CN117345344B - Intelligent prediction method and system for mine acoustic and electric signals - Google Patents
Intelligent prediction method and system for mine acoustic and electric signals Download PDFInfo
- Publication number
- CN117345344B CN117345344B CN202311564610.XA CN202311564610A CN117345344B CN 117345344 B CN117345344 B CN 117345344B CN 202311564610 A CN202311564610 A CN 202311564610A CN 117345344 B CN117345344 B CN 117345344B
- Authority
- CN
- China
- Prior art keywords
- data
- layer
- lstm
- mine
- acoustic
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 49
- 230000005670 electromagnetic radiation Effects 0.000 claims abstract description 31
- 239000011435 rock Substances 0.000 claims abstract description 22
- 238000012545 processing Methods 0.000 claims description 13
- 230000005611 electricity Effects 0.000 claims description 7
- 230000005484 gravity Effects 0.000 claims description 7
- 238000011156 evaluation Methods 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 4
- 238000013139 quantization Methods 0.000 claims description 4
- 230000006835 compression Effects 0.000 claims description 3
- 238000007906 compression Methods 0.000 claims description 3
- 239000010410 layer Substances 0.000 description 60
- 230000002159 abnormal effect Effects 0.000 description 22
- 238000012544 monitoring process Methods 0.000 description 20
- 238000001514 detection method Methods 0.000 description 17
- 238000009826 distribution Methods 0.000 description 11
- 238000010586 diagram Methods 0.000 description 7
- 239000003245 coal Substances 0.000 description 5
- 230000000875 corresponding effect Effects 0.000 description 5
- 238000010276 construction Methods 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 3
- 238000012549 training Methods 0.000 description 3
- 230000010485 coping Effects 0.000 description 2
- 238000005520 cutting process Methods 0.000 description 2
- 238000007405 data analysis Methods 0.000 description 2
- 238000007418 data mining Methods 0.000 description 2
- 238000005553 drilling Methods 0.000 description 2
- 239000002355 dual-layer Substances 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000005065 mining Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 238000012800 visualization Methods 0.000 description 2
- 230000003321 amplification Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000012854 evaluation process Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000000556 factor analysis Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000003199 nucleic acid amplification method Methods 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Classifications
-
- E—FIXED CONSTRUCTIONS
- E21—EARTH OR ROCK DRILLING; MINING
- E21F—SAFETY DEVICES, TRANSPORT, FILLING-UP, RESCUE, VENTILATION, OR DRAINING IN OR OF MINES OR TUNNELS
- E21F17/00—Methods or devices for use in mines or tunnels, not covered elsewhere
- E21F17/18—Special adaptations of signalling or alarm devices
-
- 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
- G06N3/044—Recurrent networks, e.g. Hopfield networks
- G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/30—Assessment of water resources
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Physics & Mathematics (AREA)
- Mining & Mineral Resources (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Geology (AREA)
- Biophysics (AREA)
- Geochemistry & Mineralogy (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- General Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
Abstract
The invention discloses an intelligent prediction method and system for a mine acoustic-electric signal, comprising the following steps: s1: acquiring mine acoustic emission data and electromagnetic radiation data of rock burst dangers; s2: normalizing the acoustic emission data and the electromagnetic radiation data, and comprehensively evaluating the normalized acoustic emission data and electromagnetic radiation data by adopting an entropy method to obtain acoustic-electric comprehensive data; constructing a data set based on the sound-electricity comprehensive data; s3: and constructing an LSTM-Autoencoder model, and obtaining an intelligent prediction result of the mine acousto-optic signal based on the data set and the LSTM-Autoencoder model. The invention can intelligently predict future signals so as to timely early warn rock burst disasters.
Description
Technical Field
The invention belongs to the technical field of mine safety monitoring, and particularly relates to an intelligent prediction method and system for mine acoustic and electric signals.
Background
Along with the deepening of the mining depth of the mines in China, the stress of the stopes is increased, the dynamic disasters of the coal and the rock become more complex and serious, and serious threats are caused to the safety production of the mines and the safety of personnel. The key point of effective early warning and prevention of coal and rock dynamic disasters such as rock burst, coal, gas and the like is that advanced monitoring means and methods are adopted. Acoustic emissions, electromagnetic radiation, are positively correlated with the load and deformation fracture process of the coal and rock mass, and are substantially enhanced as the load and deformation fracture strength increases. The current monitoring methods are divided into regional monitoring and local monitoring, wherein the regional monitoring method generally adopts a microseismic monitoring method, and the local monitoring method comprises a drilling cutting method, stress on-line monitoring, acoustoelectric monitoring and the like. Since rock burst occurs at and near the face where impact is a risk, continuous local impact risk monitoring of the face is necessary, and thus local monitoring is particularly important in rock burst control.
At present, the existing local monitoring technology establishes various rock burst disaster early warning criteria and early warning methods from various angles such as electromagnetic waves, coal rock strength and the like by monitoring methods such as acoustic emission, electromagnetic radiation, drilling cutting method, stress on line and the like. However, in the prior art, the early warning of rock burst disasters is mostly carried out based on a history rule, and research on data to be generated in the future is lacking. The timeliness of rock burst early warning can be greatly improved by reasonably and accurately predicting and quantitatively analyzing the future acousto-electric data. Therefore, there is a need for an intelligent prediction method and system for acoustic-electric signals of mines with high accuracy.
Disclosure of Invention
The invention aims to solve the defects of the prior art, and provides an intelligent prediction method and an intelligent prediction system for a mine acoustic-electric signal, which can intelligently predict future signals so as to realize early warning on rock burst dangers in time.
In order to achieve the above object, the present invention provides the following solutions:
an intelligent prediction method for a mine acoustic-electric signal comprises the following steps:
s1: acquiring mine acoustic emission data and electromagnetic radiation data of rock burst dangers;
S2: normalizing the acoustic emission data and the electromagnetic radiation data, and comprehensively evaluating the normalized acoustic emission data and electromagnetic radiation data by adopting an entropy method to obtain acoustic-electric comprehensive data; constructing a data set based on the sound-electricity comprehensive data;
s3: and constructing an LSTM-Autoencoder model, and obtaining an intelligent prediction result of the mine acousto-optic signal based on the data set and the LSTM-Autoencoder model.
Preferably, in step S2, the method for obtaining the electroacoustic combination data includes:
Taking the time-average acoustic emission intensity and the time-average acoustic emission ringing of the acoustic emission data and the time-average electromagnetic intensity and the time-average electromagnetic pulse of the electromagnetic radiation data as prediction indexes, and carrying out dimensionalization processing on the prediction indexes to obtain normalized data X i＝{xi1,xi2,...,xij, wherein X ij represents a j index corresponding to an i sample data;
calculating the proportion of each value in each index to all values based on the normalized data:
wherein n is the number of samples, and m is the number of indexes;
calculating information entropy of each index based on the specific gravity:
calculating index weights of the indexes based on the information entropy:
Calculating the index comprehensive score of each index based on the index weight:
based on the index comprehensive score, obtaining the sound-electricity comprehensive data: y= { Y 1,y2,...,yj }.
Preferably, in step S3, the LSTM-Autoencoder model includes an input layer, an encoder, an intermediate layer, a decoder, and an output layer;
The encoder is a double-layer LSTM, and each layer comprises a first hidden layer and a first embedded layer;
the decoder is a reverse double layer LSTM, each layer comprising a second hidden layer and a second embedded layer, opposite to the encoder arrangement.
Preferably, in step S3, the method for obtaining the predicted result of the acoustic-electric signal based on the LSTM-Autoencoder model includes:
inputting input data to the encoder based on the input layer, obtaining a compressed representation of the intermediate layer Representing the compression/>The input decoder decompresses and reconstructs the input data to obtain the prediction result of the acoustic-electric signal
Wherein the input data has the format ofT is time;
preferably, the parameters of the LSTM-Autoencoder model include input window, sliding step size, output window, hidden layer, embedded layer, learning rate, batch size, optimizer, and number of iterations.
Preferably, the method further comprises S4: and evaluating the prediction performance of the LSTM-Autoencoder model by reconstructing the difference between output data and input data of the loss quantization model, wherein the formula is as follows:
the invention also provides an intelligent prediction system for the mine acoustic-electric signal, which applies the prediction method and comprises a data acquisition module, a data processing module and a prediction module;
The data acquisition module is used for acquiring mine acoustic emission data and electromagnetic radiation data of rock burst dangers;
the data processing module is used for carrying out normalization processing on the acoustic emission data and the electromagnetic radiation data, and carrying out comprehensive evaluation on the normalized acoustic emission data and electromagnetic radiation data by adopting an entropy method to obtain acoustic-electric comprehensive data; constructing a data set based on the sound-electricity comprehensive data;
The prediction module is used for constructing an LSTM-Autoencoder model and obtaining an intelligent prediction result of the mine acoustic-electric signal based on the data set and the LSTM-Autoencoder model.
Preferably, the LSTM-Autoencoder model comprises an input layer, an encoder, an intermediate layer, a decoder and an output layer;
The encoder is a double-layer LSTM, and each layer comprises a first hidden layer and a first embedded layer;
the decoder is a reverse double layer LSTM, each layer comprising a second hidden layer and a second embedded layer, opposite to the encoder arrangement.
Compared with the prior art, the invention has the beneficial effects that: normalizing the acoustic emission data and the electromagnetic radiation data, and comprehensively evaluating the normalized acoustic emission data and electromagnetic radiation data by adopting an entropy method to obtain acoustic-electric comprehensive data; constructing a data set based on the sound-electricity comprehensive data; through the application of the entropy method, the weight and the comprehensive score of each index are scientifically determined, and more scientific and accurate sound and electricity comprehensive data are obtained; and constructing an LSTM-Autoencoder model, and obtaining an intelligent prediction result of the mine acousto-optic signal based on the data set and the LSTM-Autoencoder model. And the future signal is intelligently predicted so as to timely realize early warning on rock burst danger.
Drawings
In order to more clearly illustrate the technical solutions of the present invention, the drawings that are needed in the embodiments are briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of an intelligent prediction method for mine acoustic and electric signals according to an embodiment of the invention;
FIG. 2 is a diagram of a data comprehensive evaluation process according to an embodiment of the present invention;
FIG. 3 is a diagram of modeling in accordance with an embodiment of the present invention;
fig. 4 is a signal prediction diagram according to an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
In order that the above-recited objects, features and advantages of the present invention will become more readily apparent, a more particular description of the invention will be rendered by reference to the appended drawings and appended detailed description.
Example 1
As shown in fig. 1-4, an intelligent prediction method for mine acoustic-electric signals comprises the following steps:
s1: acquiring mine acoustic emission data and electromagnetic radiation data of rock burst dangers;
In particular, because acoustic emission data and electromagnetic radiation data are obtained through the monitoring system, however, mine signal environments are complex, and during normal production, a large number of interference signals are generated by factors such as mining procedures, personnel activities, electrical equipment and the like, and the accuracy of the monitoring data and the accuracy of the prediction results are affected by the interference signals and the duration of the interference signals. The acoustic emission intensity is generally about 200 mV-400 mV, the electromagnetic radiation intensity is generally about 100 mV-300 mV, but the signal intensity is fluctuated by interference signals, so the filtering of the interference signals is also carried out before the data normalization is carried out in the embodiment:
Establishing a normal signal data set and an interference signal data set to obtain time domain features, frequency domain features and time frequency features of normal acousto-electric signals and the acousto-electric signals with the interference signals; comparing the characteristics of the normal acousto-electric signal and the acousto-electric signal mixed with the interference signal, and marking distinguishing characteristics;
training the model by using marked distinguishing features based on a two-way long-short-term memory cyclic neural network model to realize the identification of the acoustic-electric interference signals;
And deleting or replacing the identified acousto-electric interference.
S2: filling acoustic emission data and electromagnetic radiation data according to the hours, normalizing, and comprehensively evaluating the normalized acoustic emission data and electromagnetic radiation data by adopting an entropy method to obtain acoustic-electric comprehensive data; constructing a data set based on the sound-electricity comprehensive data; specifically, the sound-electricity data is calculated according to 8:1:1 is divided into a training set, a verification set and a test set, the training set, the verification set and the test set are input into an LSTM-Autoencoder model, an input window is set, and a sliding step length and an output window divide the input data set into samples and realize the amplification of the samples.
In a further embodiment, in step S2, the method for obtaining the electroacoustic combination data includes:
Taking the time-average acoustic emission intensity, the time-average acoustic emission ringing of acoustic emission data and the time-average electromagnetic intensity and the time-average electromagnetic pulse of electromagnetic radiation data as prediction indexes, and carrying out dimensionality removal treatment on the prediction indexes to obtain normalized data X i＝{xi1,xi2,...,xij, wherein X ij represents a j index corresponding to an i sample data;
calculating the proportion of each value in each index to all values based on the normalized data:
wherein n is the number of samples, and m is the number of indexes;
Calculating information entropy of each index based on specific gravity:
If pij＝0,Ei＝0；
Based on the information entropy, calculating the index weight of each index:
calculating an index comprehensive score of each index based on the index weight:
Obtaining sound and electricity comprehensive data based on the index comprehensive score: y= { Y 1,y2,...,yj }.
S3: and constructing an LSTM-Autoencoder model, and obtaining an intelligent prediction result of the mine acoustic-electric signal based on the data set and the LSTM-Autoencoder model.
A further embodiment is that in step S3, the LSTM-Autoencoder model includes an input layer, an encoder, an intermediate layer, a decoder, and an output layer;
Encoder with a plurality of sensors For a dual-layer LSTM, each layer comprising a first hidden layer and a first embedded layer; the size of the embedded layer is smaller than the size of the input layer in order to obtain a compressed representation/>, of the intermediate layer
The decoder g θ (·) is a reverse double layer LSTM, each layer comprising a second hidden layer and a second embedded layer, reconstructing the output, contrary to the encoder setup
In a further embodiment, in step S3, the method for obtaining the prediction result of the acoustic-electric signal based on the LSTM-Autoencoder (LSTM self-encoder) model is as follows:
inputting input data to an encoder based on an input layer, obtaining a compressed representation of an intermediate layer Representation of compression/>The input decoder decompresses and reconstructs the input data to obtain the prediction result/>
Wherein the input data has the format ofT is time; the input window is the length of each sample, the sliding step is to expand the sample size, and the output window is the signal length that needs to be predicted.
A further embodiment is that the parameters of the LSTM-Autoencoder model include input window, sliding step size, output window, hidden layer, embedded layer, learning rate, batch size, optimizer, and number of iterations.
A further embodiment is that the method further comprises S4: the prediction performance of the LSTM-Autoencoder model is evaluated by reconstructing the difference between the output data and the input data of the loss quantization model, and the L1 loss is minimized by using the Mean Absolute Error (MAE), and the formula is as follows:
Particularly, according to the prediction result of the acoustic-electric signal output by the model, the embodiment also carries out the next step of result analysis, carries out the tracing of the mine dangerous signal, further realizes the rock burst danger early warning in a targeted manner, and carries out intelligent decision based on the early warning result;
specifically, the method for tracing the mine dangerous signal based on the prediction result of the acoustic-electric signal comprises the following steps:
Firstly, based on the prediction result of the acoustic-electric signal, abnormal signal detection is performed:
Establishing a signal characteristic distribution diagram based on the normal acousto-electric signal;
extracting features of the sound-electricity signal prediction result to obtain signal features;
Constructing a signal detection model, and carrying out signal detection based on the signal detection model and signal characteristics to obtain a detection result; the detection result comprises a normal signal and a suspicious signal;
the signal detection model comprises an identity mapping module and a convolution module;
analyzing the feature distribution of the suspicious signals, and finally determining whether the suspicious signals are abnormal signals or not based on the signal feature distribution map;
and secondly, tracing the signal based on the characteristic distribution of the abnormal signal. Besides time-frequency domain characteristics, the characteristic distribution also comprises the position of signal acquisition equipment (an acoustic-electric sensor) in a mine and the time period of signal acquisition, so that intelligent positioning of abnormal points of the mine and monitoring points about to be dangerous is realized.
Based on the discovery of mine abnormal points, the method for realizing intelligent decision comprises the following steps: the prior abnormal condition data and corresponding coping strategies are collected through data analysis and mining, and the prior data (characters, images and the like) of different data structures are processed through a multi-mode model, so that a decision maker is helped to obtain prior comprehensive real information, and intelligent decision is provided.
Example two
The invention also provides an intelligent prediction system for the mine acoustic-electric signal, which is applied with a prediction method and comprises a data acquisition module, a data processing module and a prediction module;
The data acquisition module is used for acquiring mine acoustic emission data and electromagnetic radiation data of rock burst dangers;
The data processing module is used for carrying out normalization processing on the acoustic emission data and the electromagnetic radiation data, and carrying out comprehensive evaluation on the normalized acoustic emission data and electromagnetic radiation data by adopting an entropy method to obtain acoustic-electric comprehensive data; constructing a data set based on the sound-electricity comprehensive data;
And the prediction module is used for constructing an LSTM-Autoencoder model and obtaining an intelligent prediction result of the mine acousto-optic signal based on the data set and the LSTM-Autoencoder model.
The further implementation mode is that the data processing module comprises a data preprocessing unit, a specific gravity calculating unit, an information entropy calculating unit, a weight calculating unit, a comprehensive scoring unit and a sound and electricity comprehensive data acquiring unit;
The data preprocessing unit is used for taking the time-average acoustic emission intensity and the time-average acoustic emission ringing of acoustic emission data and the time-average electromagnetic intensity and the time-average electromagnetic pulse of electromagnetic radiation data as prediction indexes, and carrying out dimensionalization processing on the prediction indexes to obtain normalized data X i＝{xi1,xi2,...,xij, wherein X ij represents a j index corresponding to an i sample data;
A specific gravity calculation unit for calculating the specific gravity of each value in each index to all values based on the normalized data:
wherein n is the number of samples, and m is the number of indexes;
an information entropy calculation unit for calculating information entropy of each index based on the specific gravity:
If pij＝0,Ei＝0；
a weight calculation unit for calculating the index weight of each index based on the information entropy: k＝m；
a comprehensive scoring unit for calculating an index comprehensive score of each index based on the index weights:
The sound and electricity comprehensive data acquisition unit is used for acquiring sound and electricity comprehensive data based on the index comprehensive score: y= { Y 1,y2,...,yj }.
A further embodiment is that in the prediction module, the LSTM-Autoencoder model includes an input layer, an encoder, an intermediate layer, a decoder, and an output layer;
Encoder with a plurality of sensors For a dual-layer LSTM, each layer comprising a first hidden layer and a first embedded layer; the size of the embedded layer is smaller than the size of the input layer in order to obtain a compressed representation/>, of the intermediate layer
The decoder g θ (·) is a reverse double layer LSTM, each layer comprising a second hidden layer and a second embedded layer, reconstructing the output, contrary to the encoder setup
A further embodiment is that the prediction module comprises an encoding unit and a decoding unit;
An encoding unit for inputting input data to the encoder based on the input layer of the LSTM-Autoencoder model to obtain a compressed representation of the intermediate layer
Decoding unit for representing compressed representationThe input decoder decompresses and reconstructs the input data to obtain the prediction result/>
Wherein the input data has the format ofT is time; the input window is the length of each sample, the sliding step is to expand the sample size, and the output window is the signal length that needs to be predicted.
A further embodiment is that the parameters of the LSTM-Autoencoder model include input window, sliding step size, output window, hidden layer, embedded layer, learning rate, batch size, optimizer, and number of iterations.
A further embodiment of the system further includes an evaluation module configured to evaluate a predictive performance of the LSTM-Autoencoder model by reconstructing a difference between output data and input data of the loss quantization model, and minimize an L1 loss using a Mean Absolute Error (MAE) with a formula:
Particularly, according to the prediction result of the acoustic-electric signal output by the model, the embodiment also carries out next-step result analysis, a tracing module and an intelligent decision module are arranged, the tracing module is used for tracing the mine dangerous signal, further pertinently realizing rock burst danger early warning, and the intelligent decision module is used for carrying out intelligent decision based on the early warning result;
specifically, the tracing module comprises an anomaly detection unit, a feature map construction unit, a signal feature acquisition unit, a signal detection model construction unit, an anomaly signal determination unit and a tracing unit;
The abnormal detection unit is used for detecting abnormal signals based on the prediction result of the acoustic-electric signals:
the characteristic diagram construction unit is used for establishing a signal characteristic distribution diagram based on the normal acousto-electric signal;
the signal characteristic acquisition unit is used for carrying out characteristic extraction on the sound-electricity signal prediction result to obtain signal characteristics;
The signal detection model construction unit is used for constructing a signal detection model, and carrying out signal detection based on the signal detection model and the signal characteristics to obtain a detection result; the detection result comprises a normal signal and a suspicious signal;
the signal detection model comprises an identity mapping module and a convolution module;
the abnormal signal determining unit is used for analyzing the characteristic distribution of the suspicious signals and finally determining whether the suspicious signals are abnormal signals or not based on the signal characteristic distribution map;
And the tracing unit is used for tracing the signal based on the characteristic distribution of the abnormal signal. Besides time-frequency domain characteristics, the characteristic distribution also comprises the position of signal acquisition equipment (an acoustic-electric sensor) in a mine and the time period of signal acquisition, so that intelligent positioning of abnormal points of the mine and monitoring points about to be dangerous is realized.
Based on the discovery of the mine abnormal points, the decision module carries out intelligent decision of mine abnormal information processing: the prior abnormal condition data and corresponding coping strategies are collected through data analysis and mining, and the prior data (characters, images and the like) of different data structures are processed through a multi-mode model, so that a decision maker is helped to obtain prior comprehensive real information, and intelligent decision is provided.
Specifically, the intelligent decision module comprises a decision factor analysis unit, a classification unit, a decision unit and a visualization unit;
The decision problem analysis unit is used for generating a problem text based on the early warning result and intelligent positioning of mine abnormal points, and summarizing the problem text to generate a decision problem set;
The classification unit is used for classifying the decision problem set to obtain an abnormal condition classification result;
Specifically, based on a naive bayes algorithm, classifying the problem texts in the decision problem set:
acquiring text characteristics of a question text, wherein the characteristic components comprise: the name and code number of a specific monitoring point, the node position information of a monitoring sensor, abnormal values of an acousto-electric signal, a distribution diagram and the like;
Calculating class probabilities based on the existing decision problem set and text features;
Based on the category probabilities, a classification result of the abnormal situation is obtained.
The decision unit is used for obtaining a decision result based on priori knowledge and abnormal condition classification results;
the visualization unit is used for visualizing the decision result, and based on the decision result, the intelligent prediction of the mine acousto-electric signal is realized through the prediction module again, and the feasibility of the decision result is judged.
Particularly, the system also comprises a history record module for recording early warning information, abnormal point positioning information and decision results of each time and generating a history record database for combining with priori knowledge to jointly realize intelligent decision of the abnormal acoustic-electric signals of the next mine.
The above embodiments are merely illustrative of the preferred embodiments of the present invention, and the scope of the present invention is not limited thereto, but various modifications and improvements made by those skilled in the art to which the present invention pertains are made without departing from the spirit of the present invention, and all modifications and improvements fall within the scope of the present invention as defined in the appended claims.
Claims (8)
1. An intelligent prediction method for a mine acoustic-electric signal is characterized by comprising the following steps:
s1: acquiring mine acoustic emission data and electromagnetic radiation data of rock burst dangers;
S2: normalizing the acoustic emission data and the electromagnetic radiation data, and comprehensively evaluating the normalized acoustic emission data and electromagnetic radiation data by adopting an entropy method to obtain acoustic-electric comprehensive data; constructing a data set based on the sound-electricity comprehensive data;
s3: and constructing an LSTM-Autoencoder model, and obtaining an intelligent prediction result of the mine acousto-optic signal based on the data set and the LSTM-Autoencoder model.
2. The intelligent prediction method of mine sound and electricity signals according to claim 1, wherein in step S2, the method for obtaining sound and electricity integrated data is as follows:
Taking the time-average acoustic emission intensity and the time-average acoustic emission ringing of the acoustic emission data and the time-average electromagnetic intensity and the time-average electromagnetic pulse of the electromagnetic radiation data as prediction indexes, and carrying out dimensionalization processing on the prediction indexes to obtain normalized data X i＝{xi1,xi2,...,xij, wherein X ij represents a j index corresponding to an i sample data;
calculating the proportion of each value in each index to all values based on the normalized data:
wherein n is the number of samples, and m is the number of indexes;
calculating information entropy of each index based on the specific gravity:
When p ij =0, E i =0;
calculating index weights of the indexes based on the information entropy: k＝m；
Calculating the index comprehensive score of each index based on the index weight:
based on the index comprehensive score, obtaining the sound-electricity comprehensive data: y= { Y 1,y2,...,yj }.
3. The intelligent prediction method of mine acousto-electric signals according to claim 1, wherein in step S3, the LSTM-Autoencoder model includes an input layer, an encoder, an intermediate layer, a decoder, and an output layer;
The encoder is a double-layer LSTM, and each layer comprises a first hidden layer and a first embedded layer;
the decoder is a reverse double layer LSTM, each layer comprising a second hidden layer and a second embedded layer, opposite to the encoder arrangement.
4. The intelligent prediction method of the mine acoustic-electric signal according to claim 3, wherein in step S3, the method for obtaining the acoustic-electric signal prediction result based on the LSTM-Autoencoder model is as follows:
inputting input data to the encoder based on the input layer, obtaining a compressed representation of the intermediate layer Representing the compression/>The input decoder decompresses and reconstructs the input data to obtain the prediction result of the acoustic-electric signal
Wherein the input data has the format ofT is time.
5. The method of claim 1, wherein the parameters of the LSTM-Autoencoder model include an input window, a sliding step size, an output window, a hidden layer, an embedded layer, a learning rate, a batch size, an optimizer, and a number of iterations.
6. The mine acoustoelectric signal intelligent prediction method of claim 1, wherein the method further comprises S4: and evaluating the prediction performance of the LSTM-Autoencoder model by reconstructing the difference between output data and input data of the loss quantization model, wherein the formula is as follows:
7. An intelligent prediction system for a mine acoustic-electric signal, wherein the prediction system applies the prediction method of any one of claims 1-6, and the intelligent prediction system comprises a data acquisition module, a data processing module and a prediction module;
The data acquisition module is used for acquiring mine acoustic emission data and electromagnetic radiation data of rock burst dangers;
the data processing module is used for carrying out normalization processing on the acoustic emission data and the electromagnetic radiation data, and carrying out comprehensive evaluation on the normalized acoustic emission data and electromagnetic radiation data by adopting an entropy method to obtain acoustic-electric comprehensive data; constructing a data set based on the sound-electricity comprehensive data;
The prediction module is used for constructing an LSTM-Autoencoder model and obtaining an intelligent prediction result of the mine acoustic-electric signal based on the data set and the LSTM-Autoencoder model.
8. The mine acoustoelectric signal intelligent prediction system of claim 7, wherein the LSTM-Autoencoder model comprises an input layer, an encoder, an intermediate layer, a decoder, and an output layer;
The encoder is a double-layer LSTM, and each layer comprises a first hidden layer and a first embedded layer;
the decoder is a reverse double layer LSTM, each layer comprising a second hidden layer and a second embedded layer, opposite to the encoder arrangement.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311564610.XA CN117345344B (en) | 2023-11-22 | 2023-11-22 | Intelligent prediction method and system for mine acoustic and electric signals |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311564610.XA CN117345344B (en) | 2023-11-22 | 2023-11-22 | Intelligent prediction method and system for mine acoustic and electric signals |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117345344A CN117345344A (en) | 2024-01-05 |
| CN117345344B true CN117345344B (en) | 2024-05-14 |
Family
ID=89365182
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311564610.XA Active CN117345344B (en) | 2023-11-22 | 2023-11-22 | Intelligent prediction method and system for mine acoustic and electric signals |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117345344B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101762830A (en) * | 2009-09-29 | 2010-06-30 | 中国矿业大学 | Distributed coal mine rock burst monitoring method |
| CN104088668A (en) * | 2014-06-30 | 2014-10-08 | 中国矿业大学 | Ultra-low frequency electromagnetic induction monitoring and early warning system and method for coal or rock dynamic disasters |
| CN106437853A (en) * | 2016-09-27 | 2017-02-22 | 西安科技大学 | Method for early warning against coal rock burst dynamic disasters |
| 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 |
-
2023
- 2023-11-22 CN CN202311564610.XA patent/CN117345344B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101762830A (en) * | 2009-09-29 | 2010-06-30 | 中国矿业大学 | Distributed coal mine rock burst monitoring method |
| CN104088668A (en) * | 2014-06-30 | 2014-10-08 | 中国矿业大学 | Ultra-low frequency electromagnetic induction monitoring and early warning system and method for coal or rock dynamic disasters |
| CN106437853A (en) * | 2016-09-27 | 2017-02-22 | 西安科技大学 | Method for early warning against coal rock burst dynamic disasters |
| 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 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117345344A (en) | 2024-01-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106888205B (en) | Non-invasive PLC anomaly detection method based on power consumption analysis | |
| JP5145417B2 (en) | Signal determination method, signal determination apparatus, program, and signal determination system | |
| CN106198749B (en) | A kind of data fusion method of the multiple sensor based on Metal Crack monitoring | |
| CN112884747A (en) | Automatic bridge crack detection system integrating cyclic residual convolution and context extractor network | |
| Son et al. | Deep learning-based anomaly detection to classify inaccurate data and damaged condition of a cable-stayed bridge | |
| Li et al. | Automatic recognition and classification of microseismic waveforms based on computer vision | |
| Qin et al. | Remaining useful life prediction for rotating machinery based on optimal degradation indicator | |
| CN117726181B (en) | Collaborative fusion and hierarchical prediction method for typical disaster risk heterogeneous information of coal mine | |
| CN111520192A (en) | Non-contact tunnel engineering construction rock burst real-time forecasting optimization method | |
| CN110287827A (en) | A kind of bridge strain data outliers recognition methods based on data correlation | |
| Rosso et al. | Review on deep learning in structural health monitoring | |
| CN116861369A (en) | Air compressor health assessment method and system for industrial internet heterogeneous data fusion | |
| CN114679310A (en) | Network information security detection method | |
| CN115128663A (en) | Mine safety production detection system based on Internet of things and detection method thereof | |
| CN111505064A (en) | Catalytic combustion type methane sensor service state evaluation method | |
| CN114124420A (en) | Network flow abnormity detection method based on deep neural network | |
| CN104504265B (en) | A kind of method of servicing bridges monitoring information security evaluation | |
| CN117345344B (en) | Intelligent prediction method and system for mine acoustic and electric signals | |
| CN116975728A (en) | Safety management method and system for coal bed methane drilling engineering | |
| US20230024104A1 (en) | Identification of false transformer humming using machine learning | |
| CN114580472B (en) | Large-scale equipment fault prediction method with repeated cause and effect and attention in industrial internet | |
| CN113139681A (en) | Neural network rock burst prediction method based on time series data | |
| Yi et al. | An Uncertainty-Aware Deep Learning Model for Reliable Detection of Steel Wire Rope Defects | |
| Gao et al. | Anomaly detection for time series with difference rate sample entropy and generative adversarial networks | |
| Talebi-Kalaleh et al. | Damage Detection in Bridge Structures through Compressed Sensing of Crowdsourced Smartphone Data |
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 |