KR101956294B1 - Stand-alone automatic emergency alarm system and method for mine safety management based on machine learning - Google Patents

Abstract

A machine learning based self-contained automatic emergency alert system and method for mine safety management. The emergency alarm system for the mine safety management according to the embodiment collects and analyzes the experimental data of the occurrence of the dangerous situation in the mine due to the vibration generated in the mine, and analyzes the magnitude of the vibration data, A data learning module for setting vibration data detail information including a cause of occurrence and a dangerous situation likely to occur due to vibration; A sensing module for sensing vibration data generated by an artificial or natural factor; An alarm generation module for comparing the vibration data detail information set in the data learning module with the sensed vibration data and generating a danger alarm based on the comparison result; .

Description

TECHNICAL FIELD [0001] The present invention relates to a self-standing automatic emergency alert system and method for mine safety management,

And more particularly, to an apparatus and method for automatically informing a dangerous signal after a danger signal for mine safety management.

Unless otherwise indicated herein, the contents set forth in this section are not prior art to the claims of this application and are not to be construed as prior art to be included in this section.

100% of domestic mines are located in remote mountains, and it is difficult to recruit new and skilled workers due to poor working environment compared to other industries, and the pace of aging of existing workers is gradually accelerating. In addition, disasters and major accidents continue to occur in mines. Among the 32 cases of mine disasters in 2015, 10 cases (31%) of collapses were reported. In recent years, 3 large-scale earthquake offsets in domestic mines (Uljin limestone mine in Gyeongbuk, Samcheok Baekun limestone mine in Gangwon province, Two slope failures (Gangneung Okpyeong La Paz, Halla Cement, Gyeonggi Gapyeong International Mine) occurred.

In order to prevent disasters and safety accidents that continue to occur in mines, it is necessary to develop a mine development safety management system nationwide. In addition, there is a growing need for the development of disaster prevention technologies for mines that have a higher rate of disasters than other industries, due to the increase in general mines that convert to mining in relation to environmental protection and the deepening of the workplace. Therefore, it is essential to improve the safety of mining operations and to minimize the economic damage caused by disasters by developing an emergency alarm system using machine learning technology, which is the fourth industrial technology.

1. Korean Patent Publication No. 10-2014-0080450 (June 30, 2014) 2. Korean Patent Registration No. 10-2013-0053988 (2013.05.13)

(For example, cracks such as discontinuities and joints) and artificial danger signals (for example, mine equipment movement and perforation, blasting, impact due to removal of ceiling pits, etc.) And provides an alarm device and method for mine safety management that provides preliminary predictions of the risk of collapse of mines and automatic emergency alert through mine data processing / analysis / learning based on machine learning.

The automatic running emergency alarm system based on machine learning for mine safety management according to the embodiment collects and analyzes the experimental data of a dangerous situation in the mine due to the vibration generated in the mine and analyzes the vibration data which may cause dangerous situation in the mine A data learning module that sets vibration data detail information including a size, a cause of occurrence of vibration data, and a dangerous situation that may occur due to vibration; A sensing module for sensing vibration data generated by an artificial or natural factor; An alarm generation module for comparing the vibration data detail information set in the data learning module with the sensed vibration data and generating a danger alarm based on the comparison result; .

A machine learning based automatic emergency alert method for mine safety management according to another embodiment is characterized in that (A) a data learning module collects and analyzes the vibration data generated in the mine, Setting vibration data detail information including a size of vibration data that may cause a dangerous situation, a cause of occurrence of vibration data, and a risk situation that may occur due to vibration; (B) sensing vibration data, which is installed on the inner wall of the mine in the sensing module and is generated artificially or caused by natural factors; And (C) comparing the vibration data detailed information set in the data learning module with the sensed vibration data in the alarm generating module and generating a danger alarm based on the comparison result; .

The emergency alarm apparatus and method as described above can provide a clear emergency alarm to the mine workers through the sound device and the light emitting device when detecting the mine collapse precursor phenomenon in the mine to prevent the safety accident.

Analysis of the danger signal of the emergency alarm device itself According to the final result, it can be classified as good, attention, and danger, and the risk level can be classified into the acoustic type and the luminescent color.

The present disclosure can greatly reduce the time required for alarming a dangerous situation, and it is possible to improve the safety of the worker by predicting the dangerous signs more precisely and quickly by installing a large number of machine learning based automatic emergency alarm devices inside the mine, Can be minimized.

It should be understood that the effects of the present invention are not limited to the above effects and include all effects that can be deduced from the detailed description of the present invention or the configuration of the invention described in the claims.

1 is a view showing a configuration of a mining risk warning device according to an embodiment;
2 is a view showing a schematic configuration of a mine risk warning device according to another embodiment
3 is a block diagram showing a more specific configuration of the emergency alarm device according to the embodiment
4 is a diagram showing a data processing flow of the mining emergency alarm method according to the embodiment

BRIEF DESCRIPTION OF THE DRAWINGS The advantages and features of the present invention and the manner of achieving them will become apparent with reference to the embodiments described in detail below with reference to the accompanying drawings. The present invention may, however, be embodied in many different forms and should not be construed as being limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the concept of the invention to those skilled in the art. Is provided to fully convey the scope of the invention to those skilled in the art, and the invention is only defined by the scope of the claims. Like numbers refer to like elements throughout.

In the following description of the present invention, a detailed description of known functions and configurations incorporated herein will be omitted when it may make the subject matter of the present invention rather unclear. The following terms are defined in consideration of the functions in the embodiments of the present invention, which may vary depending on the intention of the user, the intention or the custom of the operator. Therefore, the definition should be based on the contents throughout this specification.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a view showing a configuration of a mining risk warning device according to an embodiment. FIG.

Referring to FIG. 1, the mine risk warning system may include a sensor data collection unit, a machine learning algorithm based data processing unit, a data analysis unit, and an alarm unit.

The sensor data collection unit collects dangerous signals in the mine in real time and collects various experimental data to identify the cause of the risk in the mine. Hazardous signals collected by the sensor data collection unit are caused by artificial causes such as natural hazards such as cracks in discontinuities and joints and natural hazards such as mine equipment movement, perforation, blasting, It contains an artificial risk signal. Experimental data may include experimental data from relevant agencies to correlate the hazard signal from the mine with the factors that generate the hazard signal.

The data processing unit based on the machine learning algorithm uses an RNN (Recurrent Neural Network) that generates a machine learning based classification model to output the results of various risk signal generating factors generated in the mine through the accumulated data sets in the sensor data collecting unit Automatically analyze the data characteristics to learn the correlation between the hazard signal and the risk signal. In addition, the data set is analyzed to calculate the probability of occurrence of accidents based on the type and size of the dangerous signal.

The data analysis part analyzes the risk signal detected in the mine and identifies the risk signal according to the risk signal based on the correlation learned in the data processing part based on the machine learning algorithm.

The alarm section outputs the warning signs detected by the data analysis section to the mine as an output means such as a speaker or an emergency bell. For example, the alarm unit can transmit an accurate risk situation to workers in the mine by varying the display color or the sound volume output according to the degree of danger.

In an embodiment, the mine risk warning device may operate with an autonomous power supply, without batteries or batteries connected to the battery. The battery is connected in parallel with a secondary battery that supplies power to the mine hazard warning device, or it alerts when the remaining battery level is at risk, thereby causing malfunction of the mine risk warning device due to power shortage, Prevent situations where you can not sense signals.

2 is a view showing a schematic configuration of a mining risk warning device according to another embodiment.

Referring to FIG. 2, the mine risk warning apparatus 100 may include a data learning module 110, a sensing module 130, and an alarm generation module 150. The term " module ", as used herein, should be interpreted to include software, hardware, or a combination thereof, depending on the context in which the term is used. For example, the software may be machine language, firmware, embedded code, and application software. As another example, the hardware may be a circuit, a processor, a computer, an integrated circuit, an integrated circuit core, a sensor, a micro-electro-mechanical system (MEMS), a passive device, or a combination thereof.

The data learning module 110 collects and analyzes the vibration data and the experimental data generated within the mine, and sets detailed information of the vibration data that may cause a dangerous situation in the mine. The details of the vibration data may be the metadata of the vibration, such as the magnitude and frequency of the vibration data, the cause of the vibration, and the dangerous situation possibly caused by the vibration. Experimental data are experimental data carried out by related organizations in order to clarify the cause of the risk situation in the mine. The vibration data including the kind of natural vibration that can cause a specific risk situation, vibration type including amplitude, size and frequency And information on the correlation between information and risk signal generation in the mine, including cracks, cracks, and vibrations.

The sensing module 130 is installed on the inner wall of the mine, and detects vibration data generated by an artificial or natural factor. In an embodiment, the sensing module 130 may comprise a micro vibration sensor, an ultra-small, ultra-precise, ultra-light acceleration sensor of three axes (x, y, z)

The alarm generating module 150 compares the vibration data detailed information set in the data learning module 150 with the sensed vibration data and generates a danger alarm based on the comparison result. The alarm generating module 150 may generate alarms according to the similarity by comparing the vibration data detailed information set in the data learning module 110 with the sensing data. For example, the risk alarm proportional to the similarity between the vibration data detail information and the sensing data can be classified into good, cautious, and dangerous, and the danger warning can be generated by dividing into a sound type and a luminescent color.

3 is a block diagram showing a more specific configuration of the emergency alarm device according to the embodiment.

3, the data learning module 110 may include a database, a data classifier 111, and a data learning unit 113. The sensing module 130 may include an artificial generation vibration detector 131, And a naturally occurring vibration detection unit 133. The alarm generation module 150 may include a comparison unit 151 and an alarm generation unit 153. [

The database of the data learning module 110 stores sensing data necessary for machine learning to grasp the occurrence of a mining hazard signal and its cause and a series of data necessary for driving the emergency alarm device.

The data classification unit 111 classifies the vibration data into a naturally occurring vibration or an artificially generated vibration according to the cause of vibration occurrence. Naturally occurring vibrations are caused by cracks caused by self-aging of mines, vibration caused by cracks due to discontinuities or joints, vibrations caused by gravity-induced dropouts, and vibrations caused by collapses in mines. And vibration data generated by the vibration data. The artificial vibration is caused by the mining work in the mine, the vibration caused by the movement of equipment in the mine, the vibration caused by the mining drilling for the blasting work, the vibration caused by the blasting impact, And the vibration generated by the mining worker and the machine, such as the vibration caused by the excavation work in the mine.

The data learning unit 113 learns the correlation between the artificial generated vibration and the natural generated vibration based on the experimental data, and correlates the artificial generated vibration and the generated natural vibration with the cause of the danger signal using various machine learning algorithms such as RNN Learning. For example, the data learning unit 113 analyzes data characteristics using an RNN (Recurrent Neural Network), performs signal classification and pattern recognition, and calculates the amplitude of artificial generated vibration causing naturally occurring vibrations, It is possible to deduce the cause information of the natural vibration occurrence including the cycle, the cause of the artificial vibration, the number of artificial vibration occurrences, and the cumulative amplitude of the artificial vibration. The data learning unit 113 also learns about the cause of the mine accident by calculating the probability that the naturally occurring vibration will cause an accident including mine collapse or the like. That is, the correlation between the artificial generated vibration and the spontaneous generated vibration may include a causal relationship between the natural occurring vibration and the artificial generated vibration, and a causal relationship between the generated natural vibration and the danger signal generation.

 The artificial generated vibration sensing unit 131 of the sensing module 130 senses and converts the vibration generated by an artificially generated vibration among the vibrations generated in the mine into data, The vibration generated by the natural occurrence factor of the vibration is detected and converted into data.

The comparing unit 131 of the alarm generating module 150 compares the artificial generated vibration and naturally occurring vibration data sensed by the sensing module 130 and the vibration data generated by the data learning module 110 Compare the information. For example, when the data learning module 110 deduces a machine learning result that causes a specific risk signal when vibration having an amplitude equal to or greater than a predetermined magnitude occurs periodically, the comparator 131 compares the amplitude of the sensed vibration data The period and the amplitude and period of the vibration data deduced by the data learning module 110 can be compared.

The alarm generating unit 153 generates a danger notification alarm according to a result of comparing the sensing vibration data with the vibration data that is a cause of the danger signal generation in the mine. The alarm generating unit 153 can generate a danger alarm based on the similarity, which is one of the comparison results between the two data. Since the likelihood of a danger signal increases when the similarity of the two data is higher than a certain level, it is possible to classify the risk level into good, attention, and danger levels in proportion to the degree of similarity. This can provide clear emergency alarms to mine workers through sound devices (speakers, emergency bells, sirens, etc.) and light emitting devices when detecting the mine decay event.

Hereinafter, the emergency alarm method will be described in turn. Since the function (function) of the emergency alarm method according to the present invention is essentially the same as that of the emergency alarm apparatus and system, the description overlapping with those of FIGS. 1 to 3 will be omitted.

4 is a view showing a data processing flow of the mining emergency alarm method according to the embodiment.

In step S410, the data learning module 110 collects and analyzes the vibration data generated in the mine and vibration data generated in the mine due to the vibration, analyzes the vibration data using the RNN and other machine learning algorithms, .

In step S420, the data learning module 110 calculates the magnitude of the vibration data that may cause a dangerous situation in the mine, the cause of the vibration data, and the dangerous situation that may occur due to the vibration based on the data learned in step S410 And sets the vibration data detail information to be included. Specifically, in step S420, experimental data for detecting the cause of the danger signal including cracks, cracks, and discontinuities of the mine are stored, and the vibration data is classified into naturally occurring vibration or artificially generated vibration according to the cause of the generated vibration data. Thereafter, the correlation between the artificial generated vibration and the natural occurring vibration is learned based on the experimental data, and the correlation between the artificial generated vibration and the natural generated vibration and the cause of the danger signal is learned, Details can be set.

In the embodiment, the correlation between the artificial generated vibration and the naturally occurring vibration includes a natural vibration including the amplitude and period of the artificial generated vibration causing the naturally occurring vibration, the cause of the artificial generated vibration, the number of times of occurrence of the artificial vibration, The probability of occurrence of a dangerous situation based on the cause data and the details of the naturally occurring vibration, and the like.

In step S430, the sensing module 130 detects the vibration data generated on the inner wall of the mine by artificial or natural factors.

In step S440, the alert generation module 150 performs a process of comparing the vibration data detailed information set in the data learning module 110 with the sensed vibration data.

In step S450, a risk alarm is generated based on the comparison result of the alert generation module 150. [ For example, in step S450, the alarm generating module 150 may generate alerts such as good, attention, and danger in proportion to the degree of similarity between the sensed vibration data and the set vibration detail information.

The emergency alarm apparatus and method as described above can provide a clear emergency alarm to the mine workers through the sound device and the light emitting device when detecting the mine collapse precursor phenomenon in the mine to prevent the safety accident.

Analysis of the danger signal of the emergency alarm device itself According to the final result, it can be classified as good, attention, and danger, and the risk level can be classified into the acoustic type and the luminescent color.

The present disclosure can greatly reduce the time required for alarming the danger situation and improve the safety of work through more precise and quick prediction of the danger signs when installing a large number of machine learning based automatic emergency alarm devices inside the mine, Can be minimized.

It is to be understood that the invention is not limited to the disclosed embodiments, but, on the contrary, is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims. It is not limited to the embodiment.

110: Data learning module
130: sensing module
150: Alarm generation module

Claims (12)

A machine-based, self-contained, automatic emergency alert system for mine safety management,
It is important to understand the magnitude of the vibration data, the cause of the vibration data, and the possible risk of vibration due to the vibration. A data learning module for setting the vibration data detail information to be included;
A sensing module for sensing vibration data generated artificially or by natural factors in a mine; And
An alarm generating module for comparing the vibration data detailed information set in the data learning module with the sensed vibration data and generating a danger alarm based on the comparison result; And
The data learning module
A database for storing experimental data for identifying the cause of the hazard signal generation in mines including cracks, cracks, and discontinuities;
A data classifying unit for classifying the vibration data into a naturally occurring vibration or an artificially generated vibration according to a cause of occurrence; And
A data learning unit that learns a correlation between an artificial generated vibration and a naturally occurring vibration based on experimental data, and learns a correlation between an artificial generated vibration and a naturally occurring vibration and a cause of the danger signal; And an emergency alarm device for mine safety management.
delete The method of claim 1, wherein the correlation between the artificial generated vibration and the naturally occurring vibration is
Information on reason of occurrence of natural vibration including the amplitude, cycle, cause of artificial vibration, number of artificial vibration, cumulative amplitude of artificial vibration, and inference and natural vibration of artificial generated vibration causing naturally occurring vibration include safety accident including mine collapse And the probability of occurrence of the mine accident.
2. The method of claim 1,
Vibration caused by mineral excavation work in the mine, vibration caused by moving equipment in the mine, vibration caused by mining drilling for blasting work, vibration caused by blasting impact, vibration caused by mining ceiling pumice removal and mine And a vibration generated by the excavation work of the mine.
2. The method of claim 1,
A vibration caused by a crack caused by self-aging of a mine, a vibration caused by a crack due to a discontinuity or a joint, a vibration caused by a dropping phenomenon due to gravity, and a vibration caused by a collapse in a mine, Emergency alarm for management.
The method according to claim 1, wherein the experimental data of occurrence of a hazardous situation in a mine due to the vibration
Correlation between vibration detail including amplitude, size and frequency of spontaneous and artificially generated vibrations and risk signal generation in mines including cracks, cracks, and vibrations, and the relationship between risk signal generation in mines and accident occurrence Wherein the test data are data for identifying the mines.
A machine-based, self-contained, automatic emergency alarm method for mine safety management,
(A) Vibration data generated inside the mine in the data learning module and the occurrence of dangerous situation in the mine due to vibration After collecting and analyzing the experimental data, the magnitude of the vibration data, which may cause dangerous situation in the mine, Setting vibration data detail information including a dangerous situation likely to occur due to vibration;
(B) sensing vibration data generated on an inner wall of the mine by a sensing module, the vibration data being generated artificially or caused by a natural factor; And
(C) comparing the vibration data detailed information set in the data learning module with the sensed vibration data in an alarm generating module and generating a danger alarm based on the comparison result; Including the
(A) setting the vibration data detail information;
Storing experimental data for identifying a cause of a danger signal including cracks, cracks, and discontinuities of a mine;
Classifying the vibration data into a naturally occurring vibration or an artificially generated vibration according to the cause of the vibration data; And
Learning correlation between artificial generated vibration and naturally occurring vibration based on experimental data, learning correlation between artificial generated vibration and naturally occurring vibration and cause of the danger signal; Wherein the emergency alert method for mine safety management comprises:
delete 8. The method of claim 7, wherein the correlation between the artificially generated vibration and the naturally occurring vibration is
Incidence of natural vibration including information on the amplitude, cycle, cause of artificial vibration, number of artificial vibration, and cumulative amplitude of artificial vibration, which cause spontaneous vibration And the probability of occurrence of an emergency alert for mine safety management.
8. The method according to claim 7, wherein the experimental data of occurrence of a hazardous situation in the mine due to the vibration
The correlation between the vibration details including the amplitude, size and frequency of spontaneous and artificially generated vibrations and the generation of danger signals in mines including cracks, cracks and vibrations is inferred and the probability of occurrence of accidents Wherein the test data is the test data to be calculated.
8. The method of claim 7, wherein the artificially generated vibration
Vibration caused by mineral excavation work in the mine, vibration caused by moving equipment in the mine, vibration caused by mining drilling for blasting work, vibration caused by blasting impact, vibration caused by mining ceiling pumice removal and mine And a vibration generated by the excavation work of the mine.
8. The method of claim 7,
A vibration caused by a crack caused by self-aging of a mine, a vibration caused by a crack due to a discontinuity or a joint, a vibration caused by a dropping phenomenon due to gravity, and a vibration caused by a collapse in a mine, Emergency alarm method for management.

Images