You are on page 1of 6

Arabian Journal of Geosciences (2020) 13:852

https://doi.org/10.1007/s12517-020-05833-6

GMGDA 2019

The prediction of water conducted zone in coal mining by Internet


of things perception
Zhenguo Yan 1 & Xintan Chang 1 & Yanping Wang 1

Received: 15 May 2020 / Accepted: 12 August 2020


# Saudi Society for Geosciences 2020

Abstract
To ensure that water resources are not damaged during coal mining, the height of the water conducted zone in coal mining is
explored in this study. When mining coal mines, the height of the water conducted zone is complex and unstable when making
predictions. Therefore, the Internet of things perception technology is applied to solve the problem of inaccurate data collection.
Moreover, the multiple regression analytical method is introduced to obtain the prediction equation for the height of the water
conducted zone. The research results show as follows: by using the Internet of things sensing technology, when collecting data,
the redundant and error data can be excluded before the data fusion, which can improve the validity of the data. The multiple
regression analysis methods are used to summarize and conclude the data by using the linear regression method; therefore, the
obtained prediction equation can accurately predict the height of the water conducted zone, which solves the problem caused by
using traditional empirical equations in coal mining in the past that leads to a large error in the prediction height. By using the
Internet of things perception technology to predict the height of the water conducted zone in coal mining, the collected data are
more accurate and the prediction results are more effective. Therefore, this method can effectively protect water resources during
coal mining.

Keywords Water conducted zone . Perception technology . Multiple regression method . Linear regression

Introduction Research on the development law and height of the water


conducted zone is a prerequisite to ensure the safety of coal
During coal mining, the overlying strata on the coal seam will mining operations (Liu et al. 2017a). Scholars worldwide have
also be affected, and the phenomenon of movement and dam- studied the development law and height of the water conduct-
age will occur, resulting in deformation. The degree of these ed zone; in the meantime, studies in this area started earlier in
damages can be divided into fall zones, fissure zones, and foreign countries of China. Globally, scholars have put for-
curved sinking zones, in which the falling and fissure zones ward the “arch theory,” “cantilever theory,” “zone division
are called water conducted zones. Because the water conduct- theory,” and “natural slope theory.” These theories are all
ed zone is at the lowest level, the development of the water the results of research on the development of water conducted
conducted zone and the prediction of its height can be used to zone height (Yang and Lv 2017). Some other scholars have
predict the amount of water in the mine and the retention plan put forward a summary of the prediction methods for coal
of the safety coal pillars. In addition, it can also predict the mining subsidence. In the twentieth century, the British
safety of mining work under high water content coal seams. Mining Bureau also covered the overburden composition,
coal seam thickness, and mining methods during coal mining
under the sea in the relevant requirements (Wang et al. 2018).
This article is part of the Topical Collection on Geological Modeling and In China, scholars have combined the mining thickness of coal
Geospatial Data Analysis
seams and the strength types of rock masses in the early days
and obtained the relationship between coal seam thickness,
* Zhenguo Yan
zhenguoyan666@126.com crack height, and different rock mass conditions (Wang et al.
2017a). In the 1980s, Chinese scholars proposed the concept
1
of water conducted zone and caused domestic scholars to
College of Safety Science and Engineering, Xi’an University of
Science and Technology, Xi’an 710054, Shaanxi, China
study it. At the same time, the scholar also proposed related
852 Page 2 of 6 Arab J Geosci (2020) 13:852

equations to calculate its height (Beneš 2016). In recent years, information and data. The relationship between these data is
scientists in China have done a lot of research on the proper- intricate and difficult to be directly used; therefore, it is nec-
ties of rock masses and the development law of water con- essary to organize these data. The perceptual information fu-
ducted zones. sion technology is one of the core technologies of IoT tech-
With the continuous progress of society and the continuous nology. It can integrate and process the IoT data in various
development of science and technology, the Internet of things aspects and between different levels, and filter out and extract
(IoT) has also quickly attracted widespread attention from useful information from it. The method of acquiring data by a
scholars worldwide, and many scholars have studied it. single sensor is more accurate; at the same time, the relation-
American scholars have combined radio frequency identifica- ship between the data can be better obtained, and the advan-
tion (RFID) technology with the Internet to give the IoT the tage of complementary information is achieved (Mathapelo
function of identification and management (Wang and Meng et al. 2016).
2018). Furthermore, scholars have conducted in-depth re-
search on this method so that the IoT has the functions of Perception information fusion technology
object identification, tracking, and monitoring, making it ca-
pable of management work. Later, scholars have redefined the Information fusion technology is the process of data fusion.
IoT, emphasized the characteristics of the IoT, and proposed The data collected in a single sensor is used by various de-
that the IoT has intelligent functions (Wang et al. 2017b). vices. Computing technology analyzes information under cer-
Chinese scholars define the IoT as a type of sensor and sensor tain rules, comprehensively processes it, and obtains the re-
network to perceive complex information to get the informa- quired information. In the process of information acquisition,
tion that people want (Słota et al. 2016). The IoT technology this technology can continuously repair itself to make the final
has also been promoted to the coal industry. However, this obtained result more perfect and accurate (Duan et al. 2018).
technology is only used as a safety monitoring work and part Because the sensors in the system work independently when a
of the high-level data prediction work in coal mine operations, single sensor is damaged and cannot continue to work, it can-
while the application is not ideal. Usually, in China, the inte- not affect the work of other sensors; therefore, the network has
gration between the systems is ignored, which leads to their a stable working performance; second, because the sensors in
inability to cooperate well, and the data cannot be effectively the system are not unique, it is necessary to use geometric
calculated and applied during work. methods to form their conductive aperture when working so
According to the existing research, this study conducted in- that the spatial resolution of the information and data fusion is
depth research on the development law and height of the water effectively improved; the process of data fusion can avoid the
conducted zone. The IoT perception technology is used to influence of inaccurate information obtained by data errors
collect data from coal mining, which is more accurate and and improve the accuracy of the information; meanwhile, be-
effective for the height prediction of the water conducted cause the data fusion is performed in the case of multiple
zone. In addition, the linear prediction method in multiple sensors, the acquired information is better than that obtained
regression is used to derive the height prediction equation of by a single sensor. The information obtained is more compre-
the water conducted zone. hensive, preventing excessive errors of information data (Gao
et al. 2015). Data fusion is to pick out the data that users are
interested in and send it to users. Therefore, these data flows
Methods need to be managed. The management process is shown in
Fig. 1.
Application of the IoT technology in the prediction of The fusion of multi-sensor information data is a kind of
water conducted zone comprehensive analysis and processing based on certain
criteria; finally, it obtains valuable data. Considering informa-
IoT is a combination of various information sensing devices tion as abstract data, information fusion can be divided into
with the Internet and other technologies to form a huge net- three levels, i.e., the pixel-level fusion, feature-level fusion,
work. Its purpose is to combine sensors with the Internet to and decision-level fusion. The pixel-level fusion is the
identify and manage objects connected to the network. This lowest-level fusion, and the decision-level fusion is the
technology is virtual and implements the connection between highest-level fusion process.
people and things (Lu et al. 2017). IoT technology can per-
form real-time monitoring of various links in coal mining and Application of perception information fusion in
obtain huge engineering information data. It can comprehen- height prediction of water conducted zone
sively manage strict coal mining construction and have a de-
tailed understanding of the mining area. However, at the same When mining coal, the geological conditions and the working
time, IoT technology relies on massive amounts of environment are complex and changeable, which brings a lot
Arab J Geosci (2020) 13:852 Page 3 of 6 852

Sensor 1 Sensor 2 Sensor 3 Sensor n Gross error is a data error caused by unnecessary failures of
equipment. At present, there are mainly three methods for
eliminating error data, i.e., periodic method, Wright criterion
or Grubbs rule, and algorithms (Yan et al. 2016). Due to the
complex and diverse working environment of coal mining,
Storage normal load distribution cannot be achieved. The first two
methods cannot achieve the ideal result of eliminating errors.
Data
Sensor side
convergence
Furthermore, this study chooses to use algorithms to eliminate
storage errors. The algorithm selected in this study is the Kalman filter
point end storage
algorithm, which can be used in the calculation of various
linear, discrete, or finite-dimensional space systems. The
Kalman filter algorithm is to use the equations to estimate
the linear minimum variance of the probability density in the
Query
continuous updating process under the calculation state
Data (Arthur et al. 2016). This algorithm is to set the initial data
Sensor side
storage
convergence of the height of water conducted zone as the initial optimal
point end storage data. Then, it calculates the covariance at that moment. The
same calculation method is used to obtain the covariance at
the next state, and the two covariance variance data are opti-
mally compared. Therefore, the covariance data are continu-
Result
ously updated to obtain the optimal covariance data corre-
Fig. 1 Flowchart of data management sponding to the water conducted zone, which is output as
the result.
of inconvenience to the mining work. There is a great possi-
bility to cause pollution and damage to water resources when Data-level fusion
carrying out coal mining work. Therefore, the correct predic-
tion of the height and development law of the water conducted After preprocessing the measured data, it is necessary to per-
zone is very important to prevent and control water loss. The form data fusion. In this study, the Bayesian network algo-
factors affecting the high development of the water conducted rithm is selected to fuse the obtained complex data. The
zone are diverse; therefore, the data collection must be as Bayesian network algorithm is an algorithm that uses mathe-
comprehensive and accurate as possible. Then, the informa- matical probability networks. It infers a graphical network
tion fusion of the collected data is performed. The information based on an overview and processes and analyzes variable
fusion process is shown in Fig. 2. information data to obtain the size of the probability informa-
When predicting the height of the water conducted zone, tion. The calculation process of this method is shown in Fig. 3.
the collected data needs to be preprocessed, and the redundant
and error data in the information data should be processed and Application of multiple linear regression analysis
eliminated. methods in prediction of water conducted zone

Data preprocessing After processing and analyzing the data obtained through the
IoT technology, it is necessary to calculate the processed data
In the process of data preprocessing, the error data in the data and use multivariate linear regression analysis to derive the
are mainly eliminated, which are mainly the gross errors. equation for predicting the highly developed water conducted

Fig. 2 Flowchart of data fusion


Geographic
Decision
information Data fusion
perception
system
Sensor node

Factor Internet of Data


Remote acquisition Things Road preprocessing
sensing
instrument
852 Page 4 of 6 Arab J Geosci (2020) 13:852

Fig. 3 Process of Bayesian


network algorithm Sensor 1 Sensor 2 Sensor n

Kalman filter Kalman filter Kalman filter

Combined Bayesian formula

Data fusion

Result

zone. The multiple linear regression analysis methods are temperature range of 22–25 °C, and the information data col-
based on mathematics. It collects the mathematical relation- lected by the sensor is processed by the Kalman filter algo-
ships between the data, finds a regular change calculation rithm. Then, the preprocessed data is subjected to information
method of the independent variables that affect the changes, fusion, and the uncertain information in the obtained data can
and finally obtains expected results. Many factors affect the be processed. The Bayesian algorithm uses the probability
height development of the water conducted zone. However, inference method as the basis for information fusion. The data
based on the existing research results, this study mainly selects can be highly fused in the prediction of the height of the water
several factors, such as coal seam mining height, mining conducted zone in coal mining to obtain more accurate results.
depth, lithological proportion coefficient, and a working Using sensors to monitor the status of coal mining in real-time
length of the working interface, to explore their relationship can comprehensively understand the status of the mine, which
between the heights of water conducted zone (Liu et al. eliminates the need for the staff to go down to make measure-
2017b). These data are then analyzed and processed by com- ments and saves a lot of time and energy for the staff.
puter software to obtain the relationship between them and the
water conducted zone. The calculation equation of the multi-
Application of multiple regression analysis methods
ple linear regression algorithm is shown below.
in the prediction of height development of water
y ¼ β 0 þ β 1 x1 þ ⋯ þ β n x n ð1Þ conducted zone

where the value of β is a variable, which is the dependent The mining height, mining depth, rock mass proportionality
variable of the observed value and also the coefficient required factor, and working length of the four factors are mainly se-
this time. It is solved by the method of least squares. lected to determine the influence of the height development of
 2 the water conducted zone. First, the relationship between the
f ðxi Þ ¼ ∑ yi −byi ¼ ∑ðyi −β 0 −β 1 x1i −⋯−β n −xni Þ2 ¼ min ð2Þ height development of the water conducted zone and the min-
ing height is tested. The relationship between the water con-
ducted zone and the height before mining is tested in a range
of coal mining heights within 8 m. It is found that the mining
Results and analysis height began to change after 2 m and showed irregular chang-
es. However, from the perspective of the overall distribution,
Application of perception information fusion the overall trend is rising. When the mining height is 7 to 8 m,
technology in the prediction of height development the height development of the water conducted zone ap-
of water conducted zone proaches 100 m. The height development trend of water con-
ducted zone within the range of lithological scale coefficient
This method is used to eliminate the gross error of the data of 1 is calculated. It is found that the overall trend is that the
obtained by the sensor during preprocessing. The coal mining height of the water conducted zone increases with the increase
condition is monitored for a period of 100 min in the of the lithological scale factor. The height of the water
Arab J Geosci (2020) 13:852 Page 5 of 6 852

conducted zone varies greatly within the range of 0.4 to 0.6, As shown in Fig. 5, the predicted value of the height
and the height development of the water conducted zone development of the water conducted zone is closer to
shares the relationship with the oblique length and mining the actual value using the IoT technology researched in
depth, both of which change with the mining depth. The rela- this study combined with the multiple linear regression
tionship between the height development of the water con- analysis method. In addition, the predicted value obtain-
ducted zone and the mining depth is shown in Fig. 4. ed by this method is better than the traditional empirical
As shown in Fig. 4, the height development of the water equation, and the error of the value is smaller.
conducted zone will change non-linearly with the increase of Therefore, the method proposed in this study can better
the mining depth of the coal mine. The height of the water predict the height development trend of the water con-
conducted zone also reaches the highest when the mining ducted zone, reduce the prediction error, and improve
depth is 430 m. the prediction accuracy.
In addition, the propulsion speed also has a certain
effect on the height development of the water conducted
zone. However, because the propulsion speed is relatively
long, it is not involved in this study. Multivariate regres- Conclusions
sion analysis is performed on the relationship between the
above influencing factors and the water conducted zone to The research of this study shows that the IoT technology
obtain the corresponding linear relationship. This study can be used to detect and collect data in coal mining.
mainly analyzes these factors comprehensively and ana- Meanwhile, the Kalman filter algorithm can be used to
lyzes the relationship between the comprehensive factors process the information data collected by sensors. Then,
and the water conducted zone, as shown in the following the preprocessed data is subjected to information fusion,
equation. and the uncertain information in the obtained data can be
  processed. The Bayesian algorithm uses the probability
426:243
H ¼ 3:41M þ 27:12b þ 1:85Inl þ 0:13 exp 5:346− þ 0:64v þ 6:11 inference method as the basis for information fusion.
S
The data can be highly fused in the prediction of the
ð3Þ
height of the water conducted zone in coal mining to
where M is the mining height of coal mine operation, b is obtain more accurate results. Using sensors to monitor
the lithological proportion coefficient, l is the length of the status of coal mining in real time can comprehensively
the working face during coal mine operation, s is the understand the status of the mine, which eliminates the
mining depth, and v is the advancement speed. need for the staff to go down to make measurements
The obtained prediction equation of the water con- and saves a lot of time and energy for the staff. Then,
ducted zone is used to predict the height development the multiple linear regression method is used to obtain
of the water conducted zone in actual coal mining. The the prediction equation of the water conducted zone on
predicted value is compared with the result obtained by the obtained data. After monitoring the actual engineering
using the traditional empirical equation. The analysis data, it is found that the prediction value of the height
result is shown in Fig. 5. development of the water conducted zone by the IoT tech-
nology and multiple linear regression analysis methods is
70 smaller than the prediction value obtained by the tradi-
Height of water flowing fractured zone

tional empirical equation. Therefore, it can be used to


60 accurately predict the height development of water con-
ducted zone.
50
There are still some deficiencies in this study.
Currently, there are few studies on the prediction of
water conducted zones with this technology. Therefore,
40
there are some theoretical shortcomings. At the same
time, due to time and space constraints, in this study,
30 only the factors influencing the height development of
the water conducted zone are selected, which is not
20 comprehensive enough. It is hoped that in the subse-
quent study, researchers can consider the factors that
350 400 450 500 550 600 650 700 affect the height of the water conducted zone more
Mining depth value comprehensively and obtain more accurate prediction
Fig. 4 Relationship between water conducted zone and mining depth results.
852 Page 6 of 6 Arab J Geosci (2020) 13:852

Fig. 5 Comparison results of Height value of water flowing fractured zone


water conducted zone heights Error value
under different prediction
60
methods

50

40

Value
30

20

10

0
Measured value Empirical formula Prediction formula
Type

References Mathapelo PS, McCrindle RI, Combrinck S et al (2016) Hazard assess-


ment of polycyclic aromatic hydrocarbons in water and sediment in
the vicinity of coalmines. Journal of Soils & Sediments 16(12):1–13
Arthur LM, Weakley AT, Griffiths PR et al (2016) Direct-on-filter ?-
Słota K, Słota Z, Kułagowska E (2016) Thermal load at workstations in
quartz estimation in respirable coal mine dust using transmission
the underground coal mining: results of research carried out in 6 coal
Fourier transform infrared spectrometry and partial least squares
mines. Med Pr 67(4):477–498
regression. Appl Spectrosc 71(5):1014
Wang X, Meng FB (2018) Statistical analysis of large accidents in
Beneš M, Krupička L. Weak solutions of coupled dual porosity flows in China’s coal mines in 2016. Natural Hazards Journal of the
fractured rock mass and structured porous media. 2016, 433(1), pp. International Society for the Prevention & Mitigation of Natural
543–565
Hazards 92(4):311–325
Duan ZH, Wu TTH, Guo SW et al (2018) Development and trend of Wang XL, Qin QR, Fan CH (2017a) Failure characteristic and fracture
condition monitoring and fault diagnosis of multi-sensors informa- evolution law of overburden of thick coal in fully mechanized sub-
tion fusion for rolling bearings: a review. Int J Adv Manuf Technol level caving mining. J Eng Geol 46(11):2041–2048
96(1–4):803–819 Wang W, Wang Q, Sohraby K (2017b) Multimedia sensing as a service
Gao XD, Sun Y, You DY et al (2015) Multi-sensor information fusion for (MSaaS), pp. Exploring resource saving potentials of at Cloud-Edge
monitoring disk laser welding. Int J Adv Manuf Technol 85(5–8): IoT and fogs. IEEE Internet Things J 4(2):487–495
1167–1175 Wang ZC, Wang C, Wang ZC (2018) The hazard analysis of water inrush
Liu SL, Li WP, Wang QQ (2017a) Height of the water-flowing fractured of mining of thick coal seam under reservoir based on entropy
zone of the Jurassic coal seam in Northwestern China. Mine Water weight evaluation method. Geotech Geol Eng 36(5):3019–3028
& the Environment 37(3):312–321 Yan PC, Zhou MR, Liu QM, Zhang KY, He CY (2016) Research on the
Liu C, Jin Gong E et al (2017b) Prediction for the performance of gas source identification of mine water inrush based on LIF technology
turbine units using multiple linear regression. Proceedings of the and SIMCA algorithm. Spectroscopy & Spectral Analysis 36(1):
Csee 37(16):4731–4738 243–247
Lu T, Liu SD, Wang B et al (2017) A review of geophysical exploration Yang Y, Lv P (2017) Comprehensive determination of the height of water
technology for mine water disaster in China: applications and trends. flowing fractured zone in super close upward mining. Liaoning
Mine Water & the Environment 36(5):1–10 Gongcheng Jishu Daxue Xuebao 36(4):371–376

You might also like