Set a sampling ratio according to the sample imbalance ratio to determine the sampling magnification N. For each minority class sample , randomly select several samples from its k nearest neighbors. Suppose the selected nearest neighbors are .

[0159] (3) For each randomly selected nearest neighbor , construct new samples with the original sample respectively according to the following formula.

[0160] (8)

[0161] In the above formula, is the generated new sample, is the th original sample, For the neighboring samples of the th randomly selected original sample, is a random number between 0 and 1.

[0162] The model training process can be described in detail in combination with the following specific implementation manners.

[0163] In a specific implementation of the present application, the above step 105 may include:

[0164] Sub-step C1: Input the acoustic emission characteristic parameters of the target acoustic emission signal in the sample data set into the rock fracture prediction model to be trained, and the rock fracture prediction model to be trained processes the acoustic emission characteristic parameters of the target acoustic emission signal.

[0165] In this embodiment, after obtaining the sample data set, the acoustic emission characteristic parameters of the target acoustic emission signal in the sample data set can be input into the rock fracture prediction model to be trained, and the rock fracture prediction model to be trained processes the acoustic emission characteristic parameters of the target acoustic emission signal. That is, the rock fracture prediction model to be trained learns the acoustic emission characteristic parameters of the target acoustic emission signal.

[0166] Sub-step C2: Obtain the predicted rock state stage corresponding to the target acoustic emission signal output by the rock fracture prediction model to be trained.

[0167] After the rock fracture prediction model to be trained processes the acoustic emission characteristic parameters of the target acoustic emission signal, the predicted rock state stage corresponding to the target acoustic emission signal output by the rock fracture prediction model to be trained can be obtained.

[0168] Sub-step C3: Calculate the loss value of the rock fracture prediction model to be trained based on the rock state stage label and the predicted rock state stage.

[0169] Furthermore, the loss value of the rock fracture prediction model to be trained can be calculated based on the rock state stage label and the predicted rock state stage. In this embodiment, the loss value can be a cross-entropy loss value, a mean squared error loss value, or a categorical cross-entropy loss value, etc. Specifically, the specific type of the loss value can be determined according to business requirements, and this embodiment does not limit this.

[0170] After calculating the loss value of the rock fracture prediction model to be trained based on the rock state stage label and the predicted rock state stage, execute sub-step C4.

[0171] Sub-step C4: When the loss value is within a preset range, use the trained rock fracture prediction model to be trained as the final rock fracture prediction model.

[0172] After obtaining the loss value of the rock fracture prediction model to be trained based on the rock state stage label and the predicted rock state stage, it is possible to determine whether the loss value is within a preset range.

[0173] If the loss value is not within the preset range, it indicates that the rock fracture prediction model to be trained has not converged. At this time, more sample data sets can be combined to continue training the rock fracture prediction model to be trained until the model converges.

[0174] If the loss value is within the preset range, it indicates that the rock fracture prediction model to be trained has converged. At this time, the trained rock fracture prediction model to be trained can be used as the final rock fracture prediction model.

[0175] For the model training process, it can be combined with Figure 3 described as follows. As Figure 3 shown, the model training process may include:

[0176] 1. Monitor acoustic emission signals and stress-strain values.

[0177] 2. Data preprocessing: Screen data through the acoustic emission characteristic parameters of the acoustic emission signal (such as duration, ring count) to construct a data set for model training.

[0178] 3. Label generation: Determine the classification label of the acoustic emission signal through the stress-strain value, that is, draw a stress-strain curve with the stress-strain value, and use the stress-strain curve to determine the classification label, so as to generate a label reflecting the rock state.

[0179] 4. Combine the data set and the label reflecting the rock state to train a model for predicting the rock state stage.

[0180] 5. Model evaluation: Compare the predicted label and the observed label to determine the model accuracy and feature importance. Specifically, the trained model can be used to evaluate the importance of each feature.

[0181] This embodiment focuses on the relationship between mesoscopic microcracks of rocks and acoustic emission signals, uses acoustic emission technology to analyze the evolution process of mesoscopic microcracks inside rocks and the accumulation of damage, and uses machine learning technology to classify and predict different stages of rocks by combining multiple acoustic emission signals. The classification model for the stages of rocks established based on mesoscopic acoustic emission signals can accurately predict different stages of granite, and the accuracy can reach 78.6%, which is higher than the prediction accuracy using X-ray technology before.

[0182] After completing the training of the rock fracture prediction model, step 106 is executed.

[0183] Step 106: Process the to-be-detected acoustic emission signals of the target rock of a specified type in the surrounding rock of the gas storage reservoir collected in real time by using the rock fracture prediction model, to obtain the rock state stage corresponding to the target rock, where the to-be-detected acoustic emission signals are the signals of the target rock around the reinforced structure of the gas storage reservoir collected by a piezoelectric ceramic sensor.

[0184] In this embodiment, when predicting the rock state stage of the target rock of a specified type in the surrounding rock of the underground gas storage reservoir, the to-be-detected acoustic emission signals of the target rock of a specified type in the surrounding rock of the gas storage reservoir can be collected. In a specific implementation, the to-be-detected acoustic emission signals of the target rock around the reinforced structure of the gas storage reservoir can be collected by a piezoelectric ceramic sensor.

[0185] In a specific implementation, the installation of the reinforced structure of the surrounding rock of the gas storage reservoir is a complex and critical engineering process, which is crucial for ensuring the stability and safety of the gas storage reservoir. The reinforced structure of the surrounding rock of the gas storage reservoir can be, but is not limited to, rock bolts. The method of installing the reinforced structure on the surface of the surrounding rock (taking rock bolts as an example) can be: first drill holes in the surrounding rock, and then insert and fix the rock bolts. Install the sensor at a pre-designed fixed position, and use adhesives, bolts or other fixing devices to fix the sensor on the surface of the surrounding rock.

[0186] In this example, the piezoelectric ceramic sensor can be pre-installed at the surrounding position of the reinforced structure of the gas storage reservoir to collect the acoustic emission signals of the rock around the reinforced structure of the gas storage reservoir. The method of installing the piezoelectric ceramic sensor at the surrounding position of the reinforced structure of the gas storage reservoir can be: 1. Epoxy resin glue bonding: Select an epoxy resin glue with corrosion resistance, avoid using substances that may cause corrosive reactions, ensure that the glue is evenly and thinly applied to improve the bonding quality, and during the bonding process, cure it at room temperature or in a slightly heated environment, ensuring that the curing time is greater than 30 minutes. 2. Mechanical clamping fixation: For piezoelectric ceramic sensors with certain specific shapes, mechanical clamping can be used for fixation. Such as the pre-installed reinforced structure in the gas storage reservoir, etc. Taking rock bolts as an example, a fixed clamp or bracket can be used to firmly connect with the reinforced structure, and then the sensor can be installed on the clamp or bracket to achieve the fixation of the piezoelectric ceramic sensor through the reinforced structure of the gas storage reservoir, avoiding the situation that the sensor falls off and cannot collect the acoustic emission signals of the surrounding rock of the gas storage reservoir.

[0187] It can be understood that the above examples of the installation methods of the reinforced structure and the piezoelectric ceramic sensor are only examples listed for better understanding the technical solutions of the embodiments of the present application. In practical applications, any method that can install the reinforced structure and the piezoelectric ceramic sensor on the periphery of the surrounding rock of the gas storage reservoir can be applied to this embodiment, and this embodiment does not limit this.

[0188] After the acoustic emission signals to be measured of the target rock of the specified type in the surrounding rock of the gas storage reservoir are collected, the acoustic emission characteristic parameters of the acoustic emission signals to be measured can be input into the rock fracture prediction model, so that the rock fracture prediction model processes the acoustic emission characteristic parameters of the acoustic emission signals to be measured. Furthermore, the rock state stage corresponding to the target rock output by the rock fracture prediction model can be obtained. It can be understood that after the rock fracture prediction model is trained, it is only necessary to arrange piezoelectric ceramic sensors outside the target rock of the specified type in the surrounding rock of the gas storage reservoir to monitor the acoustic emission signals to be measured, and identify the acoustic emission characteristic parameters of the acoustic emission signals to be measured and input them into the model to predict the rock state stage of the target rock. This process does not require stress-strain parameters.

[0189] In the embodiment of the present application, by optimizing the acquisition and analysis system of acoustic emission signals, the real-time monitoring ability of the rock fracture process is significantly improved. It can quickly capture and process micro-fracture signals, realize timely early warning, and effectively prevent the occurrence of geological disasters. Moreover, in the embodiment of the present application, by collecting the acoustic emission signals of rock fractures at the position of the reinforced structure, the reinforcement effect of the reinforced structure on the surrounding rock can be monitored in real time. If the reinforced structure can effectively disperse and bear the stress inside the surrounding rock, the acoustic emission signals of rock fractures at this position may be relatively weak or less. The acoustic emission signals of rock fractures at the position of the reinforced structure can also be used as an important indicator for warning potential damage. When the signals show abnormal changes, it may mean that the reinforced structure is under excessive stress or damage, so as to remind relevant personnel to take timely measures for repair or reinforcement.

[0190] Of course, in specific implementation, piezoelectric ceramic sensors can also be installed at the non-reinforced structure positions of the surrounding rock of the gas storage reservoir to collect the acoustic emission signals to be measured of the rocks at the non-reinforced structure positions for predicting the rock fracture state stage.

[0191] The real-time prediction method for the rupture of the surrounding rock of a gas storage reservoir provided by the embodiment of the present application obtains the acoustic emission signals generated by a specified type of rock in the surrounding rock of the gas storage reservoir during the rupture process and the stress-strain parameters of the rock. The acoustic emission signals are the signals collected by piezoelectric ceramic sensors arranged on the surface of the rock during the continuous application of external force to the rock, and the stress-strain parameters are the parameters monitored by displacement monitoring sensors arranged on the side of the rock during the continuous application of external force to the rock. Based on the acoustic emission characteristic parameters corresponding to the acoustic emission signals, target acoustic emission signals meeting the training conditions are selected from the acoustic emission signals. Based on the stress-strain curve of the stress-strain parameters corresponding to the target acoustic emission signals, the rock state stage label corresponding to the target acoustic emission signals is determined. Based on the acoustic emission characteristic parameters of the target acoustic emission signals and the rock state stage labels corresponding to the target acoustic emission signals, a sample data set is constructed. A rock rupture prediction model for predicting the rock state stage is trained based on the sample data set, and the rock rupture prediction model is used to process the to-be-detected acoustic emission signals of the target rock of the specified type in the surrounding rock of the gas storage reservoir collected in real time, so as to obtain the rock state stage corresponding to the target rock. The to-be-detected acoustic emission signals are the signals of the target rock around the reinforced structure of the gas storage reservoir collected by piezoelectric ceramic sensors. The embodiment of the present application monitors the acoustic emission characteristic parameters at different stages of rock rupture through acoustic emission technology for training the rock rupture prediction model, so that the trained model can predict the stage of rock rupture, improves the real-time monitoring ability of the rock rupture process, can quickly capture and process micro-rupture signals, realizes timely early warning, and improves the timeliness of rock rupture monitoring. At the same time, the use of the model prediction method can greatly improve the accuracy of acoustic emission signal feature extraction and state stage classification prediction, and can more accurately identify the precursor signals and rupture modes of rock rupture.

[0192] Referring to Figure 4 , a schematic structural diagram of a real-time prediction device for the rupture of the surrounding rock of a gas storage reservoir provided by the embodiment of the present application is shown. As Figure 4 shown, the real-time prediction device 400 for the rupture of the surrounding rock of the gas storage reservoir may include the following modules:

[0193] An acoustic emission signal acquisition module 410, configured to obtain the acoustic emission signals generated by a specified type of rock in the surrounding rock of the gas storage reservoir during the rupture process and the stress-strain parameters of the rock. The acoustic emission signals are the signals collected by piezoelectric ceramic sensors arranged on the surface of the rock during the continuous application of external force to the rock, and the stress-strain parameters are the parameters monitored by displacement monitoring sensors arranged on the side of the rock during the continuous application of external force to the rock;

[0194] A target signal screening module 420, configured to select target acoustic emission signals meeting the training conditions from the acoustic emission signals based on the acoustic emission characteristic parameters corresponding to the acoustic emission signals;

[0195] A state phase label determination module 430, configured to determine a rock state phase label corresponding to the target acoustic emission signal based on a stress-strain curve of stress-strain parameters corresponding to the target acoustic emission signal;

[0196] A sample data set construction module 440, configured to construct a sample data set based on acoustic emission feature parameters of the target acoustic emission signal and the rock state phase label corresponding to the target acoustic emission signal;

[0197] A rock fracture prediction model training module 450, configured to train a rock fracture prediction model for predicting a rock state phase based on the sample data set;

[0198] A rock state real-time prediction module 460, configured to process an acoustic emission signal to be measured of a target rock of a specified type in the surrounding rock of a gas storage reservoir collected in real time by using the rock fracture prediction model, so as to obtain a rock state phase corresponding to the target rock, where the acoustic emission signal to be measured is a signal of the target rock around a reinforced structure of the gas storage reservoir collected by a piezoelectric ceramic sensor.

[0199] Optionally, the acoustic emission feature parameters include: acoustic emission hit number, acoustic emission energy, acoustic emission RA value, acoustic emission b value, signal duration, ring count, and acoustic emission rate.

[0200] Optionally, the target signal screening module includes:

[0201] A target signal acquisition unit, configured to remove acoustic emission signals in the acoustic emission signals, where a signal duration is less than a time threshold and / or a ring count is less than a count threshold, so as to obtain the target acoustic emission signal.

[0202] Optionally, the state phase label determination module includes:

[0203] A stress-strain curve determination unit, configured to determine a stress-strain curve corresponding to the target acoustic emission signal based on the stress-strain parameters corresponding to the target acoustic emission signal;

[0204] A rock elastic modulus determination unit, configured to determine a rock elastic modulus of the rock based on a slope of the stress-strain curve;

[0205] A state phase label determination unit, configured to determine a rock state phase label corresponding to the target acoustic emission signal based on the rock elastic modulus, an instantaneous secant modulus corresponding to the stress-strain curve, a strain parameter of the rock, and a peak strain parameter.

[0206] Optionally, the rock elastic modulus determination unit includes:

[0207] A target curve acquisition subunit, configured to acquire a target stress-strain curve that is linearly related on the stress-strain curve;

[0208] A rock elastic modulus calculation subunit, configured to calculate the rock elastic modulus of the rock based on the stress parameters and strain parameters corresponding to the starting point and the ending point of the target stress-strain curve respectively.

[0209] Optionally, the state stage label determination unit includes:

[0210] An instantaneous secant modulus calculation subunit, configured to calculate the instantaneous secant modulus corresponding to each point based on the stress parameters and strain parameters of each point on the stress-strain curve;

[0211] An elastic stage label determination subunit, configured to determine that the rock state stage label of the target acoustic emission signal corresponding to the first point on the stress-strain curve is an elastic deformation stage label when the ratio of the instantaneous secant modulus corresponding to the first point on the stress-strain curve to the rock elastic modulus is greater than a first threshold and the strain parameter of the first point is less than the peak strain parameter, where the first threshold is a value greater than 0 and less than 1;

[0212] A compaction stage label determination subunit, configured to determine that the rock state stage label of the target acoustic emission signal corresponding to the second point on the stress-strain curve is a compaction stage label when the ratio of the instantaneous secant modulus corresponding to the second point on the stress-strain curve to the rock elastic modulus is greater than 0 and the strain parameter of the second point is less than a set multiple of the peak strain parameter, where the set multiple is a value greater than 0 and less than 1;

[0213] A post-peak stage label determination subunit, configured to determine that the rock state stage label of the target acoustic emission signal corresponding to the third point on the stress-strain curve is a post-peak failure stage label when the ratio of the instantaneous secant modulus corresponding to the third point on the stress-strain curve to the rock elastic modulus is less than 0;

[0214] A plastic stage label determination subunit, configured to determine that the rock state stage labels of the remaining acoustic emission signals in the target acoustic emission signal are plastic stage labels, where the remaining acoustic emission signals refer to the acoustic emission signals in the target acoustic emission signal that are not the elastic deformation stage label, the compaction stage label, and the post-peak failure stage label.

[0215] Optionally, the rock fracture prediction model training module includes:

[0216] An acoustic emission feature parameter input unit, configured to input the acoustic emission feature parameters of the target acoustic emission signal in the sample dataset into a rock fracture prediction model to be trained, and the rock fracture prediction model to be trained processes the acoustic emission feature parameters of the target acoustic emission signal;

[0217] A rock state stage prediction unit, configured to obtain a predicted rock state stage corresponding to the target acoustic emission signal output by the to-be-trained rock fracture prediction model;

[0218] A loss value calculation unit, configured to calculate a loss value of the to-be-trained rock fracture prediction model based on the rock state stage label and the predicted rock state stage;

[0219] A rock fracture prediction model acquisition unit, configured to use the trained to-be-trained rock fracture prediction model as the final rock fracture prediction model when the loss value is within a preset range.

[0220] The real-time prediction device for surrounding rock fracture of a gas storage reservoir provided by the embodiments of the present application obtains acoustic emission signals generated during the fracture process of a specified type of rock in the surrounding rock of the gas storage reservoir and stress-strain parameters of the rock. The acoustic emission signals are signals collected by piezoelectric ceramic sensors arranged on the surface of the rock during the continuous application of external force to the rock, and the stress-strain parameters are parameters monitored by displacement monitoring sensors arranged on the side of the rock during the continuous application of external force to the rock. Based on the acoustic emission characteristic parameters corresponding to the acoustic emission signals, target acoustic emission signals meeting the training conditions are screened out from the acoustic emission signals. Based on the stress-strain curve of the stress-strain parameters corresponding to the target acoustic emission signals, a rock state stage label corresponding to the target acoustic emission signals is determined. Based on the acoustic emission characteristic parameters of the target acoustic emission signals and the rock state stage label corresponding to the target acoustic emission signals, a sample data set is constructed. A rock fracture prediction model for predicting the rock state stage is trained based on the sample data set, and the to-be-measured acoustic emission signals of the target rock of the specified type in the surrounding rock of the gas storage reservoir collected in real time are processed by using the rock fracture prediction model to obtain the rock state stage corresponding to the target rock. The to-be-measured acoustic emission signals are signals of the target rock around the reinforced structure of the gas storage reservoir collected by piezoelectric ceramic sensors. The embodiments of the present application monitor the acoustic emission characteristic parameters at different stages of rock fracture through acoustic emission technology to train the rock fracture prediction model, so that the trained model can predict the stage of rock fracture, improve the real-time monitoring ability of the rock fracture process, can quickly capture and process micro-fracture signals, realize timely early warning, and improve the timeliness of rock fracture monitoring. At the same time, the use of the model prediction method can greatly improve the accuracy of acoustic emission signal feature extraction and state stage classification prediction, and can more accurately identify the precursor signals and fracture modes of rock fracture.

[0221] Additionally, an embodiment of the present application further provides an electronic device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, where when the computer program is executed by the processor, it implements the above-mentioned real-time prediction method for the rupture of the surrounding rock of the gas storage reservoir.

[0222] Figure 5 FIG. shows a schematic structural diagram of an electronic device 500 according to an embodiment of the present invention. As Figure 5 shown, the electronic device 500 includes a central processing unit (CPU) 501, which can execute various appropriate actions and processes according to computer program instructions stored in a read-only memory (ROM) 502 or computer program instructions loaded from a storage unit 508 into a random access memory (RAM) 503. In the RAM 503, various programs and data required for the operation of the electronic device 500 can also be stored. The CPU 501, ROM 502, and RAM 503 are connected to each other through a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0223] Multiple components in the electronic device 500 are connected to the I / O interface 505, including: an input unit 506, such as a keyboard, a mouse, a microphone, etc.; an output unit 507, such as various types of displays, speakers, etc.; a storage unit 508, such as a magnetic disk, an optical disc, etc.; and a communication unit 509, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 509 allows the electronic device 500 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0224] Each of the processes and processes described above can be executed by the processing unit 501. For example, the method of any of the above embodiments can be implemented as a computer software program, which is tangibly included in a computer-readable medium, such as the storage unit 508. In some embodiments, part or all of the computer program can be loaded and / or installed onto the electronic device 500 via the ROM 502 and / or the communication unit 509. When the computer program is loaded into the RAM 503 and executed by the CPU 501, one or more actions of the method described above can be executed.

[0225] The embodiment of the present application also provides a computer-readable storage medium. A computer program is stored on the computer-readable storage medium. When the computer program is executed by a processor, it implements each process of the above embodiment of the real-time prediction method for the rupture of the surrounding rock of the gas storage reservoir and can achieve the same technical effects. To avoid repetition, it will not be elaborated here. Among them, the computer-readable storage medium is, for example, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disc, etc.

[0226] It should be noted that in this text, the term "including", "comprising" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "including one..." does not exclude the existence of another identical element in the process, method, article or device including that element.

[0227] The above is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of changes or substitutions, which should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A real-time prediction method for surrounding rock fracture of a gas storage reservoir, characterized in that: The method comprises: Acquiring acoustic emission signals generated by a specified type of rock in the surrounding rock of the gas storage reservoir during the fracture process and stress and strain parameters of the rock, wherein the acoustic emission signals are signals collected by a piezoelectric ceramic sensor arranged on the surface of the rock during the process of continuously applying external force to the rock, and the stress and strain parameters are parameters monitored by a displacement monitoring sensor arranged on the side of the rock during the process of continuously applying external force to the rock; Based on the acoustic emission characteristic parameters corresponding to the acoustic emission signals, a target acoustic emission signal meeting the training conditions is screened out from the acoustic emission signals; Determine a rock state stage label corresponding to the target acoustic emission signal based on a stress-strain curve of stress-strain parameters corresponding to the target acoustic emission signal; Constructing a sample data set based on the acoustic emission characteristic parameters of the target acoustic emission signal and the rock state stage label corresponding to the target acoustic emission signal; A rock fracture prediction model for predicting the rock state stage is obtained by training based on the sample data set; The rock fracture prediction model is used to process the acoustic emission signals to be measured of the target rock of the specified type in the surrounding rock of the gas storage reservoir collected in real time to obtain the rock state stage corresponding to the target rock, wherein the acoustic emission signals to be measured are signals of the target rock around the reinforced structure of the gas storage reservoir collected by the piezoelectric ceramic sensor; The step of determining the rock state stage label corresponding to the target acoustic emission signal based on the stress-strain curve of the stress-strain parameter corresponding to the target acoustic emission signal comprises: Determining a stress-strain curve corresponding to the target acoustic emission signal based on the stress-strain parameter corresponding to the target acoustic emission signal; determining a rock elastic modulus of the rock based on the slope of the stress-strain curve; Determine a rock state stage label corresponding to the target acoustic emission signal based on the rock elastic modulus, the instantaneous secant modulus corresponding to the stress-strain curve, the strain parameter and the peak strain parameter of the rock; The step of determining the rock state stage label corresponding to the target acoustic emission signal based on the rock elastic modulus, the instantaneous secant modulus corresponding to the stress-strain curve, the strain parameter and the peak strain parameter of the rock comprises: Based on the stress parameter and strain parameter of each point on the stress-strain curve, the instantaneous secant modulus corresponding to each point is calculated; When the ratio of the instantaneous secant modulus to the rock elastic modulus corresponding to the first point on the stress-strain curve is greater than a first threshold value, and the strain parameter of the first point is less than the peak strain parameter, it is determined that the rock state stage label of the target acoustic emission signal corresponding to the first point is an elastic deformation stage label, and the first threshold value is a value greater than 0 and less than 1; When the ratio of the instantaneous secant modulus to the rock elastic modulus corresponding to the second point on the stress-strain curve is greater than 0, and the strain parameter of the second point is less than a set multiple of the peak strain parameter, it is determined that the rock state stage label of the target acoustic emission signal corresponding to the second point is a compaction stage label, and the set multiple is a value greater than 0 and less than 1; When the ratio of the instantaneous secant modulus corresponding to the third point on the stress-strain curve to the rock elastic modulus is less than 0, determining that the rock state stage label of the target acoustic emission signal corresponding to the third point is a post-peak failure stage label; The rock state stage labels of the remaining acoustic emission signals in the target acoustic emission signal are determined as plastic stage labels, and the remaining acoustic emission signals refer to the acoustic emission signals in the target acoustic emission signal that are not the elastic deformation stage label, the compaction stage label, and the post-peak failure stage label.

2. The method according to claim 1, characterized in that The acoustic emission characteristic parameters include: acoustic emission impact number, acoustic emission energy, acoustic emission RA value, acoustic emission b value, signal duration, ringing count and acoustic emission rate.

3. The method according to claim 2, characterized in that The step of selecting a target acoustic emission signal that meets a training condition from the acoustic emission signal based on the acoustic emission characteristic parameters corresponding to the acoustic emission signal comprises: The target acoustic emission signal is obtained by eliminating the acoustic emission signals whose signal duration is less than a time threshold and / or whose ring count is less than a count threshold.

4. The method according to claim 1, characterized in that: Determining the rock elastic modulus of the rock based on the slope of the stress-strain curve includes: Obtaining a target stress-strain curve having a linear relationship on the stress-strain curve; The rock elastic modulus of the rock is calculated based on the stress parameters and strain parameters corresponding to the starting point and the end point of the target stress-strain curve, respectively.

5. The method according to claim 1, characterized in that The rock fracture prediction model for predicting the rock state stage obtained by training based on the sample data set includes: Inputting the acoustic emission characteristic parameters of the target acoustic emission signal in the sample data set into the rock fracture prediction model to be trained, and processing the acoustic emission characteristic parameters of the target acoustic emission signal by the rock fracture prediction model to be trained; Acquiring a predicted rock state stage corresponding to the target acoustic emission signal output by the rock fracture prediction model to be trained; Based on the rock state stage label and the predicted rock state stage, a loss value of the rock fracture prediction model to be trained is calculated; When the loss value is within a preset range, the trained rock fracture prediction model to be trained is used as the final rock fracture prediction model.

6. A real-time prediction device for gas storage surrounding rock fracture, characterized in that: The device comprises: An acoustic emission signal acquisition module, used to acquire acoustic emission signals generated by a specified type of rock in the surrounding rock of the gas storage reservoir during the fracture process and stress and strain parameters of the rock, wherein the acoustic emission signals are signals collected by a piezoelectric ceramic sensor arranged on the surface of the rock during the process of continuously applying external force to the rock, and the stress and strain parameters are parameters monitored by a displacement monitoring sensor arranged on the side of the rock during the process of continuously applying external force to the rock; A target signal screening module, used for screening target acoustic emission signals that meet training conditions from the acoustic emission signals based on acoustic emission characteristic parameters corresponding to the acoustic emission signals; A state stage label determination module, used to determine the rock state stage label corresponding to the target acoustic emission signal based on a stress-strain curve of stress-strain parameters corresponding to the target acoustic emission signal; A sample data set construction module, used to construct a sample data set based on the acoustic emission characteristic parameters of the target acoustic emission signal and the rock state stage label corresponding to the target acoustic emission signal; A rock fracture prediction model training module, used for training a rock fracture prediction model for predicting a rock state stage based on the sample data set; A rock state real-time prediction module is used to process the acoustic emission signals to be measured of the target rock of the specified type in the surrounding rock of the gas storage reservoir collected in real time by using the rock fracture prediction model to obtain the rock state stage corresponding to the target rock, wherein the acoustic emission signals to be measured are signals of the target rock around the reinforcement structure of the gas storage reservoir collected by the piezoelectric ceramic sensor; The state stage label determination module includes: A stress-strain curve determining unit, configured to determine a stress-strain curve corresponding to the target acoustic emission signal based on the stress-strain parameter corresponding to the target acoustic emission signal; A rock elastic modulus determination unit, used to determine the rock elastic modulus of the rock based on the slope of the stress-strain curve; A state stage label determination unit, used to determine the rock state stage label corresponding to the target acoustic emission signal based on the rock elastic modulus, the instantaneous secant modulus corresponding to the stress-strain curve, the strain parameter and the peak strain parameter of the rock; The state stage label determination unit comprises: An instantaneous secant modulus calculation subunit, used to calculate the instantaneous secant modulus corresponding to each point on the stress-strain curve based on the stress parameter and the strain parameter of each point; An elastic stage label determination subunit is used to determine that the rock state stage label of the target acoustic emission signal corresponding to the first point on the stress-strain curve is an elastic deformation stage label when the ratio of the instantaneous secant modulus to the rock elastic modulus corresponding to the first point on the stress-strain curve is greater than a first threshold value and the strain parameter of the first point is less than the peak strain parameter, and the first threshold value is a value greater than 0 and less than 1; A compaction stage label determination subunit is used to determine that the rock state stage label of the target acoustic emission signal corresponding to the second point on the stress-strain curve is a compaction stage label when the ratio of the instantaneous secant modulus to the rock elastic modulus corresponding to the second point on the stress-strain curve is greater than 0 and the strain parameter of the second point is less than a set multiple of the peak strain parameter, and the set multiple is a value greater than 0 and less than 1; A post-peak stage label determination subunit is used to determine that the rock state stage label of the target acoustic emission signal corresponding to the third point on the stress-strain curve is a post-peak failure stage label when the ratio of the instantaneous secant modulus to the rock elastic modulus corresponding to the third point on the stress-strain curve is less than 0; The plastic stage label determination subunit is used to determine the rock state stage labels of the remaining acoustic emission signals in the target acoustic emission signal as plastic stage labels, and the remaining acoustic emission signals refer to the acoustic emission signals in the target acoustic emission signal that are not the elastic deformation stage label, the compaction stage label, and the post-peak failure stage label.

7. The device according to claim 6, characterized in that The acoustic emission characteristic parameters include: acoustic emission impact number, acoustic emission energy, acoustic emission RA value, acoustic emission b value, signal duration, ringing count and acoustic emission rate.

8. The device according to claim 7, characterized in that The target signal screening module comprises: The target signal acquisition unit is used to eliminate the acoustic emission signals whose signal duration is less than a time threshold and / or whose ring count is less than a count threshold in the acoustic emission signals, so as to obtain the target acoustic emission signal.

9. The device according to claim 6, characterized in that The rock elastic modulus determination unit comprises: A target curve acquisition subunit, used to acquire a target stress-strain curve having a linear relationship on the stress-strain curve; The rock elastic modulus calculation subunit is used to calculate the rock elastic modulus of the rock based on the stress parameters and strain parameters corresponding to the starting point and the end point of the target stress-strain curve respectively.

10. The device according to claim 6, characterized in that The rock fracture prediction model training module includes: An acoustic emission characteristic parameter input unit, used to input the acoustic emission characteristic parameters of the target acoustic emission signal in the sample data set into the rock fracture prediction model to be trained, and the rock fracture prediction model to be trained processes the acoustic emission characteristic parameters of the target acoustic emission signal; A rock state stage prediction unit, used to obtain the predicted rock state stage corresponding to the target acoustic emission signal output by the rock fracture prediction model to be trained; A loss value calculation unit, used for calculating the loss value of the rock fracture prediction model to be trained based on the rock state stage label and the predicted rock state stage; The rock fracture prediction model acquisition unit is used to use the trained rock fracture prediction model to be trained as the final rock fracture prediction model when the loss value is within a preset range.

11. An electronic device, characterized in that: include: The memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the real-time prediction method for gas storage surrounding rock rupture according to any one of claims 1 to 5 by executing corresponding computer instructions.

12. A readable storage medium, characterized in that: Computer instructions are stored, which are used to enable a computer to execute the real-time prediction method for gas storage surrounding rock fracture according to any one of claims 1 to 5.

Images