CN116842854A - Intelligent prediction and reducing pressure relief method for coal body stress based on optimized neural network - Google Patents
Intelligent prediction and reducing pressure relief method for coal body stress based on optimized neural network Download PDFInfo
- Publication number
- CN116842854A CN116842854A CN202311121892.6A CN202311121892A CN116842854A CN 116842854 A CN116842854 A CN 116842854A CN 202311121892 A CN202311121892 A CN 202311121892A CN 116842854 A CN116842854 A CN 116842854A
- Authority
- CN
- China
- Prior art keywords
- pressure relief
- drilling
- stress
- hole
- neural network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 239000003245 coal Substances 0.000 title claims abstract description 60
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 47
- 238000000034 method Methods 0.000 title claims abstract description 33
- 238000005553 drilling Methods 0.000 claims abstract description 109
- 238000010276 construction Methods 0.000 claims abstract description 22
- 239000011435 rock Substances 0.000 claims abstract description 22
- 238000009826 distribution Methods 0.000 claims abstract description 15
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 13
- 230000008569 process Effects 0.000 claims abstract description 8
- 238000005065 mining Methods 0.000 claims abstract description 5
- 238000012549 training Methods 0.000 claims description 24
- 238000012795 verification Methods 0.000 claims description 24
- 238000003062 neural network model Methods 0.000 claims description 13
- 238000012544 monitoring process Methods 0.000 claims description 12
- 238000005070 sampling Methods 0.000 claims description 11
- 238000012937 correction Methods 0.000 claims description 7
- 230000002068 genetic effect Effects 0.000 claims description 5
- 230000005484 gravity Effects 0.000 claims description 4
- 239000002245 particle Substances 0.000 claims description 4
- 238000011156 evaluation Methods 0.000 claims description 3
- 238000002474 experimental method Methods 0.000 claims description 3
- 230000001419 dependent effect Effects 0.000 claims description 2
- 238000005516 engineering process Methods 0.000 abstract description 9
- 238000012545 processing Methods 0.000 abstract description 4
- 239000011148 porous material Substances 0.000 abstract description 2
- 230000002265 prevention Effects 0.000 abstract description 2
- 238000013461 design Methods 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 4
- 238000001514 detection method Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 240000007594 Oryza sativa Species 0.000 description 1
- 235000007164 Oryza sativa Nutrition 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 239000002817 coal dust Substances 0.000 description 1
- 230000008303 genetic mechanism Effects 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 230000008092 positive effect Effects 0.000 description 1
- 238000003825 pressing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 235000009566 rice Nutrition 0.000 description 1
- 210000000225 synapse Anatomy 0.000 description 1
- 238000005303 weighing Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/14—Force analysis or force optimisation, e.g. static or dynamic forces
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Computer Hardware Design (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Medical Informatics (AREA)
- Geometry (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Earth Drilling (AREA)
Abstract
The invention discloses an intelligent prediction and reducing pressure relief method for coal body stress based on an optimized neural network, and belongs to the field of prevention and control of coal mine rock burst by utilizing an intelligent prediction technology. The method specifically comprises the steps of collecting drilling parameters when a working face drills a pressure relief hole when the pressure relief hole is drilled, and constructing a prediction model by utilizing stress values obtained by accompanying holes of the pressure relief hole, so that the prediction model is utilized to predict the mining stress values when other working faces drill the pressure relief hole, and the mining stress distribution rule is analyzed according to the prediction result, so that the reaming position, the reaming section length and the pore size of the pressure relief hole are determined. According to the invention, while the drilling parameters are monitored in the construction process of the pressure relief hole, the stress value of the coal body corresponding to the drilling depth of the adjacent accompanying hole is monitored, and the intelligent algorithm such as a neural network is adopted for data processing, so that the stress value of the pressure relief drilling hole can be rapidly and accurately predicted in real time, the construction parameters of the pressure relief hole needing reaming are predicted in advance, and the accurate pressure relief is realized.
Description
Technical Field
The invention belongs to the field of underground engineering safety, in particular to the field of preventing and controlling coal mine rock burst by utilizing an intelligent prediction model combining a signal while drilling detection technology and an intelligent prediction technology, and particularly relates to an intelligent prediction and reducing pressure relief method for coal body stress based on an optimized neural network.
Background
Dynamic disasters such as rock burst and the like occur as the stress of the coal body is dynamically developed and changed along with the pushing of the working surface, and the dynamic disasters are the results of development and accumulation of mining stress. At present, a drilling pressure relief method is one of the more common methods for reducing the impact risk of coal and rock, and pressure relief holes with equal diameters are generally adopted, but if the diameters of the pressure relief holes are too small, the expected pressure relief effect cannot be achieved, and if the diameters of the pressure relief holes are too large, surrounding rock is damaged, so that the stability of the surrounding rock is affected. In order to achieve accurate pressure relief, the reducing pressure relief technology is developed at the present stage, when the pressure relief is carried out from the surface of surrounding rock to the depth of the surrounding rock, the pressure relief drilling needs to be reamed in a high stress area along the depth of the hole, and the stress distribution of the surrounding rock is changed in a manual mode. Therefore, how to conveniently and rapidly measure the distribution condition of the stress of the coal becomes a bottleneck problem for restricting the development of the reducing accurate pressure relief technology.
The traditional drilling pressure relief method adopts the coal dust amount discharged by drilling construction to indirectly reflect the coal body stress, and errors exist in the numerical conversion process; the discharged pulverized coal needs to be manually weighed, so that the rapid measurement of the stress of the coal body cannot be realized; weighing pulverized coal, suspending drilling construction, and causing discontinuous monitoring of parameters while drilling, wherein the parameters while drilling which are easy to cause monitoring when the drilling machine is frequently started are mostly abnormal values; the data volume of the parameter while drilling is huge, and the traditional method has low processing efficiency. For the reasons, the traditional drilling pressure relief method lacks accurate guiding measures, and cannot achieve accurate and efficient pressure relief effects.
The parameters while drilling have the characteristics of huge quantity and various types, have complex nonlinear relations with the stress of the coal body, and the traditional linear and nonlinear methods can not establish a correlation model between the two. The artificial neural network is a mathematical model for information processing by applying a structure similar to brain nerve synapse connection, is an operation model, and has been widely used for pattern recognition, signal processing, knowledge engineering, expert system, optimization combination, robot control and the like at present. However, the artificial neural network technology is prone to over-fitting, and it is difficult to ensure prediction accuracy. The genetic algorithm is an optimization algorithm for simulating natural selection and genetic mechanism, searches for an optimal solution through simulating an evolution process, and can be used for automatically adjusting network structure and weight parameters so as to improve the prediction performance of the neural network. However, the technical problem of intelligent prediction while drilling of rock mechanical parameters in a coal mine well is not realized at present by utilizing a neural network and an optimized neural network model, and the key technology puzzling the realization of the technology is as follows: on the one hand, the hardware detection equipment for detecting the while-drilling signals is required, on the other hand, the acquired while-drilling signals are processed by a computer system to establish a corresponding intelligent prediction model, and on the other hand, the key precondition of the establishment of the hardware detection equipment and the intelligent prediction model is to comprehensively plan the drilling construction of surrounding rocks under a coal mine, the acquisition of while-drilling parameters and the reasonable distribution design of the acquisition time and the space of the while-drilling parameters for training and prediction of a calculation model.
Therefore, it is necessary to research a highly-operative and accurate pressure relief method combining the signal while drilling detection and intelligent prediction technologies.
Disclosure of Invention
In order to conveniently and rapidly measure the distribution condition of the stress of the coal body and realize accurate pressure relief for the mine with impact tendency, the invention provides an intelligent prediction and reducing unloading method for the stress of the coal body based on an optimized neural network
And (5) pressing.
In order to achieve the above purpose, the technical scheme adopted by the invention is as follows:
an intelligent prediction and reducing pressure relief method for coal body stress based on an optimized neural network is characterized by comprising the following steps:
s1: training set data acquisition
S1.1: selecting a sampling working surface in the advancing direction of the stoping working surface, designing pressure relief holes on the sampling working surface according to the rock burst dangerous degree, and designing at least one accompanying hole for monitoring stress near each pressure relief hole;
s1.2: drilling construction and data acquisition
S1.2.1: firstly, drilling the accompanying holes, wherein the drilling method of the accompanying holes is constructed under the following two conditions:
in the first case, when the Prussian coefficient of the coal is more than or equal to 3.0, drilling an accompanying hole parallel to each pressure relief hole near the pressure relief hole, wherein the accompanying hole and the pressure relief hole are in the same horizontal plane and have the same hole depth, and a plurality of drilling stress meters are arranged in the accompanying hole at equal intervals along the hole depth;
in the second case, when the Prussian coefficient of the coal is less than 3.0, a plurality of accompanying holes with the depth increased by equal amount are drilled at equal intervals around the pressure relief hole, the depth of the deepest accompanying hole is equal to the depth of the pressure relief hole, and a drilling stress gauge is arranged at the bottom of each accompanying hole;
s1.2.2: after the drilling of the accompanying hole is completed, starting to drill a pressure relief hole, acquiring a drilling parameter a of a drilling machine when the pressure relief hole is drilled to the same depth position as a drilling stress meter in the accompanying hole, and acquiring a stress value b of the stress meter;
s2: establishing a predictive model
S2.1: sample dataset collation
Arranging all drilling parameters a acquired during drilling of the pressure relief holes and stress values b of the stress meters in the accompanying holes to form a training set;
s2.2: neural network prediction model establishment
By determining the coefficient R 2 Setting different numbers of training samples, hidden layers, hidden layer nodes and different parameter combinations while drilling as independent variables, establishing a comparison experiment of a corresponding artificial neural network model, examining the influence of the numbers of the training samples, the hidden layers, the hidden layer nodes and the different input characteristic parameter combinations on the model prediction accuracy, optimizing the optimal numbers of the training samples, the hidden layers and the hidden layer nodes and the optimal characteristic parameter combinations according to influence results, and establishing a neural network prediction model according to the optimal numbers of the optimized training samples, the optimal numbers of the hidden layers and the hidden layer nodes and the optimal characteristic parameter combinations;
s2.3. optimized neural network model establishment
Optimizing the artificial neural network model established in the step S2.2 by adopting a genetic algorithm and a particle swarm algorithm respectively;
s2.4: the optimal neural network prediction model is optimized by comparing the decision coefficients of the three neural network prediction models obtained in the step S2.2 and the step S2.3;
s3: intelligent prediction and reducing pressure relief of coal body stress
S3.1: intelligent prediction of coal stress
Selecting an adjacent working face of a sampling working face as a prediction working face, collecting parameters while drilling when the prediction working face is subjected to pressure relief hole construction, taking the parameters while drilling as input parameters, and predicting the stress of the coal body in real time by adopting the optimal neural network prediction model constructed in the step S2;
s3.2: determining reducing pressure relief scheme according to coal body stress distribution rule
Analyzing a coal body stress distribution rule according to a coal body stress prediction value, distinguishing a high stress area and a low stress area, determining a reaming position and an aperture section length of a pressure relief drilling hole according to the distribution condition of the high stress area, determining a reaming aperture according to the stress of the high stress area, and implementing reaming construction according to the determined reaming section length and the determined aperture size when the pressure relief drilling hole is drilled to the reaming position, so as to finally realize the purpose of accurate pressure relief;
s3.3: predictive model correction
When the pressure relief drilling is carried out on the predicted working surface, the optimal neural network prediction model is corrected at intervals, and the corrected neural network prediction model is utilized to predict the stress value of the coal body.
Further, the specific process of the step S3.3 prediction model correction is as follows:
when the pressure relief drilling is carried out on the predicted working surface, selecting a plurality of pressure relief hole drilling stress monitoring accompanying holes at intervals to serve as verification accompanying holes, constructing the verification accompanying holes according to the method of the step S1.2, collecting stress values and drilling parameters of the pressure relief holes at corresponding positions to form a verification data set, and carrying out error verification on the optimal neural network prediction model of the step S2.4 according to the verification data set; if the error exceeds the set range (the error value can be set according to the actual self-setting, if the set decision coefficient is greater than 0.8), the verification data set is added into the training sample set in the step S1, so as to correct the model, the corrected neural network prediction model is utilized to predict the coal body stress value during mining, and if the error is within the set range, the model does not need to be corrected.
Further, the while-drilling parameters a in step S1.2.2 include drilling parameters including a rotational speed, a feed pressure, a motor torque, a current voltage, and a motor real-time power, and vibration parameters including a mean value, a standard deviation, a mean square error, and a center of gravity frequency.
The Prussian coefficient f is also called as the firmness coefficient and the fastening coefficient of the rock, the calculation formula is f=R/10, wherein R is the uniaxial compressive strength of the rock, and the unit is MPa; the threshold value of the Prussian coefficient for distinguishing the hardness of the coal can be selected according to the specific situation of the site.
The decision coefficient of the present invention is an evaluation index commonly used for evaluating the performance of the neural network model, and the calculation method is common knowledge and will not be described in detail herein.
The positive effects of the present invention with respect to the prior art are described below.
For a mine with impact tendency, drilling and pressure relief are common anti-impact measures, in the prior art, equal-diameter pressure relief drilling and pressure relief are generally drilled on site according to working conditions and experiences, the construction method is equivalent to taking the stress of surrounding rock as a constant value, in fact, as the depth increases, the stress difference between the deep and shallow positions of the surrounding rock is larger, the pressure relief holes are easy to be too small or too large, the expected pressure relief effect cannot be achieved when the diameter of the pressure relief holes is too small, and the surrounding rock is damaged when the diameter of the pressure relief holes is too large, so that the stability of the surrounding rock is affected. When the pressure relief holes are drilled, drilling parameters are collected when one working face drills the pressure relief holes, stress values are obtained by utilizing the accompanying holes of the pressure relief holes to construct a prediction model, so that the stress values can be predicted in real time, quickly and accurately by utilizing the constructed prediction model when other working faces drill the pressure relief holes, meanwhile, the predicting model is corrected by arranging the accompanying holes at intervals, the accuracy of predicting the stress of the coal body is improved, the positions of the pressure relief holes needing reaming and the pore sizes of the reaming are predicted in advance, and the accurate pressure relief is realized.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic diagram of a sample face drilling layout according to an embodiment of the present invention;
FIG. 2 is a diagram of a differential pressure relief prediction scheme implemented in the vicinity of a sampling face in accordance with an embodiment.
FIG. 3 is a schematic diagram of a two sample face drilling layout in accordance with an embodiment of the present invention;
FIG. 4 is a perspective view of a pressure relief vent and stress monitoring companion vent arrangement of example two;
FIG. 5 is a side view of a pressure relief vent and stress monitoring companion vent arrangement of embodiment two;
FIG. 6 is a top view of a pressure relief vent and stress monitoring companion vent arrangement of embodiment two;
fig. 7 is a diagram of a differential pressure relief prediction scheme (verification accompanying holes omitted) implemented on the adjacent working surface of the second sampling working surface of the embodiment.
In the figure:
1-pressure relief holes, 2-accompanying holes, 3-stress gauges, 4-reducing pressure relief holes and 5-verification accompanying holes.
2 1 、2 2 、2 3 …2 i Representing the first, second, and third companion apertures … ith companion aperture in that order.
Detailed Description
The construction process of the embodiments of the present invention will be described in detail with reference to the accompanying drawings, so that the advantages and features of the present invention will be more readily understood by those skilled in the art, and thus the scope of the present invention will be more clearly defined.
The technical scheme of the invention is described in detail below by taking two mine with rock burst risk as an example.
Example one taking a mine with a Pursh coefficient of 3.0 or more with harder coal
S1: training set data acquisition
S1.1: drilling design
As shown in fig. 1, a certain working surface is taken as a sampling working surface, the positions (including the hole pitch and the row pitch of the pressure relief holes 1) and the hole diameter and the hole depth of the pressure relief holes are designed according to the dangerous degree of the field rock burst in the coal slope of a transportation gate way along the advancing direction of the stope working surface, meanwhile, a companion hole 2 is designed near each pressure relief hole 1, for example, at a position which is one meter away from the pressure relief hole 1, the companion hole 2 and the corresponding pressure relief hole 1 are identical and parallel in height from a bottom plate, and the hole depths of the companion hole 2 and the corresponding pressure relief hole 1 are identical;
s1.2: drilling construction and data acquisition
Firstly drilling a concomitant hole 2 before drilling a pressure relief hole 1, arranging a stress meter 3 in the concomitant hole 2 at intervals of one meter along the depth of the hole, then constructing the pressure relief hole 1, wherein the pressure relief hole 1 is an equal-diameter drilling hole, acquiring drilling parameters a including drilling parameters (parameters such as rotation speed, feeding pressure, motor torque, current voltage, real-time power of a motor and the like) and vibration parameters (parameters such as average value, standard deviation, mean square error, gravity center frequency and the like) at intervals of one meter in the drilling process of the pressure relief hole 1, and acquiring stress values b of the stress meter in the concomitant hole 2 corresponding to the pressure relief hole 1;
s2: establishing a predictive model
S2.1: sample dataset collation
And arranging all drilling parameters a acquired during drilling of the pressure relief holes 1 and stress values b in the accompanying holes 2 to form a training set.
S2.2: neural network prediction model establishment
By determining the coefficient R 2 As an evaluation index for model prediction accuracy
Setting training samples, hidden layers, hidden layer nodes and different input characteristic parameter combinations with different numbers as independent variables, and determining model prediction accuracy, namely determining coefficient R 2 As dependent variables, a corresponding comparison experiment of the artificial neural network model is established, and the influence of the number of training samples, hidden layers, hidden layer nodes and different input characteristic parameter combinations on the model identification precision is investigated, so that an optimal neural network model is obtained;
the combination of different input characteristic parameters can be a total combination mode of parameters while drilling, such as a total of 511 combination samples of 9 parameters while drilling, including a rotation speed, a feeding pressure, a motor torque, a current voltage, a motor real-time power, an average value, a standard deviation, a mean square error, a center of gravity frequency and the like. When the parameters while drilling are relatively large, the combined sample quantity is large, the calculated amount of the model is large, calculation resources are wasted, at the moment, the main component analysis method is firstly utilized to reduce the dimension of the parameters while drilling, for example, the dimension is reduced to 5 dimensions, namely 5 types of characteristic parameters, the 5 types of characteristic parameters are used as independent variables to be input into the neural network model for training, and the characteristic parameters in the combined sample corresponding to the optimal neural network model after training are the parameters while drilling which need to be collected for follow-up prediction.
S2.3: optimized neural network model building
And (2) respectively optimizing the optimal neural network model established in the step (2.2) by adopting algorithms such as a genetic algorithm, a particle swarm algorithm and the like to respectively obtain the optimal neural network model optimized by the genetic algorithm、An optimal neural network model optimized by a particle swarm algorithm;
s2.4: the optimal neural network model established in the step 2.2 is compared with the decision coefficients of the two optimal neural network models obtained by optimizing the two algorithms in the step 2.3, and an optimal neural network prediction model is optimized;
s3, coal body stress prediction and reducing pressure relief construction
S3.1 coal stress prediction
As shown in fig. 2, selecting an adjacent working surface of a sampling working surface as a prediction working surface, and performing drilling construction on the prediction working surface, wherein the stress of surrounding rock surfaces and shallow coal bodies is small, so that during construction, equal-diameter drilling holes (see equal-diameter sections of the different-diameter pressure relief holes 4 in fig. 2) are drilled for the first time, drilling parameters while drilling are collected in the drilling process, the parameters while drilling are taken as input parameters, and the stress of the coal bodies is predicted in real time by adopting the optimal neural network prediction model constructed in the second step;
s3.2: determining different-diameter pressure relief scheme according to analysis of coal stress distribution rule
According to the predicted value of the coal body stress, analyzing the coal body stress distribution rule, distinguishing a high stress area and a low stress area, determining the diameter expansion position of the reducing pressure relief hole 4 (see the diameter expansion section of the reducing pressure relief hole 4 in fig. 2), determining the diameter expansion position and the length of the diameter expansion section of the reducing pressure relief hole 4 according to the distribution range of the high stress area, determining the diameter expansion construction reducing pressure relief according to the stress of the high stress area, and finally realizing the purpose of accurate pressure relief.
S3.3: predictive model correction
When the pressure relief drilling is carried out on the predicted working surface, a plurality of pressure relief holes (which can be correspondingly adjusted according to the actual situation of the site) are drilled at intervals, namely a verification accompanying hole 5, the verification accompanying hole 5 is provided with a stress meter 3 according to the mode of step S1.2, when the pressure relief hole corresponding to the verification accompanying hole 5 is drilled, the stress value of the stress meter in the verification accompanying hole 5 and the parameter while drilling are collected simultaneously to form a verification data set, and according to the verification data set, error verification is carried out on the optimal neural network prediction model in the step S2.4. If the error exceeds a certain range (the error value can be set according to the actual self-setting, if the set decision coefficient is greater than 0.8), the verification data set is added into the training sample set in the step S1, and then the optimal neural network prediction model is corrected. If the error is within a certain range, the optimal neural network prediction model does not need to be corrected.
Example two taking a softer coal mine with a Prussian coefficient of < 3.0 as an example
The second embodiment is different from the first embodiment mainly in that firstly, the design and construction of the accompanying hole are different during the data acquisition of the training set in step S1, and secondly, the drilling method of the accompanying hole 5 is verified to be different during the correction of the prediction model in step S3.3. The concrete steps are as follows:
s1: training set data acquisition
S1.1: in connection with fig. 3-6, during drilling design and construction, a circle of accompanying holes 2 for monitoring stress are designed and constructed around the pressure relief holes 1 near each pressure relief hole 1, for example, a plurality of accompanying holes 2 are drilled at equal intervals on a circumference taking the pressure relief hole 1 as a circle center and taking 1 meter as a radius, the depths of the plurality of accompanying holes 2 are sequentially increased by equal amounts, and the depth of the deepest accompanying hole 2 is equal to the depth of the pressure relief hole 1; wherein the number and depth of satellite boreholes are determined as follows:
(1) determination of the number of accompanying holes 2
As shown in FIG. 5, the depth of the relief hole 1 is set to an integer L meter (when L is not an integer, the integer is rounded off), and each drill is designed in advanceOne meter collects parameters while drilling, and the number of the accompanying holes 2 is L, so that one stress value is collected per meter of drilling. The L companion holes 2 in fig. 5 are sequentially referred to as first companion holes 2 by counterclockwise numbering 1 Second companion hole 2 2 Third companion hole 2 3 … ith companion well 2 i Wherein i is 1,2,3, … L. For example, when the depth of the relief hole 1 is 10 meters, the number of the accompanying holes 2 is 10, the depth of the relief hole 1 is 9.4 meters, the number of the accompanying holes 2 is 9, and when the depth of the relief hole 1 is 10.5 meters, the number of the accompanying holes 2 is 11.
(2) Depth L of companion hole 2 i Is determined by (a)
As shown in FIG. 6, if a strain gauge length L is used y Rice, then the first accompanying hole 2 1 Second companion hole 2 2 Third companion hole 2 3 … ith companion well 2 i Depth ofL i Bits are meters) is calculated as:where i is 1,2,3 … L, i.e. the centre of the strain gauge 3 is guaranteed to be located at a position corresponding to the accompanying hole depth minus half the length of the strain gauge when the strain gauge 3 is installed;
s1.2: drilling construction
First, the first pilot hole 2 is constructed according to the number of the designed pilot holes 2 and the depth of each pilot hole 2 1 First associated hole 2 2 Third companion hole 2 3 … ith companion well 2 i A stress meter is arranged at the bottom of each accompanying hole 2; after the construction of the auxiliary holes 2 is finished, the pressure relief holes 1 are drilled, when the pressure relief holes 1 are drilled to 1 meter and 2 meters … L (10 meters), the parameters a while drilling are respectively collected once, and the corresponding first auxiliary holes 2 are collected 1 Second companion hole 2 2 Third companion hole 2 3 … ith companion well 2 i Stress value b of the intermediate stress meter 3.
Step S3.3: predictive model correction
In the correction of the neural network prediction model, the stress monitoring accompanying holes in which a plurality of pressure relief holes 1 are drilled at intervals are called verification accompanying holes 5 according to the method of the first embodiment, and the drilling method of the verification accompanying holes 5 is performed according to the design construction method of step S1. Example two prediction methods according to the present invention a precise pressure relief prediction scheme of reducing diameter implemented on a surface adjacent to a sampling surface is shown in fig. 7.
The invention has the beneficial effects that: (1) Based on the while-drilling parameters, the stress change of the coal body is intelligently predicted, and according to the existing pressure relief theory, the pressure relief drilling parameters such as the drilling aperture, the depth and the like are dynamically adjusted and optimized according to the pressure relief requirement, so that the purpose of accurate pressure relief is realized; (2) Accurately sensing the stress concentration degree of the coal body, pre-judging the impact disaster risk degree, and providing support for disaster prevention; (3) The coal body stress distribution rule is clarified, support is provided for the deployment of the working face roadway, and the stoping roadway is deployed at a position avoiding stress peaks.
It should be noted that the innovation of the invention is mainly that an accompanying hole is arranged on a sampling working surface by utilizing each pressure relief hole, stress at different depth positions of each pressure relief hole is obtained by utilizing the accompanying hole, while drilling parameters of the pressure relief holes are collected, training is carried out through a neural network by utilizing a one-to-one correspondence relation between the while drilling parameters and stress values, so that an optimal neural network prediction model of the coal body stress is obtained, a great number of stress monitoring holes are not needed to be beaten in the later working surface by utilizing the model, and the stress values of different depth positions of each pressure relief hole can be quickly and accurately collected in real time only by correcting the model at intervals, so that parameters of different diameter holes are quickly and accurately obtained. How to train the while-drilling parameters by using the neural network is not an important point of the present invention, such as how to examine the influence of the training samples, the hidden layers, the number of hidden layer nodes, and the combination of different input characteristic parameters on the model recognition accuracy, and how to set the training samples, the hidden layers, the hidden layer nodes, and the different input characteristic parameters, which are all known in the art, is not excessively stated in the present invention. In addition, in the embodiment, the number and depth of the accompanying holes 2 are determined based on the stress value and the while-drilling parameter acquired by taking one meter into consideration, which is just a scale convenient for operation and data processing under the condition of considering the cost and the accuracy of the prediction result, and is not a limitation of the technical scheme of the present invention, in practice, as long as any scale of the while-drilling parameter and the stress value acquired under the condition of ensuring the same depth is within the protection scope of the present invention, therefore, several improvements made under the condition of not departing from the principle of the present invention should be considered as the protection scope of the present invention.
Claims (3)
1. An intelligent prediction and reducing pressure relief method for coal body stress based on an optimized neural network is characterized by comprising the following steps:
s1: training set data acquisition
S1.1: selecting a sampling working surface in the advancing direction of the stoping working surface, designing pressure relief holes according to the rock burst dangerous degree, and designing at least one accompanying hole for monitoring stress near each pressure relief hole;
s1.2: drilling construction and data acquisition
S1.2.1: firstly, drilling the accompanying holes, wherein the drilling method of the accompanying holes is constructed under the following two conditions:
in the first case, when the Prussian coefficient of the coal is more than or equal to 3.0, drilling an accompanying hole parallel to each pressure relief hole near the pressure relief hole, wherein the accompanying hole and the pressure relief hole are in the same horizontal plane and have the same hole depth, and a plurality of drilling stress meters are arranged in the accompanying hole at equal intervals along the hole depth;
in the second case, when the Prussian coefficient of the coal is less than 3.0, a plurality of accompanying holes with the depth increased by equal amount are drilled at equal intervals around the pressure relief hole, the depth of the deepest accompanying hole is equal to the depth of the pressure relief hole, and a drilling stress gauge is arranged at the bottom of each accompanying hole;
s1.2.2: after the drilling of the accompanying hole is completed, starting to drill a pressure relief hole, acquiring a drilling parameter a of a drilling machine when the pressure relief hole is drilled to the same depth position as a drilling stress meter in the accompanying hole, and acquiring a stress value b of the stress meter;
s2: establishing a predictive model
S2.1: sample dataset collation
Arranging all drilling parameters a acquired during drilling of the pressure relief holes and stress values b of the stress meters in the accompanying holes to form a training set;
s2.2: neural network prediction model establishment
By determining the coefficient R 2 As an evaluation index of model prediction precision, setting training samples, hidden layers, hidden layer nodes and different parameter combinations while drilling with different numbers as independent variables, setting model prediction precision as a dependent variable, establishing a comparison experiment of a corresponding artificial neural network model, and establishing a neural network prediction model;
s2.3. optimized neural network model establishment
Optimizing the neural network prediction model established in the step S2.2 by adopting a genetic algorithm and a particle swarm algorithm respectively;
s2.4: the optimal neural network prediction model is optimized by comparing the decision coefficients of the three neural network prediction models obtained in the step S2.2 and the step S2.3;
s3: intelligent prediction and reducing pressure relief of coal body stress
S3.1: intelligent prediction of coal stress
Selecting an adjacent working face of a sampling working face as a prediction working face, collecting parameters while drilling when the prediction working face is subjected to pressure relief hole construction, and predicting the stress of the coal body in real time by adopting the optimal neural network prediction model constructed in the step S2;
s3.2: determining reducing pressure relief scheme according to coal body stress distribution rule
Analyzing the coal stress distribution rule according to the predicted value of the coal stress, and determining a reducing pressure relief scheme according to the coal stress distribution rule;
s3.3: predictive model correction
When the pressure relief drilling is carried out on the predicted working surface, the optimal neural network prediction model is corrected at intervals, and the corrected neural network prediction model is utilized to predict the stress value of the coal body.
2. The intelligent prediction and reducing pressure relief method for coal body stress based on the optimized neural network as claimed in claim 1, wherein the specific process of modifying the prediction model in step S3.3 is as follows: when the pressure relief drilling is carried out on the predicted working surface, selecting a plurality of pressure relief hole drilling stress monitoring accompanying holes at intervals to serve as verification accompanying holes, constructing the verification accompanying holes according to the method of the step S1.2, collecting stress values and drilling parameters of the pressure relief holes at corresponding positions to form a verification data set, and carrying out error verification on the optimal neural network prediction model of the step S2.4 according to the verification data set; if the error exceeds the set range, the verification data set is added into the training sample set in the step S1, the model is further corrected, the corrected neural network prediction model is utilized to predict the coal body stress value during mining, and if the error is within the set range, the model does not need to be corrected.
3. The method for intelligent prediction and reducing pressure relief of coal body stress based on optimized neural network as claimed in claim 1, wherein the while drilling parameters a in step S1.2.2 include drilling parameters including rotational speed, feed pressure, motor torque, current voltage and motor real-time power, and vibration parameters including mean value, standard deviation, mean square error and center of gravity frequency.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311121892.6A CN116842854B (en) | 2023-09-01 | 2023-09-01 | Intelligent prediction and reducing pressure relief method for coal body stress based on optimized neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311121892.6A CN116842854B (en) | 2023-09-01 | 2023-09-01 | Intelligent prediction and reducing pressure relief method for coal body stress based on optimized neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116842854A true CN116842854A (en) | 2023-10-03 |
| CN116842854B CN116842854B (en) | 2023-11-07 |
Family
ID=88160299
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311121892.6A Active CN116842854B (en) | 2023-09-01 | 2023-09-01 | Intelligent prediction and reducing pressure relief method for coal body stress based on optimized neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116842854B (en) |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105260599A (en) * | 2015-09-30 | 2016-01-20 | 山东黄金矿业(莱州)有限公司三山岛金矿 | Rockburst dynamic prediction method based on BP neural network modeling |
| CN108843331A (en) * | 2018-07-06 | 2018-11-20 | 山东科技大学 | Slim hole joint release method for arranging under the equivalent drilling area of one kind |
| CN110500095A (en) * | 2019-09-20 | 2019-11-26 | 西安科技大学 | Slim hole pressure relief method under a kind of equivalent area |
| CN110852018A (en) * | 2019-10-21 | 2020-02-28 | 中国石油集团长城钻探工程有限公司 | PSO drilling parameter optimization method based on neural network |
| CN110905402A (en) * | 2019-11-13 | 2020-03-24 | 山东科技大学 | Pressure relief hole construction method based on mining induced stress dynamic monitoring |
| CN111291997A (en) * | 2020-02-18 | 2020-06-16 | 山东科技大学 | Coal seam impact risk real-time evaluation method based on measurement while drilling technology |
| CN113008440A (en) * | 2021-03-10 | 2021-06-22 | 山东科技大学 | Flexible liquid injection sensor detection method based on genetic algorithm optimization neural network |
| CN114320268A (en) * | 2021-12-20 | 2022-04-12 | 山东唐口煤业有限公司 | Large-diameter drilling pressure relief effect evaluation method based on drilling stress monitoring |
| CN115405363A (en) * | 2022-08-30 | 2022-11-29 | 华北科技学院 | Coal mine rock burst monitoring and early warning system based on LSTM neural network |
| US20230266500A1 (en) * | 2020-07-31 | 2023-08-24 | Hamed Soroush | Geomechanics and wellbore stability modeling using drilling dynamics data |
-
2023
- 2023-09-01 CN CN202311121892.6A patent/CN116842854B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105260599A (en) * | 2015-09-30 | 2016-01-20 | 山东黄金矿业(莱州)有限公司三山岛金矿 | Rockburst dynamic prediction method based on BP neural network modeling |
| CN108843331A (en) * | 2018-07-06 | 2018-11-20 | 山东科技大学 | Slim hole joint release method for arranging under the equivalent drilling area of one kind |
| CN110500095A (en) * | 2019-09-20 | 2019-11-26 | 西安科技大学 | Slim hole pressure relief method under a kind of equivalent area |
| CN110852018A (en) * | 2019-10-21 | 2020-02-28 | 中国石油集团长城钻探工程有限公司 | PSO drilling parameter optimization method based on neural network |
| CN110905402A (en) * | 2019-11-13 | 2020-03-24 | 山东科技大学 | Pressure relief hole construction method based on mining induced stress dynamic monitoring |
| CN111291997A (en) * | 2020-02-18 | 2020-06-16 | 山东科技大学 | Coal seam impact risk real-time evaluation method based on measurement while drilling technology |
| US20230266500A1 (en) * | 2020-07-31 | 2023-08-24 | Hamed Soroush | Geomechanics and wellbore stability modeling using drilling dynamics data |
| CN113008440A (en) * | 2021-03-10 | 2021-06-22 | 山东科技大学 | Flexible liquid injection sensor detection method based on genetic algorithm optimization neural network |
| CN114320268A (en) * | 2021-12-20 | 2022-04-12 | 山东唐口煤业有限公司 | Large-diameter drilling pressure relief effect evaluation method based on drilling stress monitoring |
| CN115405363A (en) * | 2022-08-30 | 2022-11-29 | 华北科技学院 | Coal mine rock burst monitoring and early warning system based on LSTM neural network |
Non-Patent Citations (1)
| Title |
|---|
| 王继成;: "七星煤矿东四区八层右部冲击矿压分析与防治", 民营科技, no. 06 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116842854B (en) | 2023-11-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111291934B (en) | Surrounding rock real-time grading prediction and self-checking method in tunnel construction process | |
| Rajesh Kumar et al. | Prediction of uniaxial compressive strength, tensile strength and porosity of sedimentary rocks using sound level produced during rotary drilling | |
| CN107292467A (en) | A kind of drilling risk Forecasting Methodology | |
| CN109033504A (en) | A kind of casing damage in oil-water well prediction technique | |
| CN112987125B (en) | Shale brittleness index prediction method based on logging data | |
| CN116595809B (en) | Underground engineering surrounding rock drilling pressure relief-detection evaluation method | |
| CN111291997A (en) | Coal seam impact risk real-time evaluation method based on measurement while drilling technology | |
| CN112504838B (en) | TBM-loaded rock mechanics comprehensive test and information evaluation system | |
| CN116451013B (en) | Deep stratum rock in-situ drillability grade value prediction method | |
| CN104153768A (en) | Granite reservoir stratum reservoir performance evaluation method | |
| CN116522692A (en) | Underground engineering surrounding rock structural feature in-situ detection and classification method | |
| Gowida et al. | Synthetic well-log generation: New approach to predict formation bulk density while drilling using neural networks and fuzzy logic | |
| Bajolvand et al. | Optimization of controllable drilling parameters using a novel geomechanics-based workflow | |
| CN116842854B (en) | Intelligent prediction and reducing pressure relief method for coal body stress based on optimized neural network | |
| CN113283182A (en) | Method, device, medium and equipment for predicting and analyzing formation pressure | |
| CN111946397B (en) | Rapid method for on-site evaluation of integrity of tunnel face rock and soil mass of heading machine | |
| CN111340275B (en) | Tunnel support mode selection real-time prediction method based on detection while drilling technology | |
| CN116658157B (en) | Stratum pressure prediction method and system for tight sandstone gas reservoir | |
| US20230323770A1 (en) | Systems, apparatuses, and methods for determining rock mass properties based on blasthole drill performance data including compensated blastability index (cbi) | |
| CN117350144A (en) | Rock mass strength characteristic prediction method based on machine learning | |
| CN112329255A (en) | Rock burst prediction method based on tendency degree and uncertain measure | |
| CN111625916A (en) | Method and system for calculating stability value of well wall | |
| CN111709129B (en) | Method for determining safety coefficient of tunnel excavation in surrounding rock with rock-like pile body fracture | |
| CN114066271A (en) | Tunnel water inrush disaster monitoring and management system | |
| CN112381283A (en) | Tunnel disease treatment method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |