A method for intelligently determining sectional coal pillar size based on while-drilling parameters

Abstract

The present application provides a kind of section coal pillar size intelligent determination method based on while drilling parameter, it is related to coal mining technical field, carries out while drilling measurement drilling experiment under different confining pressure and different elastic-plastic state condition and records data;Stress meter is installed in situ to carry out stress monitoring, obtains the stress data of coal pillar before and after working face mining;The experimental data and stress test data obtained are preprocessed, and the data is divided into training set and test set;Intelligent model for determining section coal pillar size is established, and the model is trained using training set data;Test the model using test set data, i.e. parameter adjustment;Drilling rig is used to drill coal pillar, and intelligent model meeting the requirements is used to determine the size of section coal pillar;Collect the while drilling parameter and coal pillar data fed back in the use process, optimize the intelligent model and discriminant model.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of coal mining, in particular to a sectional coal pillar size intelligent determination method based on drilling parameters. BACKGROUND

[0002] Setting a coal pillar to protect the sectional roadway is the main method adopted by many coal mines, and the reasonable size of the sectional coal pillar plays an extremely important role in the stability of the coal pillar itself and the roadway and the improvement of the coal recovery rate. If the sectional coal pillar size is too large, it will lead to waste of coal resources, and if the sectional coal pillar size is too small, it may cause damage to the coal pillar itself and the adjacent roadway, posing a threat to the safety of workers, causing damage to the equipment of the roadway and working face, and unnecessary losses to the coal mine economy. Therefore, the rationality of the sectional coal pillar size determination is very important.

[0003] Different coal mines have different geological conditions, and the surrounding rock properties of different working faces in the same coal mine also differ. Therefore, the sectional coal pillar size required to be set for working faces under different geological conditions and different stress conditions should also be different. At present, there are various methods for determining the sectional coal pillar size, but the main methods can be roughly summarized into four kinds: one is to measure the stress distribution state in the coal body under the action of mining through field point arrangement, and to determine the specific recovery roadway layout and reasonable sectional coal pillar size in combination with the related theory of mine pressure; two is to use the elastic-plastic theory to deduce the sectional coal pillar size when the coal pillar is in a stable state, to obtain a reasonable sectional coal pillar size calculation formula, and to determine the reasonable sectional coal pillar size by assigning different coefficients according to different geological conditions; three is to use computer software for numerical simulation, to input the related geological information and parameters into the simulation software for calculation and simulation, and to analyze and determine the reasonable size of the sectional coal pillar according to the simulation results; four is to use the traditional field production experience, to select the sectional coal pillar size under similar field conditions in the past as a reference for the setting of the sectional coal pillar.

[0004] The commonly used sectional coal pillar size determination methods have certain limitations, for example, the field point arrangement measurement operation is relatively complex, consumes a large amount of manpower and material resources, and increases the production cost of the coal mine; the calculation formula has the problem of idealized assumption conditions, and the accuracy of the coal pillar size determination is low; the model establishment and parameter selection process of the computer simulation is relatively complex, and the result depends on personal interpretation; the field production experience is easily limited by the region, and when determining the coal pillar size, it is often designed conservatively, causing waste of coal resources. As can be seen, the commonly used sectional coal pillar size determination methods are difficult to realize the unity of convenience and accuracy.

[0005] Therefore, it is of great significance to develop a sectional coal pillar size determination method integrating convenience and accuracy for the current roadway coal pillar design. As a cutting-edge technology, the measurement-while-drilling technology can capture and monitor a series of key drilling parameters such as drilling speed, rotation speed, torque and drilling pressure in real time during drilling operation, and then accurately analyze and predict the lithology and stress state based on these parameters. After the working face is mined, the stress state in the coal pillar will change significantly, and when the drilling rig passes through the coal pillar, the stress changes will be indirectly reflected through the drilling parameters generated by drilling, and the sectional coal pillar size to be designed can be determined through the stress changes in the coal pillar. Based on this principle, the measurement-while-drilling technology can play an important role in the determination of the sectional coal pillar size. SUMMARY

[0006] The purpose of the present application is to provide an intelligent sectional coal pillar size determination method based on drilling parameters, which first determines the relationship between drilling parameters and stress in the coal pillar through laboratory experiments, discriminates the stress changes before and after the coal pillar is mined through real-time stress monitoring on site, combines the theoretical method for determining the coal pillar size, and uses artificial intelligence algorithm to associate the three, and finally intelligently determines the coal pillar size.

[0007] An intelligent sectional coal pillar size determination method based on drilling parameters, comprising:

[0008] Step one, perform measurement-while-drilling experiments under different confining pressures and record data;

[0009] Step two, install a stress meter on site to monitor stress and obtain stress test data inside the coal pillar before and after the working face is mined;

[0010] Step three, preprocess the obtained experimental data and stress test data, and divide the data into a training set and a test set;

[0011] Step four, establish an intelligent sectional coal pillar size prediction model for determining the sectional coal pillar size, and use the training set data to train the intelligent sectional coal pillar size prediction model;

[0012] Step five, test the trained intelligent sectional coal pillar size prediction model using the test set data, adjust the parameters, and obtain an intelligent sectional coal pillar size determination model;

[0013] Step six, use a drilling rig to drill the sectional coal pillar, and use the intelligent sectional coal pillar size determination model to predict the sectional coal pillar size;

[0014] Step seven, collect the feedback drilling parameters and coal pillar size data during use, and optimize the intelligent sectional coal pillar size determination model.

[0015] In the preferred embodiment, the step four, the section coal pillar size determination intelligent model construction process is as follows:

[0016] S41, taking the drilling parameters obtained from the drilling measurement drilling experiment as input parameters and the stress state of the test piece as output parameters, a model A of the relationship between the drill bit pressure, torque and the stress state of the test piece during drilling is constructed based on the DQN reinforcement learning model, and the model A is trained;

[0017] S42, the pre-mining coal pillar stress data x and the post-mining coal pillar stress data y are fitted to obtain a fitting formula;

[0018] S43, the drilling experiment test piece stress data is taken as the post-mining stress data y1, and the corresponding pre-mining stress data x1 in the state of the test piece is solved out;

[0019] S44, the pre-mining stress data x1 is taken into the model A to obtain the theoretical pre-mining drilling parameters under the laboratory parameter conditions;

[0020] S45, taking the theoretical pre-mining drilling parameters obtained in step S44 as input parameters and the post-mining stress as output parameters, a relationship model B of the pre-mining drilling parameters and the post-mining stress is constructed based on the DQN reinforcement learning model, and the model B is trained;

[0021] S46, the stress value when the stress is initially stable and the drilling length d0 at this time and the stress value when the stress is secondly stable and the drilling length d1 at this time are recorded, and the calculation formula of the coal pillar size M is:

[0022] ;

[0023] Wherein, ;

[0024] The formula is built into the model B, the output stress y1 of the model B is mapped as the drilling length d output at the stress, and the coal pillar size M calculation formula built into the model B is input, and finally the section coal pillar size M to be left is output;

[0025] The training process in S41 is repeated again to train the section coal pillar size intelligent prediction model, and after completion, the trained section coal pillar size intelligent prediction model is output.

[0026] In the preferred embodiment, the step S41 comprises:

[0027] S411, determining the state space : ;

[0028] Wherein, is the rotation speed at time t; is the torque at time t; is the coal pillar width; is the number of damaged specimens;

[0029] S412, determining an action space;

[0030] S413, designing a reward function ;

[0031] ;

[0032] wherein, is the original data; is the corrected data;

[0033] S414, -greedy action selection:

[0034] ;

[0035] wherein, is the action; is the exploration rate;

[0036] S415, calculating a target Q value:

[0037] ;

[0038] wherein, is the target Q value; is the immediate reward; is the current state; is the discount factor; is the state at next time; is the target network output; is the target network parameter; is the maximization operation; is all possible actions of the next state.

[0039] S416, gradient descent update:

[0040] ;

[0041] wherein, is the parameter; is the learning rate; is the gradient operator; is the loss function.

[0042] In a preferred embodiment, the step five comprises:

[0043] S51, import the test set data into the trained section coal pillar size intelligent prediction model;

[0044] S52, check whether the output parameters processed by the section coal pillar size intelligent prediction model are consistent with the theoretical output parameters, and determine the error thereof;

[0045] S53, if the error is too large and cannot meet the accuracy requirement, adjust the parameters of the section coal pillar size intelligent prediction model, and then test the test set data after completion;

[0046] S54, if the error meets the accuracy requirement, output the section coal pillar size intelligent determination model.

[0047] In the preferred embodiment, the step six comprises:

[0048] S61, drill 50m ahead of the working face, and the initial drilling is located at the coal pillar side roadway side of the open-off cut 50m away;

[0049] S62, drill at three points, i.e., upper, middle and lower points, at the drilling position, and take the average value of the final results;

[0050] S63, drill the section coal pillar at intervals of 50m using a drilling machine;

[0051] S64, record the coal seam mining depth h, mining thickness m, and the pre-mining coal pillar broken zone width a0, plastic zone width a1, and drilling length b0 at the initial stress stabilization.

[0052] In the preferred embodiment, the step one comprises:

[0053] S11, prepare a test piece, and determine the lithology parameters of the test piece;

[0054] S12, group the test pieces, apply different confining pressures to each group of test pieces, and construct the internal stress state;

[0055] S13, perform a while-drilling measurement drilling experiment on different test piece groups, and record the while-drilling parameters.

[0056] In the preferred embodiment, the step two comprises:

[0057] S21, determine the station and measuring point positions;

[0058] S22, install the stress device and debug;

[0059] S23, collect coal pillar stress test data, and compare and analyze the coal pillar stress state changes at different depths and the stress state relationship of the coal pillar before and after mining.

[0060] In the preferred embodiment, the step seven comprises:

[0061] S71, collecting the drilling parameters generated by drilling and the pre-mining stress collected by the stress device;

[0062] S72, data cleaning on the drilling parameters and the pre-mining stress;

[0063] S73, introducing the measured pre-mining drilling parameters and the pre-mining stress into the sectional coal pillar size intelligent prediction model to obtain the predicted sectional coal pillar size;

[0064] S74, setting the coal pillar according to the predicted sectional coal pillar size, observing the deformation of the coal pillar and the surrounding roadway after the setting, and correcting the sectional coal pillar size intelligent prediction model according to the deformation;

[0065] S75, predicting the coal pillar size by using the corrected sectional coal pillar size intelligent prediction model, and circulating the steps S74 and S75 to improve the accuracy.

[0066] In the preferred embodiment, a correction parameter is inserted into the sectional coal pillar size intelligent prediction model The output result of the sectional coal pillar size intelligent prediction model is corrected, and when the deformation of the coal pillar and the surrounding roadway is large, the correction parameter is set to be greater than 1 >1, when the deformation of the coal pillar and the surrounding roadway is small or no deformation, the correction parameter is set to be less than 1 .

[0067] The present application has the following beneficial technical effects:

[0068] Compared with the traditional sectional coal pillar size determination method, the sectional coal pillar size determination method based on the drilling parameters has the following advantages: only drilling is needed on the coal pillar side roadway before the working face is mined to obtain the sectional coal pillar size to be set, and the operation is more convenient; the drilling parameters are analyzed and processed by the intelligent model, the influence of personal subjectivity is avoided, and the accuracy of the sectional coal pillar size determination is enhanced; the drilling machine can correct the obtained sectional coal pillar size as the working face advances, and the accuracy of the sectional coal pillar size determination can be further improved. BRIEF DESCRIPTION OF DRAWINGS

[0069] Figure 1 It is a structure schematic diagram of a while-drilling measuring tester;

[0070] Figure 2 It is a shape and size schematic diagram of a test piece;

[0071] Figure 3 It is a pre-pressing elastic-plastic state and stress distribution schematic diagram of a while-drilling measuring experiment;

[0072] Figure 4 It is a field stress monitoring station and measuring point layout schematic diagram;

[0073] Figure 5 A schematic diagram of field drilling and coal pillar stress zoning;

[0074] Figure 6 A flow chart of a sectional coal pillar size intelligent determination method based on drilling parameters;

[0075] Figure 7 A specific implementation flowchart of a sectional coal pillar size intelligent determination method based on drilling parameters.

[0076] Reference signs: 1 - drilling while measuring drilling hole; 2 - sectional coal pillar; 3 - non-coal pillar side roadway; 4 - sectional roadway; 5 - roadway roof; 6 - stress curve. DETAILED DESCRIPTION

[0077] In order to make the purpose, technical solutions and advantages of the embodiments clearer, the technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0078] In the drawings of the specific embodiments of the present application, in order to better and more clearly describe the working principles of the elements in the system and show the connection relationship of the parts in the device, only the relative position relationship between the elements is distinguished, and the signal transmission direction, connection order, and the position, size and shape of the parts in the structure cannot be limited.

[0079] Secondly, the "one embodiment" or "embodiment" referred to here means that the specific features, structures or characteristics can be included in at least one implementation of the present application. The "in one embodiment" appearing in different places in the present application does not mean the same embodiment, nor is it an independent or alternative embodiment that excludes other embodiments.

[0080] The present application will be further described below by embodiments and in combination with the drawings, but is not limited to this. Based on the embodiments in the present application:

[0081] The present embodiment provides a sectional coal pillar size intelligent determination method based on drilling parameters, and the specific steps are as follows:

[0082] Step one, drilling while measuring drilling experiment under different confining pressures and recording relevant data.

[0083] The drilling while measuring drilling experiment under different confining pressures is as follows:

[0084] S11, prepare a test piece and determine the lithology parameters of the test piece.

[0085] As Figure 2 shown, the test piece preparation material is cement, river sand and water, the test piece shape is a cubic test piece, and the size is 300 mm*300 mm*300 mm;

[0086] A part of the poured test piece is cored, uniaxial compression experiment is carried out, and the mechanical property parameters of the test piece are obtained, including uniaxial compressive strength and elastic modulus.

[0087] S12, the test pieces are grouped, different confining pressures are applied to each group of test pieces, and the internal stress state is constructed.

[0088] The test pieces are divided into 5 groups, each group has 3 test pieces, the coal pillar form is simulated, each group of test pieces is labeled, and the specific test piece combination form is shown in Figure 3 , the simulated coal seam mining thickness is n=300 mm, the simulated section coal pillar width is 3n=900 mm, the three test pieces in the same group are placed horizontally and closely, so as to reduce the influence of the gap between the test pieces on the drilling experiment;

[0089] Each group of test pieces is pre-pressed by applying different confining pressures, which can press the test pieces on the one hand, so that they will not move due to drilling, and on the other hand, it can make the internal stress state and elastic-plastic state different, so as to simulate the stress state of the coal pillar under different buried depth conditions. After the pressurization is completed, the simulated coal pillar state is shown in Figure 3 , wherein the unilateral broken zone width of the coal pillar is c0, the unilateral plastic zone width of the coal pillar is c1, and the stress state of the coal pillar is shown as the stress curve 6 in Figure 3 .

[0090] S13, the drilling-while-measuring drilling experiment is carried out on different test piece groups, and the drilling parameters are recorded.

[0091] The drilling-while-measuring drilling rig shown in Figure 1 is used to carry out the drilling-while-measuring drilling experiment on different test piece groups, and the drilling parameters such as drilling speed, drilling length, drilling pressure, torque and test piece related parameters in the drilling process are recorded. Preferably, the drilling experiment in step one is carried out on the basis that the coal pillar is in the state of having been left, that is, the drilling-while-measuring parameters obtained by drilling are determined by the stress state of the coal pillar after mining. This obviously does not meet the demand of predicting the coal pillar width before mining. To solve this problem, the stress state of the coal pillar before and after mining needs to be studied on site, and a stress relationship model of the coal pillar before and after mining is constructed. The specific method is shown in step two.

[0092] Preferably, as shown in Figure 1The figure shows a schematic diagram of a measurement while drilling (MWD) rig. Its main structure includes a drill pipe, a propulsion device, and a pressure device. The sensing device attached to the drill pipe can obtain real-time MWD parameters during drilling. The propulsion device can control the propulsion force, propulsion speed, and rotation speed of the drill pipe. The pressure device is equipped with a non-uniformly distributed loading cylinder, which can load different parts of the specimen with different forces to achieve pre-compression at different positions of the specimen, so that it reaches different stresses and different elastic-plastic states.

[0093] Step 2: Install a stress gauge on site to monitor stress and obtain stress test data inside the coal pillar before and after mining at the working face.

[0094] The specific steps of on-site stress monitoring are as follows:

[0095] S21, determine the location of the measuring station and measuring points;

[0096] like Figure 4 As shown, a stress gauge is installed by drilling holes in the side of the coal pillar 50m ahead of the working face. The stress gauge is divided into three measuring stations with a spacing of 50m. Each measuring station has 10 measuring points. The depth difference of each measuring point is set to 5m. The depth difference between measuring point 9 and measuring point 10 is 10m, and the spacing between measuring points is 2m.

[0097] S22. Install the stress device and debug it;

[0098] Install the strain gauge at the selected measuring point and debug it using relevant software after installation.

[0099] S23. Collect coal pillar stress test data, compare and analyze the changes in coal pillar stress state at different depths and the relationship between the stress states of the coal pillar before and after mining.

[0100] Stress data at different depths of the coal pillar are collected before and after mining until all stress gauges at the measuring points enter the goaf for more than 100 m. The collected data are imported into relevant software to compare and analyze the changes in the stress state of the coal pillar at different depths and the relationship between the stress state of the coal pillar before and after mining.

[0101] Step 3: Preprocess the acquired experimental data and stress test data, and divide the data into training set and test set.

[0102] The specific steps of data preprocessing are as follows:

[0103] S31. Data cleaning.

[0104] Delete duplicate data, fill missing data, delete data with too much missing data, and use the Z-score statistical method to preliminarily repair abnormal data.

[0105] S32. Calculate the mean :

[0106] ;

[0107] wherein: is the total number of data; is the initial data value of the th.

[0108] S33, calculate the standard deviation :

[0109] ;

[0110] wherein: is the mean value of the data; is the total number of data; is the initial data value of the th.

[0111] S34, calculate the Z-score of each data:

[0112] ;

[0113] wherein: is the standardized value; is the original data value.

[0114] S35, data segmentation.

[0115] The data is divided into a training data set and a test data set, wherein the training data set accounts for 70%, and the test data set accounts for 30%.

[0116] S36, adding feature values.

[0117] Add new features such as products and differences to highlight the representative data in the data set, making it easier to image fit the data later.

[0118] S37, data augmentation.

[0119] Randomly insert and delete data to increase the diversity of data and prevent overfitting.

[0120] S38, data batch processing.

[0121] The data is divided into 64 groups for easy input of data in the model later.

[0122] Step four, establish a section coal pillar size determination intelligent prediction model, and use the training set data to train the section coal pillar size determination intelligent prediction model.

[0123] The construction process of the section coal pillar size determination intelligent model is as follows:

[0124] S41, taking the DQN reinforcement learning model as a basic model, taking the while-drilling parameters obtained from the while-drilling measurement drilling experiment as input parameters, and taking the stress state of the test piece as an output parameter, the DQN reinforcement learning model is imported, and a model of the relationship between the while-drilling parameters such as drill bit pressure and torque and the stress state of the test piece during drilling is constructed, which is named as model A and is trained, and the specific process is as follows.

[0125] S411, determining the state space : ;

[0126] Among them, is the rotating speed at time t; is the torque at time t; is the coal pillar width; is the number of damaged test pieces;

[0127] S412, determining the action space : ;

[0128] S413, designing a reward function :

[0129] ;

[0130] In the formula, is the original data; is the corrected data;

[0131] S414, -greedy action selection

[0132] ;

[0133] In the formula, is the exploration rate;

[0134] S415, calculating the Target Q-value

[0135] ;

[0136] In the formula, is the target Q value; is the immediate reward; is the current state; is the discount factor; is the state at the next time; is the target network output; is the target network parameter; is the maximum operation; is all possible actions of the next state.

[0137] S416, gradient descent update: ;

[0138] wherein, are parameters of model A; is learning rate; is gradient operator; is loss function.

[0139] Preferably, learning rate is 0.0005 to control the update step of parameters, discount factor is 0.95 to balance the current and future rewards, exploration rate decay is 0.5 to gradually change the model from exploration to utilization, network update frequency C is 500 steps to synchronize to stabilize the model.

[0140] S42, import the coal pillar stress data obtained before and after mining into the related software, and use the software to perform data fitting with the pre-mining stress data as x and the post-mining stress data as y to obtain a fitting formula;

[0141] S43, the stress data of the drilling experiment sample is taken as the post-mining stress data y1 and brought into the fitting formula to solve the pre-mining stress data x1 corresponding to the state of the sample.

[0142] S44, the pre-mining stress data x1 is brought into model A to obtain the theoretical pre-mining drilling parameters under the conditions of related parameters in the laboratory.

[0143] S45, again taking the DQN reinforcement learning model as the basic model, taking the theoretical pre-mining drilling parameters as the input parameters and the post-mining stress as the output parameters, a relationship model of pre-mining drilling parameters and post-mining stress is constructed, which is named as model B, and the specific training process of model B is consistent with the training process of model A in S41.

[0144] S46, record the stress value when the stress is initially stable in the drilling experiment and the drilling length d0 at this time, and the stress value when the stress is secondly stable and the drilling length d1 at this time, according to the sectional coal pillar size calculation formula wherein , the sectional coal pillar size M required to be set can be obtained. n refers to Figure 3 the height of the sample in the formula, that is, the height of the simulated coal seam mining.

[0145] The formula The built-in model B maps the command of the output stress y1 of the model B when the model B outputs to the drilling length d at the stress when it outputs, inputs the built-in coal pillar size calculation formula, and presents the final output content of the model B as the section coal pillar size M required to be left. The model is a section coal pillar size intelligent prediction model. The training process in S41 is repeated again to train the section coal pillar size intelligent prediction model, and after completion, the trained section coal pillar size intelligent prediction model is output.

[0146] Step five, test the trained section coal pillar size intelligent prediction model using the test set data, adjust the parameters, and obtain the section coal pillar size intelligent determination model, the steps are as follows:

[0147] S51, import the test set data into the trained section coal pillar size intelligent prediction model.

[0148] S52, check whether the output parameters obtained after processing by the section coal pillar size intelligent prediction model are consistent with the theoretical output parameters and determine the error.

[0149] S53, if the error is too large and cannot meet the accuracy requirement, adjust the parameters of the section coal pillar size intelligent prediction model, and after completion, use the test set data for testing again.

[0150] S54, if the error meets the accuracy requirement, output the section coal pillar size intelligent determination model.

[0151] Step six, use the drilling machine to drill the section coal pillar, and use the section coal pillar size intelligent determination model to predict the size of the section coal pillar.

[0152] The steps of section coal pillar drilling and size determination are as follows:

[0153] S61, determine the drilling position

[0154] Drill 50m ahead of the working face, that is, the initial drilling is located at the coal pillar side lane side of the open-off cut 50m away.

[0155] S62, determine the drilling point

[0156] As shown in the drilling measurement drilling hole 1 in Figure 5 To ensure the reliability of the section coal pillar size value prediction, drilling should be carried out at three points of upper, middle and lower in the drilling position, and the average value is taken as the final result.

[0157] S63, use the drilling machine to drill the section coal pillar.

[0158] The drilling interval is 50m, which is convenient for real-time adjustment of the section coal pillar size according to different geological conditions.

[0159] S64, data acquisition

[0160] As Figure 5 , record the coal seam mining depth h, mining thickness m and other related geological parameters and the pre-mining coal pillar broken zone width a0, plastic zone width a1, drilling length b0 at the initial stress stabilization and other drilling data, so as to facilitate further optimization of the subsequent model.

[0161] Step seven, collect the feedback of the drilling parameters and coal pillar size data in the use process, and optimize the intelligent determination model of the section coal pillar size.

[0162] The pre-mining drilling parameters and the pre-mining stress of the coal pillar in S43 and S44 in step four are calculated, so there will be some errors in the section coal pillar size intelligent determination model output by the laboratory drilling measurement drilling experiment when predicting the section coal pillar size. Therefore, the section coal pillar size intelligent determination model can be further optimized, and the specific optimization process is as follows.

[0163] S71, collect the drilling parameters generated during drilling and the pre-mining stress collected by the stress device.

[0164] S72, data cleaning of the drilling parameters and the pre-mining stress.

[0165] S73, import the measured pre-mining drilling parameters and pre-mining stress into the section coal pillar size intelligent prediction model after the test is completed, and obtain the predicted section coal pillar size.

[0166] S74, set the coal pillar according to the predicted size, and observe the deformation of the coal pillar and the surrounding roadway after the setting is completed. According to the deformation, the section coal pillar size intelligent prediction model is corrected, and the correction method is to insert a correction parameter in the section coal pillar size intelligent prediction model to correct the output result of the model. When the deformation of the coal pillar and the surrounding roadway is large, the correction parameter should be greater than 1, and when the deformation of the coal pillar and the surrounding roadway is small or no deformation, the correction parameter .

[0167] S75, use the corrected section coal pillar size intelligent prediction model to predict the coal pillar size, and cycle S74 and S75 steps to further improve the accuracy of the section coal pillar size intelligent prediction model.

[0168] Preferably, the geological parameters and drilling data obtained from the field construction are used to further optimize the model in order to expand the model database, so that it has the ability to cope with various geological conditions and disturbances, enhances the universality of the model and further improves the accuracy of the model.

[0169] The specific embodiments of the present application are described above. It needs to be understood that the present application is not limited to the specific embodiments described above, and various changes or modifications can be made by those skilled in the art within the scope of the claims, which do not affect the essential content of the present application. The embodiments of the present application and the features in the embodiments can be combined with each other arbitrarily without conflict.

Claims

1. A method for intelligent determination of section coal pillar size based on while-drilling parameters, characterized in that, The application relates to a method for determining the size of a section coal pillar. Step one: performing a measurement-while-drilling experiment under different confining pressures and recording data; Step two: installing a stress meter on site to monitor stress and obtain stress test data of the coal pillar before and after mining of the working face; Step three: preprocessing the obtained experimental data and stress test data, and dividing the data into a training set and a test set; Step four: establishing a section coal pillar size determination intelligent prediction model, and training the section coal pillar size determination intelligent prediction model using the training set data; the section coal pillar size determination intelligent model is constructed as follows: S41: taking a DQN reinforcement learning model as a basic model, taking the measurement-while-drilling parameters obtained through the measurement-while-drilling experiment as input parameters, and taking the stress state of the test piece as output parameters, a model A of the relationship between the drill bit pressure, torque and the stress state of the test piece during drilling is constructed, and the model A is trained; S42: data fitting is performed on the pre-mining coal pillar stress data x and the post-mining coal pillar stress data y to obtain a fitting formula; S43: the stress data of the test piece in the drilling experiment are taken as the post-mining stress data y1, and the pre-mining stress data x1 corresponding to the stress state of the test piece is solved by the fitting formula; S44: the pre-mining stress data x1 is taken into the model A to obtain the theoretical pre-mining measurement-while-drilling parameters under the laboratory parameter conditions; S45: the DQN reinforcement learning model is taken as the basic model again, the theoretical pre-mining measurement-while-drilling parameters obtained in step S44 are taken as the input parameters, and the post-mining stress is taken as the output parameter, a relationship model B of the pre-mining measurement-while-drilling parameters and the post-mining stress is constructed, and the model B is trained; S46: the stress value when the stress is initially stable in the drilling experiment and the drilling length d0 at the time and the stress value when the stress is secondly stable and the drilling length d1 at the time are recorded, and the calculation formula of the coal pillar size M is as follows: M >= 2d0 + (d1-d0) = d0+d1, Wherein, (d1-d0) >= n; n is the height of the simulated coal seam mining; The formula M >= d0+d1 is embedded in the model B, the command of outputting the stress y1 of the model B is mapped to the drilling length d when the stress is output, and the calculation formula of the coal pillar size M embedded in the model B is input, and finally the section coal pillar size M required to be set is output; The training process in S41 is repeated again to train the section coal pillar size intelligent prediction model, and the trained section coal pillar size intelligent prediction model is output after completion; Step five: the test set data are used to test the trained section coal pillar size intelligent prediction model, parameters are adjusted, and the section coal pillar size intelligent determination model is obtained; Step six: a drilling machine is used to drill the section coal pillar, and the section coal pillar size intelligent determination model is used to predict the size of the section coal pillar; Step seven: the measurement-while-drilling parameters and the coal pillar size data fed back in the use process are collected, and the section coal pillar size intelligent determination model is optimized.

2. The method for intelligent determination of section coal pillar size based on while-drilling parameters according to claim 1, characterized in that, The step S41 comprises: S411, determine state space S t : S t = [ROP t , τ t , ω t , N]; Where, ROP t is the rotating speed at time t; τ t is the torque at time t; ω t is the coal pillar width; N is the number of damaged test pieces; S412, determine action a t space; S413, design reward function r t ; wherein ω 原 is the original data; ω 新 is the corrected data; S414, ε-greedy action a t Selection: where a t is the action; ε is the exploration rate; S415: calculating a target Q value: where: y i is the target Q-value; r i is the immediate reward; s i is the current state; γ is the discount factor; s i+1 is the next state; Q is the target network output; θ - is the target network parameters; max a' is the maximization operation; a' is all possible actions for the next state. S416: gradient descent update: where θ is a parameter; a is a learning rate; is a gradient operator; ξ(θ) is a loss function.

3. The method of claim 1, wherein, The step five comprises: S51: the test set data are input into the trained section coal pillar size intelligent prediction model; S52, check whether the output parameters processed by the sectional coal pillar size intelligent prediction model are consistent with the theoretical output parameters, and determine the error thereof; S53, if the error is too large and cannot meet the accuracy requirement, adjust the parameters of the sectional coal pillar size intelligent prediction model, and then test the test set data after completion; S54, if the error meets the accuracy requirement, output the sectional coal pillar size intelligent determination model.

4. The method of claim 1, wherein, The step six comprises: S61, drilling is performed at the advanced working face 50 m away, and the initial drilling hole is located at the coal pillar side roadway side of the coal pillar at a distance of 50 m from the open-off cut; S62, drilling is performed at three points, i.e., upper, middle and lower points, and the final result is obtained by taking the average value thereof; S63, drilling is performed at an interval of 50 m by using a drilling machine; S64, the coal seam mining depth h, the mining thickness m, and the pre-mining coal pillar broken zone width a0, the plastic zone width a1, and the drilling length b0 at the initial stress stability are recorded.

5. The method of claim 1, wherein, The step one comprises: S11, preparing a test piece and determining the lithology parameters of the test piece; S12, grouping the test pieces, applying different confining pressures to each group of test pieces, and constructing the internal stress state; S13, performing a while-drilling measurement drilling experiment on different test piece groups, and recording the while-drilling parameters.

6. The method of claim 1, wherein, The step two comprises: S21, determining the station and measuring point positions; S22, installing a stress device and debugging; S23, collecting coal pillar stress test data, and comparing and analyzing the coal pillar stress state changes at different depths and the stress state relationship of the coal pillar before and after mining.

7. The method of claim 1, wherein, The step seven comprises: S71, collecting the while-drilling parameters generated by drilling and the pre-mining stress collected by the stress device; S72, data cleaning is performed on the while-drilling parameters and the pre-mining stress; S73, the measured pre-mining while-drilling parameters and pre-mining stress are introduced into the sectional coal pillar size intelligent prediction model, and the predicted sectional coal pillar size is obtained; S74, the sectional coal pillar size is set according to the predicted sectional coal pillar size, and after the setting is completed, the deformation of the coal pillar and the surrounding roadway is observed, and the sectional coal pillar size intelligent prediction model is corrected according to the deformation; S75, the sectional coal pillar size is predicted by using the corrected sectional coal pillar size intelligent prediction model, and the steps S74 and S75 are cycled to improve the accuracy.

8. The method of claim 7, wherein, The correction parameter ξ is inserted into the sectional coal pillar size intelligent prediction model to correct the output result of the sectional coal pillar size intelligent prediction model, when the deformation of the coal pillar and the surrounding roadway is large, the correction parameter ξ>1, and when the deformation of the coal pillar and the surrounding roadway is small or no deformation, the correction parameter ξ≤1.

Images