sign in sign up
<<
Home , Coal gas, Coal gasification, Linc Energy

sustainability

Review Monitoring and Control in Underground Gasification: Current Research Status and Future Perspective

Yuteng Xiao 1, Jihang Yin 1, Yifan Hu 1, Junzhe Wang 2, Hongsheng Yin 1,* and Honggang Qi 3

1 School of Information and Control Engineering, China University of Mining and Technology, Xuzhou 221116, China; [email protected] (Y.X.); [email protected] (J.Y.); [email protected] (Y.H.) 2 Bell Honors School, Nanjing University of Posts and Telecommunications, Nanjing 210023, China; [email protected] 3 School of Computer and Control Engineering, University of Chinese Academy of Sciences, Beijing 101408, China; [email protected] * Correspondence: [email protected]; Tel.: +86-137-7589-2769

 Received: 23 October 2018; Accepted: 29 December 2018; Published: 4 January 2019 

Abstract: By igniting in the coal seam and injecting gas agent, underground coal gasification (UCG) causes coal to undergo thermochemical reactions in situ and, thus, to be gasified into for power generation, production, and storage. Compared with traditional mining technology, UCG has the potential sustainable advantages in energy, environment, and the economy. The paper reviewed the development of UCG projects around the world and points out that UCG faces difficulties in the field of monitoring and control in UCG. It is expounded for the current research status of monitoring and control in UCG, and clarified that monitoring and control in UCG is not perfect, remaining in the stage of exploration. To improve the problem of low coal gasification rate and gas production, and then to make full use of the potential sustainable advantages, the paper offers a perception platform of a UCG monitoring system based on the Internet-of-Things (IoT) and an optimal control model for UCG based on deep learning, and has an outlook on breakthrough directions of the key technologies related to the package structure design for moisture-proof and thermal insulation, antenna design, the strategy for energy management optimization, feature extraction and classification design for the network model, network structure design, network learning augmentation, and the control of the network model, respectively.

Keywords: underground coal gasification; monitoring and control; Internet of Things; deep learning

1. Introduction Coal is one of the most important fossil in the world, playing a vital role in human production, especially in industrial heating, urban gas production, power generation, and other fields. However, the use of coal, especially the emission of sulfides and oxides during coal combustion, causing a series of environmental problems, has seriously affected the physical and psychological health and living quality of human beings all over the world [1]. One way to solve these problems is UCG. It converts coal into gas, followed by the operation of removing sulfides and nitrogen oxides, bypassing the traditional coal-burning process, and has the characteristics of low pollution emission [2]. In addition, UCG is also characterized by a high resource utilization rate [3] (it can exploit the unmineable coal compared with the traditional mining technology, such as coal that is too deep, of low grade, and in thin seams) and low cost [4] ( and coal in situ can be used directly during

Sustainability 2019, 11, 217; doi:10.3390/su11010217 www.mdpi.com/journal/sustainability Sustainability 2019, 11, 217 2 of 14

Sustainability 2019, 11, x FOR PEER REVIEW 2 of 14 UCG process, reducing operating costs). Therefore, UCG is one of the most promising sustainable sustainable technologies in energy, environment, and the economy, having broad application technologies in energy, environment, and the economy, having broad application prospects [5]. prospects [5]. Figure1 illustrates the schematic diagram for UCG. With the compressor injecting a gas agent Figure 1 illustrates the schematic diagram for UCG. With the compressor injecting a gas agent (air, , oxygen enrichment with different concentration, dioxide, or steam) through an (air, oxygen, oxygen enrichment with different concentration, , or steam) through an injection well, the coal in the coal seam, deeply buried under overburden, is gasified into syngas by injection well, the coal in the coal seam, deeply buried under overburden, is gasified into syngas by igniting in the active cavity. Then the harmful gases, such as sulfides and nitrogen oxides, in syngas are igniting in the active cavity. Then the harmful gases, such as sulfides and nitrogen oxides, in syngas in the operation of gas cleaning and the residue is left underground. Finally, the syngas with H2, CO, are in the operation of gas cleaning and the residue is left underground. Finally, the syngas with H2, and CH4, which is output through the production well, becomes a clean feed gas for power generation, CO, and CH4, which is output through the production well, becomes a clean feed gas for power , and storage [6,7]. generation, hydrogen production, and storage [6,7].

Figure 1. Schematic diagram for UCG. Figure 1. Schematic diagram for UCG.

InIn thethe nearly nearly 100 100 years years of research of research and practice and inpractice the world, in the many world, experiences many andexperiences achievements and haveachievements been accumulated have been in accumulated the methods in and the technologies methods and of technologies UCG. However, of UCG. in practice However, the gasification in practice processthe needs to beprocess monitored needs and to controlledbe monitored to ensure and thatcontrolled the underground to ensure gasificationthat the underground has a stable combustiblegasification has gas a composition stable combustible and calorific gas composition value, as well and as calorific a higher value, gas yield as well and as gasification a higher gas rate yield [1]. UCGand gasification is a very complex rate [1]. physical UCG is and a very chemical complex process, physical and manyand chemical factors affectprocess, the and composition many factors and qualityaffect the of syngas.composition Considering and quality high of temperature syngas. Considering humidity high and closedtemperature environment, humidity it and is difficult closed toenvironment, effectively monitor it is difficult and controlto effectively the UCG monito processr and to control improve the the UCG quality process of theto improve syngas. the Therefore, quality scholarsof the syngas. in various Therefore, countries scholars have in carried various out countries a series have of meaningful carried out research a series onof meaningful the monitoring research and controlon the inmonitoring UCG, such and as control the GPR in (Ground UCG, such Penetrating as the GPR Radar) (Ground technique Penetratin for theg monitoringRadar) technique [8,9], and for studythe monitoring on controlling [8,9], theand combustion study on controlling state in the the gasification combustion process state in [10 the–14 gasification]. There are process few related [10– studies,14]. There technical are few means related remaining studies, on technical paper, or means at the experimentalremaining on teaching paper, level,or at andthe basicexperimental research isteaching almost level, blank. and Therefore, basic research there are is almost still some blank. obstacles Therefore, standing there inare the still way some to actualobstacles application standing atin present.the way to In actual recent application years, with at present. the rapid In developmentrecent years, with of the the Internet-of-Things rapid development (IoT) of the and Internet- deep learningof-Things research (IoT) and [15– 19deep], great learning success research has been [15–19], achieved great in solving success complex has been monitoring achieved and in controlsolving problemscomplex monitoring [20–25]. This and undoubtedly control problems provides [20–25]. a new This idea undoubtedly for the monitoring provides and a new control idea of for UCG. the monitoring and control of UCG. Based on this, the paper puts forward a new idea about monitoring and control of UCG based on IoT and deep learning to improve the problem of the low rate and gas production, and then makes full use of the potential sustainable advantages. Sustainability 2019, 11, 217 3 of 14

Based on this, the paper puts forward a new idea about monitoring and control of UCG based on IoT and deep learning to improve the problem of the low coal gasification rate and gas production, and then makes full use of the potential sustainable advantages. This paper reviews the development of UCG projects worldwide and the research progress related to monitoring and control for UCG. Taking the development of IoT and deep learning as an opportunity, it is a new idea of applying the IoT and deep learning into monitoring and control in UCG to be put forward, and the key technical breakthrough directions related to them are also discussed.

2. The Development of UCG Projects Worldwide In 1868, William Siemens creatively proposed the technical idea of coal gasification in the underground, and then the Russian chemist Dimitri I. Mendeleev proposed the process conception of UCG in 1888 [26]. By the beginning of the 20th century, as many cities were covered with smoke from the side effects of the Industrial Revolution, William Ramsay opened the first UCG test in Durham, to try to solve the problem [27]. Since then, the world’s major coal powers have been competing for nearly a hundred years of research and engineering experiments. Table1 shows the primary projects of overseas UCG in the 20th century [2]. Since the beginning of this century, UCG has been identified as an important source of energy in the context of a low-carbon economy [28], and UCG projects have set off a new upsurge [29]. Table2 shows the primary UCG projects in the 21st century. Due to space limitations, the paper only briefly introduces the relevant gasification experiments in countries around the world since the 21st century.

Table 1. Primary projects of overseas UCG in the 20th century.

Seam Seam Coal Gasified Syngas cv Test Site Country Year Thickness (m) Depth (m) (ton) (mj/m3) Lisichansk Russia 1934–1936 0.75 24 3–4 Lisichansk Ukraine 1943–1963 0.4 400 3.2 Gorlovka Russia 1935–1941 1.9 40 6–10 Podmoskova Russia 1940–1962 2 40 6 with O2 Bois-la-Dame Belgium 1948 1 Newman Spinney UK 1949–1959 1 75 180 2.6 Yuzhno-Abinsk Russia 1955–1989 2-Sep 138 2 × 106 9–12.1 Angren Uzbekistan 1965–now 4 110 >1 × 107 3.6 Hanna 1 USA 1973–1974 9.1 120 3130 Hanna 2 USA 1975–1976 9.1 84 7580 5.3 Hoe Creek 1 USA 1976 7.5 100 112 3.6 Hanna 3 USA 1977 9.1 84 2370 4.1 Hoe Creek 2A USA 1977 7.5 100 1820 3.4 Hoe Creek 2B USA 1977 7.5 100 60 9 Hanna 4 USA 1977–1979 9.1 100 4700 4.1 Hoe Creek 3A USA 1979 7.5 100 290 3.9 Hoe Creek 3B USA 1979 7.5 100 3190 6.9 Pricetown USA 1979 1.8 270 350 6.1 Rawlins 1A USA 1979 18 105 1330 5.6 Rawlins 1B USA 1979 18 105 169 8.1 Rawlins 2 USA 1979 18 130–180 7760 11.8 Brauy-en-Artois France 1981 1200 Thulin Belgium 1982–1984 860 Centralia Tono A USA 1984–1985 6 75 190 9.7 Centralia Tono B USA 1984–1985 6 75 390 8.4 Haute-Duete France 1985–1986 2 880 Thulin Belgium 1986–1987 6 860 157 Rocky Mountain 1A USA 1987–1988 7 110 11,200 9.5 Rocky Mountain 1B USA 1987–1988 7 110 4440 8.8 EI Tremedal Spain 1997 2 600 Sustainability 2019, 11, 217 4 of 14

In order to reduce the pollution of coal-fired power plants, the Chinese government decided to gradually develop the UCG project. There are about 15 underground coal gasification test centers in China. The UCG Research Center of China University of Mining and Technology (Beijing) has conducted theoretical research, model tests, and field tests. At the same time, the coal industry has led the development of UCG in China. In 2011, Seamwell International Limited (UK) and China Energy Conservation and Environmental Protection Corporation launched a $1.5 billion business partnership to gasify 6 × 106 tons of coal buried in the Yihe coalfield in Inner Mongolia, with a plan to produce 1000 MW of electricity over 25 years. In 2013, Zhengzhou Coal Industry Group Co., Ltd. as the sole technical cooperation partner of Carbon Energy’s UCG project in China, owned the technology licensing agreement for UCG [26,30,31]. In 2005, Neyveli Company, funded by the Ministry of Education of , launched a new UCG project in India to study and evaluate the calorific value of the gases produced by the UCG test using suitable lignite blocks. Subsequently, the Central Mine Planning and Design Institute Limited (CMPDI) prepared five UCG coalfields and appointed the Russian Skochinsky Institute of Mining (SIM) as a consultant. The Raniganj coalfield, one of the five coalfields, has completed the UCG trial and drilling. In July 2014, two other UCG coalfields under the responsibility of Central Coalfields Ltd. and Western Coalfields Ltd. were identified by Limited (CIL) for commercial use [29,32,33].

Table 2. Primary UCG projects in the 21st century.

Year of Project Countries Company Organization Objective Commencement UCG Research Centers (Beijing) Seamwell, China Energy Conservation and Power Generation China 2011 Environmental Protection Corporation H2 for cells. Zhengzhou Coal Industry Co., Ltd. Neyvell Lignite Corporation Limited Power Generation Central Mine Planning and Design Institute India 2005 Study and evaluate the calorific Limited value of the gas generated. Central Coalfields Ltd., Western Coalfields Ltd. Pakistan 2009 Thar Coal & Energy Board Power Generation. liquefaction, Lawrence Livermore National Laboratories the U.S. 2005 Developing 3D cavity growth Linc Energy, Carbon Energy and Ergo Energy simulators. UCG-CCS, UCG-IGCC 2007 Linc Energy company Power Generation. environmental and safety issues 2007 Central Mining Institute of Poland related to UCG processes.

In December 2009, Thar Coal and Energy Board partnered with Chinese, German, and South African companies to build a 6.5 × 106 ton/a coal mine in Pakistan that is expected to generate 1200 MW of electricity. A UCG project was launched in Thar Coal deposits in 2010, and the plant is still under construction, generating electricity in 2018 [4]. The United States and Canada have carried out field trials and theoretical studies on UCG for decades. The major developments of UCG in North America took place in 2005, mainly in Lawrence Livermore National Laboratories, which received Department of Energy (DOE) funding in 2006 to summarize the best test methods for UCG. Subsequently, Lawrence Livermore National Laboratories continued to study the UCG project, focusing on the development of a complete three-dimensional cavity growth simulator (reaction, geothermal, hydrological, hydrodynamics). The commercial development of UCG in the United States is mainly undertaken by Australian and Canadian companies (Linc Energy, Carbon Energy, Ergo Energy, and Laurus). In September 2014, Linc Energy was granted permission to develop UCG projects in Wyoming and to develop UCG-related from natural gas liquefaction. In the same year, Linc Energy also carried out UCG project exploration with the relevant state departments in Cooks Bay Alaska. Canada Company Laurus plans to develop a Sustainability 2019, 11, 217 5 of 14 series of UCG projects. Currently, Canada’s most advanced UCG project is a Swan Hills Synfuels pilot project supported by the Alberta Energy Research Institute. Its mining depth is 1400 m thick, the coal seam is 4 to 5.2 m, and it used to be the world’s deepest UCG project [34]. Australia’s UCG project, the world leader in 2000–2012, has made considerable progress and development. In 1999, Linc Energy Company set up a UCG-IGCC (Integrated Gasification Combined Cycle) pilot project in Chinchilla, Queensland. Five underground coal gasification reactors were tested. After running for 21 months, the total output power is 67 MWh. In addition, Australia also actively conducts UCG syngas conversion into fuel research experiments. In 2007, Australia built a coal underground gasification and liquefaction plant. At present, the Australian Government has indicated its willingness to undertake carbon capture and storage (CCS) projects involving power generation and is interested in cooperating with European research institutions in UCG-CCS projects [4,26,31,35–37]. The Polish government believes that its vast coal resources can be exploited by UCG technology and generate electricity. Since 2007, Poland has assessed the feasibility of implementing UCG projects in the country through new exploration and field tests. The Central Mining Institute of Poland has undertaken the European Union project, Hydrogen Oriented Underground Coal Gasification (2007–2010), funded by the Research Fund for Coal and . Its main focus is the production of hydrogen-rich gases in situ by underground gasification, with a total gasification of 22 tons of , an average gasification rate of 6.2 × 10−2 ton/h, and a total gas production of 71,764 m3. The follow-up project HUGE2 (2011–2014) funded by the Research Fund for Coal and Steel focuses on environmental and safety issues related to UCG processes. A total of 5.36 tons of anthracite was gasified with a gasification rate of 3.78 × 10−2 ton/h and a total gas production of 11,043 m3 [38]. In addition, underground coal gasification projects in Britain, Bulgaria, South Africa, and other countries are also gradually developing [39–42]. To summarize, many valuable practices have been carried out in the field of UCG all over the world, which proves that UCG is one of the important ways of coal production technology revolution. However, UCG has not been able to form a large-scale development compared with the surface gasification [43]. The main reasons are: (1) difficulty in monitoring. (2) hard to control. It is difficult to perceive and monitor the combustion state of the gasification face because gasification is carried out in a closed underground and the underground conditions are complex and changeable as well as high temperature and humidity [44]. The UCG process is complex and dynamic [45]. On the other hand, the position of active cavity moves forward continuously in the gasification process [46,47], it is not fixed like the surface gasification position. The gasified coal seam is superimposed and solidified, burning in situ, unlike surface gasification, which can break the coal body according to requirements, add feeding materials according to requirements, clean up coal ash in time, and be easily controlled.

3. The Research Progress Related to Monitoring and Control for UCG The coal gasification rate is low and the gas production as well as calorific value is unstable due to the difficulty in monitoring and controlling in UCG process. In order to solve these two kinds of problems, and then make full use of the potential sustainable advantages, scholars in various countries have worked out a series of meaningful research. The following discusses research status of UCG monitoring and control in two aspects.

3.1. Research Status of UCG Monitoring Early UCG tests used thermocouples, flowmeters and gas analysis equipment to monitor underground temperature and combustion status [48]. Although these measurements effectively track combustion status and related underground and ground subsidence, they only have good effect on shallow seams and poor resolution for deeper seams. However, facing the high temperature and humidity environment of UCG, such sensors are not very accurate and may not work properly. In addition, Wang [49] studied and designed monitoring and control system of UCG based on wireless Sustainability 2019, 11, 217 6 of 14 senor self-organizing network. However, the system has been tested in the laboratory and has not considered the high temperature and humidity environmental factors in the design. Li [50] designed a real-time monitoring and control system for UCG by using Siemens S7-300 PLC and Fame View configuration software. However, the system is used in the teaching experiment, and the high temperature and humidity environment factors are also not considered in the design. Barnwa et al. [51] proposed a UCG monitoring system based on wireless sensor networks to monitor groundwater pollution in the process of gasification. However, its sensor nodes are not suitable for working in typical high temperature and humidity UCG environments. Kostúr et al. [44] proposed a new method for the low-calorific coal with a higher humidity to improve the UCG technology from the point of syngas quality. However, these methods are costly or temporarily impossible to scale. Based on this, scholars from various countries explore other effective methods to monitor the UCG reaction process: Su et al. [43,52] monitored the temperature distribution, gas generation rate and gas content in the active cavity by the acoustic emission technique. Mellors et al. [48] applied electrical resistivity tomography and interferometric synthetic aperture radar techniques into UCG monitoring. Kotyrba et al. [8] used ground penetrating radar to remote monitor the image of mass loss (porosity, crack) area in coal during and after UCG process. Unfortunately, such methods are in the simulation stage. On the other hand, Nurzynska et al. [5] developed a UCG visualization information system to visualize the combustion state of UCG. On this basis, Kotyrba [9] monitored UCG data by GPR and visualized them in 3D. However, the standard commercial radar system is used to test the in situ and non-in situ combustion conditions of coal. The results show that the method can be used to monitor and visualize the coal gasification process under the non-in situ combustion conditions. It is not applicable in other cases, that is, it cannot monitor the in situ combustion state of coal in the actual gasification process.

3.2. Research Status of UCG Control The control for UCG is an emerging field with rare references and limited to theoretical exploration and laboratory simulation. Based on model test, according to mass conservation, energy conservation, and chemical thermodynamics theory, Yang et al. [53–56] established and deduced the mathematical expression of moving velocity of gasification face, discussing the relationship between moving velocity of gasification face and oxygen supply as well as coal particle size. The calculated value is basically consistent with the model test value. On the basis of mathematical models, some scholars have studied the combustion state in gasifier [10–12,57]. These studies are effective for predicting combustion state. However, it is needed for real-time and accurate evaluation of combustion state in practice. Daggupati et al. [13] gave the relevant characteristics that affect the combustion state of UCG through laboratory simulation experiments. Subsequently, the team determined the optimal operating conditions for syngas conversion through experiments, and studied the effects of different operating variables on the evolution of the gasification face [14]. However, it did not determine an accurate mathematical model to characterize such processes. Therefore, Perkins et al. [58] developed a one-dimensional numerical model to investigate the effects of operating conditions (e.g., temperature, pressure, water influx, gas composition) and coal properties (e.g., thermo-mechanical spalling behavior, reactivity, composition) on the rate of local cavity growth and the effectiveness of energy utilization. The thermo-mechanical spalling behavior of coal, the behavior of the ash, and the amount of fixed carbon in coal were found to mostly affect the combustion rate. Uppal et al. [59] proposed a control-oriented UCG one-dimensional packed bed model that maintains the expected heating value of the outlet gas mixture by controlling the flow rate of the injected gas. The model can also predict important data parameters such as gas composition and combustion rate. However, in these control models, it is necessary to assume that the total concentration of all gases remains constant throughout the active cavity. Subsequently, Uppal made improvements to design a sliding mode control algorithm in the UCG simplified model to ensure the stability of the output heat value of the entire system [60]. However, the proposed method has simplified the modelling of the UCG process. In reality the Sustainability 2019, 11, 217 7 of 14 chemical reactions, heat, and mass transfer, pore pressure, moisture content variations, evaporation, reactor void pressure, water influx, and development of cracks also affect the total behavior of a system. A three-dimensional thermal–mechanical modelling around UCG cavities using the computational software ABAQUS has been performed. The model was able to simulate the heat propagation, stress distribution, and surface subsidence in the UCG process [61]. However, this method only studied the thermo-mechanical behavior. The control of UCG process on site scale is a difficult task, especially considering process nonlinearity and ground disturbance [62]. Some other factors also make UCG control a challenging task. In summary, these two kinds of research results will undoubtedly provide strong technical support for the research UCG monitoring and control. However, the research is relatively rare and in the theoretical exploration stage. To ensure the stability of the gasification process and to improve the composition and calorific value of the combustible gas, new technologies and methods are urgently needed to solve the problems of monitoring and control in the UCG process.

4. Future Perspectives The IoT is an important part of the new generation of information technology. The related technology based on the IoT in the field of monitoring are developing rapidly. Saravanan et al. [21] put forward a real-time water quality monitoring SCADA system based on IoT. The real-time system generates, collects, transacts, and stores sensor data through the GSM module in the Web server. Its data is analyzed, and then instant reports are generated and displayed at any time on the Web browser. The system is designed to reduce manpower and costs, improve water distribution and monitoring efficiency. Marques et al. [22] proposed a real-time PM exposure monitoring system based on the IoT architecture. By monitoring indoor air quality, the system can accurately perceive the air quality status and, if necessary, plan interventions to reduce the level of PM exposure. It is an effective tool for monitoring indoor air particulate matter, aiming to ensure the permanent classification of particulate matter. There are other success stories, and this article does not list them one by one. Successful cases have shown that the IoT has a more thorough perceptive capability, i.e., it can use any device, system, or process that can sense, measure, capture, and transmit information at anytime, anywhere to monitor the environment in real-time. There is no doubt that the new impetus and imagination are brought to monitoring in the field of UCG. On the other hand, with the rise of artificial intelligence, deep learning has developed rapidly in the field of control in recent years. Wang et al. [24], who transformed flame images into multi-layer DNN (depth neural network) or CNN (convolution neural network) by a deep learning model, predicted and identified the combustion state and heat release rate of a furnace. At the same time, the technology of stabilization and adjustment can balance the stability and sensitivity to ensure the accurate prediction of combustion state. The experimental results show that this method can predict the combustion state of the image. The recognition speed of each image is less than 1 ms, and the average accuracy is 99.91%. Therefore, it has great potential in industrial applications. Zhang et al. [25] proposed a new method to identify caving coal and rock accurately and quickly by means of bimodal deep learning and the Hilbert-Huang transform for mechanical vibration and acoustic signals of a hydraulic support tail beam. In this study, the bimodal learning method is used to make a DNN fully characterize the characteristics of coal and rock. It is used for transfer learning to solve the problem that training a DNN needs a large number of samples, and the extracted acceleration and sound pressure signal characteristics are combined to extract the most effective features. Experimental results show that the method has obvious advantages in recognition accuracy. Successful application cases show that the deep learning method can fully utilize the autonomy of the model, actively learn the surrounding environment information, constantly adjust its own parameters and network structure, and does not need to design specific algorithms for the data. It can explore its expression in low-dimensional space by high-dimensional learning, and obtain the self-learning features which are totally different from Sustainability 2019, 11, 217 8 of 14

manual features, so as to guide the physical model to make correct action choices. This also brings new impetus and imagination to control research in the field of UCG. In conclusion, we intend to apply the IoT into monitoring in UCG in order to accurately perceive the site information of UCG. We intend to apply deep learning into the control in UCG in order to accurately control the combustion speed and combustion state of UCG. Next, we will focus on the two aspects that we should study in the future. Sustainability 2019, 11, x FOR PEER REVIEW 8 of 14 4.1. Perception Platform of UCG Monitoring System Based on IoT Figure 2 shows the perception platform of UCG monitoring system based on IoT. The wireless nodes Figureare distributed2 shows thereasonably perception to monitor platform the of inpu UCGt gasifier, monitoring the systemoutput syngas, based on the IoT. three-zone The wireless gas componentnodes are distributedof the combustion reasonably furnac to monitore, the gasification the input gasifier,channel, the the output pressure, syngas, the temperature, the three-zone and gas thecomponent flow rate. of The the wireless combustion infrared furnace, cameras the gasificationare installed channel, reasonably the pressure,for flame themonitoring temperature, and the and wirelessthe flow AP rate. nodes The are wireless placed infraredreasonably. cameras These are thre installede devices reasonably transmit perceptual for flame monitoringinformation and to the the groundwireless database AP nodes server are placed for unified reasonably. storage. These Based three on devices perceptive transmit information, perceptual the information factor analysis to the servesground to databaseconstruct server an analysis for unified model storage. of influe Basedncing on perceptivefactors in information,the UCG process the factor and analysisdynamically serves analyzesto construct the influencing an analysis modelfactors ofon influencing the stable combustion factors in the of UCG coal processgasification. and dynamically It is also used analyzes to guide the theinfluencing optimized factors control on server the stable to rapidly combustion utilize of the coal deep gasification. learning method. It is also A used guidance to guide model the optimized for coal gasification,control server which to rapidly is constructed utilize the based deep on learning the perceptive method. Ainformation, guidance model can precisely for coal gasification,adjust the injectionwhich is of constructed gasifier components based on the and perceptive the output information, of , can so precisely as to achieve adjust precise the injection control of gasifierof the combustioncomponents speed and theand output combustion of coal state gas, for so asUCG. to achieve precise control of the combustion speed and combustion state for UCG.

Figure 2. Perception platform of UCG monitoring system based on the IoT. Figure 2. Perception platform of UCG monitoring system based on the IoT. However, with the complexity and dynamic of UCG process and the high temperature and However, with the complexity and dynamic of UCG process and the high temperature and humidity environment, the mature ground IoT technology and products will not work well under humidity environment, the mature ground IoT technology and products will not work well under UCG environment. Based on this, the three key directions are proposed to solve the problem of reliable UCG environment. Based on this, the three key directions are proposed to solve the problem of information monitoring in the UCG process for the perceptive information and communication nodes. reliable information monitoring in the UCG process for the perceptive information and communication nodes.

4.1.1. Package Structure Design for Moisture Proof and Thermal Insulation The traditional wireless sensors for mine monitoring products will not work because of the high temperature and humidity environment of UCG. Therefore, perceptive information based on the IoT of UCG is facing the problem of the package structure design for moisture and thermal insulation. It is found for suitable high-temperature sensitive monitoring element and low-power semiconductor chips to design the intrinsic safety circuit of a WSN perceptive information and communication node, so as to achieve the purpose of active high-temperature resistance. It is designed for packaging internal cooling pipe networks to ensure the isothermal temperature inside the node, so as to achieve the purpose of resisting high temperature and high humidity. It is searched for suitable high- temperature resistant materials and surface thermal control coatings to study special packaging technology of nodes, so as to achieve the purpose of moisture and heat insulation. Sustainability 2019, 11, 217 9 of 14

4.1.1. Package Structure Design for Moisture Proof and Thermal Insulation The traditional wireless sensors for mine monitoring products will not work because of the high temperature and humidity environment of UCG. Therefore, perceptive information based on the IoT of UCG is facing the problem of the package structure design for moisture and thermal insulation. It is found for suitable high-temperature sensitive monitoring element and low-power semiconductor chips to design the intrinsic safety circuit of a WSN perceptive information and communication node, so as to achieve the purpose of active high-temperature resistance. It is designed for packaging internal cooling pipe networks to ensure the isothermal temperature inside the node, so as to achieve the purpose of resisting high temperature and high humidity. It is searched for suitable high-temperature resistant materials and surface thermal control coatings to study special packaging technology of nodes, so as to achieve the purpose of moisture and heat insulation.

4.1.2. Antenna Design After the perceptive information and communication nodes are packaged by the intrinsic safety design of moisture and thermal insulation, it is difficult to communicate for nodes due to shielding of packaged metal, and it is intricate for the physical and chemical changes of the UCG process because of the restricted physical space and many changed dynamic factors affecting the gasification process, such as gasifier properties, gasification channel pressure, temperature, and flow rate injection. Therefore, the theoretical model of radio wave propagation is explored under high temperature and high humidity environments, and it is designed for the node antenna that meets the requirements of UCG communication to solve such problems. In addition, due to the small size and limited battery capacity of the wireless sensor nodes, it is necessary to reduce the antenna size and improve the radiation resistance and antenna radiation efficiency. All of these lays a foundation for node energy optimization.

4.1.3. The Strategy for Energy Management Optimization In the UCG process, it is difficult for people to reach the perceptive information and communication nodes so that most of the nodes will be powered by unreplaceable batteries. Therefore, based on the theory of radio wave propagation in high temperature and high humidity environments, the strategy for energy management optimization of nodes are studied to ensure their stable operation and reliable transmission and, at the same time, energy is used efficiently to maximize the network life cycle. On the other hand, to save energy and consumption, UCG uses dynamic voltage regulation technology, which refers to reducing the processing power by reducing the operating voltage and frequency of the microprocessor when the load is light.

4.2. Optimal Control Model for UCG Based on Deep Learning Figure3 shows the optimal control model for UCG based on deep learning. There are two kinds of input in deep learning model, which are monitored by the perception platform of UCG monitoring system based on the IoT. One is the perceptive factors affecting gasification result, such as the gasifying agent component, the gasification channel pressure, the gasification channel temperature, the gasification channel velocity, and so on, the other is the perception of gasification result, which is divided into video perception and syngas perception. The complex high-dimensional nonlinear function between the input and output is characterized by deep learning model, which accurately predicts the gasification combustion and combustion speed. Subsequently, the proportion of the syngas component is precisely controlled by the controller, so as to achieve the purpose of controlling the velocity and output of the gas, and guiding the development of gasification in a good direction. Sustainability 2019, 11, x FOR PEER REVIEW 9 of 14

4.1.2. Antenna Design After the perceptive information and communication nodes are packaged by the intrinsic safety design of moisture and thermal insulation, it is difficult to communicate for nodes due to shielding of packaged metal, and it is intricate for the physical and chemical changes of the UCG process because of the restricted physical space and many changed dynamic factors affecting the gasification process, such as gasifier properties, gasification channel pressure, temperature, and flow rate injection. Therefore, the theoretical model of radio wave propagation is explored under high temperature and high humidity environments, and it is designed for the node antenna that meets the requirements of UCG communication to solve such problems. In addition, due to the small size and limited battery capacity of the wireless sensor nodes, it is necessary to reduce the antenna size and improve the radiation resistance and antenna radiation efficiency. All of these lays a foundation for node energy optimization.

4.1.3. The Strategy for Energy Management Optimization In the UCG process, it is difficult for people to reach the perceptive information and communication nodes so that most of the nodes will be powered by unreplaceable batteries. Therefore, based on the theory of radio wave propagation in high temperature and high humidity environments, the strategy for energy management optimization of nodes are studied to ensure their stable operation and reliable transmission and, at the same time, energy is used efficiently to maximize the network life cycle. On the other hand, to save energy and consumption, UCG uses dynamic voltage regulation technology, which refers to reducing the processing power by reducing the operating voltage and frequency of the microprocessor when the load is light.

4.2. Optimal Control Model for UCG Based on Deep Learning Figure 3 shows the optimal control model for UCG based on deep learning. There are two kinds of input in deep learning model, which are monitored by the perception platform of UCG monitoring system based on the IoT. One is the perceptive factors affecting gasification result, such as the gasifying agent component, the gasification channel pressure, the gasification channel temperature, the gasification channel velocity, and so on, the other is the perception of gasification result, which is divided into video perception and syngas perception. The complex high-dimensional nonlinear function between the input and output is characterized by deep learning model, which accurately predicts the gasification combustion and combustion speed. Subsequently, the proportion of the Sustainabilitysyngas component2019, 11, 217 is precisely controlled by the controller, so as to achieve the purpose of controlling10 of 14 the velocity and output of the gas, and guiding the development of gasification in a good direction.

Figure 3. Optimal control model for UCG based on deep learning.

The key directions of optimal control model for UCG based on deep learning are mainly composed of feature extraction and classification design for network model, network structure design, network learning augmentation, and the control of the network model.

4.2.1. Feature Extraction and Classification Design for Network Model The input of deep learning network used in video analysis is usually the original video image, and the convolutional neural networks are not always ideal for non-image signal processing. However, non-video perceptive information is also obtained by the perception platform of UCG monitoring system based on IoT. In order to realize the fusion of video and other non-video perceptual signal features of coal gasification, it is necessary to construct a hybrid deep learning network with video convolution-signal local connection by extracting the input data features of different attributes in one network and fusing the features through full connection in the later stage of the deep network. Thus, it can be determined for the classification of combustion state and the fitting of combustion speed characterized by the output node. Then the control parameters of coal gasification are also determined.

4.2.2. Network Structure Design The accuracy for the training and control of deep neural networks depends on the deep of network structure to a certain extent. Theoretically, the deeper the structure is, the more complex the nonlinear mapping relation of the function can be expressed, that is, the more sufficient feature extraction is, the stronger the classification and fitting ability of the network is. However, in practical applications, the training algorithm of existing deep learning networks will expand rapidly with the increase of network depth, and may cause the gradient disappearance, which will to the convergence of the training process and is prone to over-fitting. Therefore, while improving the network training algorithm, it is also adjusted the structure of the deep learning network to determine the most suitable network structure and the depth of the network hierarchy, as well as the nonlinear activation function and the output objective function. Sustainability 2019, 11, 217 11 of 14

4.2.3. Network Learning Augmentation Based on the past UCG test experience, the risk of failure is high and the deployed perceptive nodes may disappear after combustion so that the number of effective learning samples may be very limited. The collected data of video, gasification component, temperature, and pressure are even more precious. It is necessary to expand and augment the learning samples for deep neural network training through some means. Therefore, it is particularly important for the study of the augmentation methods of learning samples. The deep learning network can be quickly converged in the case of limited learning samples, reducing the need for precious learning samples.

4.2.4. The Control for Network Model After obtaining a good deep learning model, it is also necessary to study the control strategy in the UCG process to obtain more accurate classification results. The control for network model is used to realize the control prediction and feedback iterative correction by deep learning. In this way, feedback function is added to a classification network, so that the classification results are fed back to correct input and the output classification results are obtained again. It is precisely adjusted for the injection of gasification component and the output of gas to achieve the purpose of precise controlling of the combustion state and combustion speed.

5. Conclusions Compared with traditional technology, UCG has the potential sustainable advantages of high energy efficiency and low operating cost, and also has the sustainable potential to improve global air pollution problems. Among them, monitoring and control, which help to improve the quality of UCG syngas and maintain the stability of the calorific value and then make full use of the potential sustainable advantages, is a very important technology in the field of UCG. Researchers have conducted a series of meaningful studies in the field of monitoring and control for UCG, but this kind of research is relatively rare and in the stage of theoretical exploration, which prompted us to think about and explore new methods of monitoring and control in the field of UCG. The paper had on outlook on the new monitoring and control in UCG, which can be realized from two models: the perception platform of a UCG monitoring system based on IoT and the optimal control model for UCG based on deep learning. These methods are expected to enable the injection of a UCG gas agent component and the output of gas to be accurately adjusted, thereby achieving the purpose of accurately controlling the combustion state and combustion speed of UCG, and finally make full use of the potential sustainable advantages. It should be pointed out that the new monitoring and control methods still face many problems, such as the package structure design for moisture-proof and thermal insulation, antenna design, the strategy for energy management optimization, and feature extraction and classification design for the network model, network structure design, network learning augmentation, and the control for network model. These are also the direction that the authors continue to explore.

Author Contributions: Conceptualization: Y.X. and H.Y.; data curation: J.Y.; writing—original draft preparation: Y.H.; writing—review and editing: J.W.; supervision: H.Q. Funding: This research was funded by the National Natural Science Foundation of China 61379100, 61472388, and 51574232. Conflicts of Interest: The authors declare no conflict of interest. Sustainability 2019, 11, 217 12 of 14

References

1. Mcinnis, J.; Singh, S.; Huq, I. Mitigation and adaptation strategies for global change via the implementation of underground coal gasification. Mitig. Adapt. Strateg. Glob. Chang. 2016, 21, 479–486. [CrossRef] 2. Bhutto, A.W.; Bazmi, A.A.; Zahedi, G. Underground coal gasification: From fundamentals to applications. Prog. Energy Combust. Sci. 2013, 39, 189–214. [CrossRef] 3. Green, M. Recent developments and current position of underground coal gasification. Proc. Inst. Mech. Eng. Part A J. Power Energy 2018, 232, 39–46. [CrossRef] 4. Shafirovich, E.; Varma, A. Underground Coal Gasification: A Brief Review of Current Status. Ind. Eng. Chem. Res. 2009, 48, 7865–7875. [CrossRef] 5. Nurzynska, K.; Iwaszenko, S.; Choroba, T. Database Application in Visualization of Process Data. In Beyond Databases, Architectures and Structures—Proceedings of the 10th International Conference BDAS 2014; Springer: Cham, Switzerland, 2014; Volume 424, pp. 537–546. 6. Doucet, D.; Perkins, G.; Ulbrich, A.; Toit, E. Du Production of power using underground coal gasification. Energy Sources 2016, 38, 3653–3660. [CrossRef] 7. Kostúr, K.; Laciak, M.; Durdan, M. Some influences of Underground Coal Gasification on the environment. Sustainability 2018, 10, 1512. [CrossRef] 8. Kotyrba, A.; Sta´nczyk,K. Application of a GPR Technique for Monitoring of Simulated Underground Coal Gasification in a Large Scale Model. Near Surf. Geophys. 2013, 11, 505–515. [CrossRef] 9. Kotyrba, A.; Sta´nczyk,K. Sensing underground coal gasification by ground penetrating radar. Acta Geophys. 2017, 65, 1185–1196. [CrossRef] 10. Khan, M.M.; Mmbaga, J.P.; Shirazi, A.S.; Trivedi, J.; Liu, Q.; Gupta, R. Modelling Underground Coal Gasification—A Review. Energies 2015, 8, 12603–12668. [CrossRef] 11. Nourozieh, H.; Kariznovi, M.; Chen, Z.; Abedi, J. Simulation Study of Underground Coal Gasification in Alberta Reservoirs: Geological Structure and Process Modeling. Energy Fuels 2010, 24, 3540–3550. [CrossRef] 12. Cui, Y.; Liang, J.; Wang, Z.; Zhang, X.; Fan, C.; Liang, D.; Wang, X. Forward and reverse combustion gasification of coal with production of high-quality syngas in a simulated pilot system for in situ gasification. Appl. Energy 2014, 131, 9–19. [CrossRef] 13. Daggupati, S.; Mandapati, R.N.; Mahajani, S.M.; Ganesh, A.; Mathur, D.K.; Sharma, R.K.; Aghalayam, P. Laboratory studies on combustion cavity growth in lignite coal blocks in the context of underground coal gasification. Energy 2010, 35, 2374–2386. [CrossRef] 14. Daggupati, S.; Mandapati, R.N.; Mahajani, S.M.; Ganesh, A.; Sapru, R.K.; Sharma, R.K.; Aghalayam, P. Laboratory studies on cavity growth and product gas composition in the context of underground coal gasification. Energy 2011, 36, 1776–1784. [CrossRef] 15. Lecun, Y.; Bengio, Y.; Hinton, G. Deep learning. Nature 2015, 521, 436–444. [CrossRef][PubMed] 16. Levine, S.; Finn, C.; Darrell, T.; Abbeel, P. End-to-End Training of Deep Visuomotor Policies. J. Mach. Learn. Res. 2016, 17, 1–40. 17. Karpathy, A.; Li, F.-F. Deep Visual-Semantic Alignments for Generating Image Descriptions. IEEE Trans. Pattern Anal. Mach. Intell. 2017, 39, 664–676. [CrossRef][PubMed] 18. Silver, D.; Huang, A.; Maddison, C.J.; Guez, A.; Sifre, L.; Van Den Driessche, G.; Schrittwieser, J.; Antonoglou, I.; Panneershelvam, V.; Lanctot, M.; et al. Mastering the game of Go with deep neural networks and tree search. Nature 2016, 529, 484–489. [CrossRef] 19. Graves, A.; Wayne, G.; Reynolds, M.; Harley, T.; Danihelka, I.; Grabska-Barwi´nska,A.; Colmenarejo, S.G.; Grefenstette, E.; Ramalho, T.; Agapiou, J.; et al. Hybrid computing using a neural network with dynamic external memory. Nature 2016, 538, 471–476. [CrossRef] 20. Yi, L.; Deng, X.; Wang, M.; Ding, D.; Yan, W. Localized Confident Information Coverage Hole Detection in Internet of Things for Radioactive Pollution Monitoring. IEEE Access 2017, 5, 18665–18674. [CrossRef] 21. Saravanan, K.; Anusuya, E.; Kumar, R.; Le, H.S. Real—time water quality monitoring using Internet of Things in SCADA. Environ. Monit. Assess. 2018, 190, 556. [CrossRef] 22. Marques, G.; Ferreira, C.R.; Pitarma, R. A System Based on the Internet of Things for Real-Time Particle Monitoring in Buildings. Int. J. Environ. Res. Public Health 2018, 15, 821. [CrossRef][PubMed] Sustainability 2019, 11, 217 13 of 14

23. Mnih, V.; Kavukcuoglu, K.; Silver, D.; Rusu, A.A.; Veness, J.; Bellemare, M.G.; Graves, A.; Riedmiller, M.; Fidjeland, A.K.; Ostrovski, G. Human-level control through deep reinforcement learning. Nature 2015, 518, 529–533. [CrossRef][PubMed] 24. Wang, Z.; Song, C.; Chen, T. Deep learning based monitoring of furnace combustion state and measurement of heat release rate. Energy 2017, 131, 106–112. [CrossRef] 25. Zhang, G.; Wang, Z.; Zhao, L.; Qi, Y.; Wang, J. Coal-Rock Recognition in Top Coal Caving Using Bimodal Deep Learning and Hilbert-Huang Transform. Shock Vib. 2017, 2017, 3809525. [CrossRef] 26. Vyas, D.U.; Singh, R.P. Worldwide Developments in UCG and Indian Initiative. Procedia Earth Planet. Sci. 2015, 11, 29–37. [CrossRef] 27. Klimenko, A.Y. Early Ideas in Underground Coal Gasification and Their Evolution. Energies 2009, 2, 456–476. [CrossRef] 28. Imran, M.; Kumar, D.; Kumar, N.; Qayyum, A.; Saeed, A.; Bhatti, M.S. Environmental concerns of underground coal gasification. Renew. Sustain. Energy Rev. 2014, 31, 600–610. [CrossRef] 29. Khadse, A.; Qayyumi, M.; Mahajam, S.; Aghalayam, P. Underground coal gasification: A new clean coal utilization technique for India. Energy 2007, 32, 2061–2071. [CrossRef] 30. Li, Y.; Liang, X.; Liang, J. An overview of the Chinese UCG program. Data Sci. J. 2007, 6, S460–S466. [CrossRef] 31. Yang, D.; Koukouzas, N.; Green, M.; Sheng, Y. Recent development on underground coal gasification and

subsequent CO2 storage. J. Energy Inst. 2016, 89, 469–484. [CrossRef] 32. Singan, A.; Ranade, V.V. Deployment of underground coal gasification in India. Curr. Sci. 2017, 113, 218–227. [CrossRef] 33. Khadse, A.N. Resources and economic analyses of underground coal gasification in India. Fuel 2015, 142, 121–128. [CrossRef] 34. Olateju, B.; Kumar, A. Techno-economic assessment of hydrogen production from underground coal gasification (UCG) in Western Canada with carbon capture and sequestration (CCS) for upgrading bitumen from . Appl. Energy 2013, 111, 428–440. [CrossRef] 35. Perkins, G.; du Toit, E.; Cochrane, G.; Bollaert, G. Overview of underground coal gasification operations at Chinchilla, Australia. Energy Sources Part A Recovery Util. Environ. Eff. 2016, 38, 3639–3646. [CrossRef] 36. Mallett, C.W. Environmental controls for underground coal gasification. Proc. Inst. Mech. Eng. Part A J. Power Energy 2018, 232, 47–55. [CrossRef] 37. Walker, L. UCG commercialization and the Cougar Energy project at Kingaroy, Queensland, Australia. Undergr. Coal Gasif. Combust. 2018, 503–526. [CrossRef] 38. Kapusta, K.; Stanczyk, K. Development conditions and limitations of the underground coal gasification in Poland. Przem. Chem. 2009, 88, 331–338. 39. Shackley, S.; Mander, S.; Reiche, A. Public perceptions of underground coal gasification in the . Energy Policy 2006, 34, 3423–3433. [CrossRef] 40. Sheng, Y.; Benderev, A.; Bukolska, D.; Eshiet, K.I.-I.; da Gama, C.D.; Gorka, T.; Green, M.; Hristov, N.; Katsimpardi, I.; Kempka, T.; et al. Interdisciplinary studies on the technical and economic feasibility of deep

underground coal gasification with CO2 storage in bulgaria. Mitig. Adapt. Strateg. Glob. Chang. 2016, 21, 595–627. [CrossRef] 41. Zieleniewski, M.; Brent, A.C. Evaluating the costs and achievable benefits of extending technologies for uneconomical coal resources in South Africa: The case of underground coal gasification. J. Energy S. Afr. 2008, 19, 21–31. 42. Sarhosis, V.; Lavis, S.; Mostade, M.; Thomas, H.R. Towards commercialising underground coal gasification in the EU. Environ. Geotech. 2017, 4, 113–122. [CrossRef] 43. Su, F.; Hamanaka, A.; Itakura, K.; Zhang, W.; Deguchi, G.; Sato, K.; Takahashi, K.; Kodama, J. Monitoring and evaluation of simulated underground coal gasification in an ex-situ experimental artificial coal seam system. Appl. Energy 2018, 223, 82–92. [CrossRef] 44. Kostúr, K.; Laciak, M.; Durdán, M.; Kaˇcur, J.; Flegner, P. Low-calorific gasification of underground coal with a higher humidity. Measurement 2015, 63, 69–80. [CrossRef] 45. Perkins, G. Underground coal gasification—Part I: Field demonstrations and process performance. Prog. Energy Combust. Sci. 2018, 67, 158–187. [CrossRef] 46. Kaˇcur, J.; Kostúr, K. Approaches to the gas control in UCG. Acta Polytech. 2017, 57, 182–200. [CrossRef] Sustainability 2019, 11, 217 14 of 14

47. Perkins, G. Underground coal gasification—Part II: Fundamental phenomena and modeling. Prog. Energy Combust. Sci. 2018, 67, 234–274. [CrossRef] 48. Mellors, R.; Yang, X.; White, J.A.; Ramirez, A.; Wagoner, J.; Camp, D.W. Advanced geophysical underground coal gasification monitoring. Mitig. Adapt. Strateg. Glob. Chang. 2016, 21, 487–500. [CrossRef] 49. Wang, J.H.; Wang, H.J.; Wang, Z.T. Design of UCG Intelligent Monitoring System Based on Wireless Ad Hoc Network. Coal Mine Mach. 2017, 38, 9–11. [CrossRef] 50. Guan, L.I.; Wen, L.I.; Jia, W.E.I. Monitoring and control system of UCG process based on S7—300 PLC. J. North China Inst. Sci. Technol. 2016, 13, 86–91. (In Chinese) 51. Barnwal, R.P.; Bharti, S.; Misra, S.; Obaidat, M.S. UCGNet: Wireless sensor network-based active aquifer contamination monitoring and control system for underground coal gasification. Int. J. Commun. Syst. 2017, 30, e2852. [CrossRef] 52. Su, F.; Itakura, K.; Deguchi, G.; Ohga, K. Monitoring of coal fracturing in underground coal gasification by acoustic emission techniques. Appl. Energy 2017, 189, 142–156. [CrossRef] 53. Yang, L.H.; Liu, S.Q. Temperature-control Blasting Underground Coal Gasification in the Inclined Coal Seams with a Short Wall. Energy Sources Part A Recovery Util. Environ. Eff. 2010, 32, 532–541. [CrossRef] 54. Yang, L.H.; Pang, X.L.; Liu, S.Q.; Chen, F. Temperature and Gas Pressure Features in the Temperature-control Blasting Underground Coal Gasification. Energy Sources Part A Recovery Util. Environ. Eff. 2010, 32, 1737–1746. [CrossRef] 55. Yang, L. Three-Dimensional Unstable Non-linear Numerical Analysis of the Underground Coal Gasification with Free Channel. Energy Sources Part A Recovery Util. Environ. Eff. 2006, 28, 1519–1531. [CrossRef] 56. Yang, L. Study on the model experiment and numerical simulation for underground coal gasification. Fuel 2004, 83, 573–584. [CrossRef] 57. Kaˇcur, J.; Durdán, M.; Laciak, M.; Flegner, P. Impact analysis of the oxidant in the process of underground coal gasification. Measurement 2014, 51, 147–155. [CrossRef] 58. Perkins, G.; Sahajwalla, V. A Numerical Study of the Effects of Operating Conditions and Coal Properties on Cavity Growth in Underground Coal Gasification. Energy Fuels 2006, 20, 596–608. [CrossRef] 59. Uppal, A.A.; Bhatti, A.I.; Aamir, E.; Samar, R.; Khan, S.A. Control oriented modeling and optimization of one dimensional packed bed model of underground coal gasification. J. Process Control 2014, 24, 269–277. [CrossRef] 60. Uppal, A.A.; Alsmadi, Y.M.; Utkin, V.I.; Bhatti, A.I.; Khan, S.A. Sliding Mode Control of Underground Coal Gasification Energy Conversion Process. IEEE Trans. Control Syst. Technol. 2018, 26, 587–598. [CrossRef] 61. Yang, D.; Sarhosis, V.; Sheng, Y. Thermal–mechanical modelling around the cavities of underground coal gasification. J. Energy Inst. 2014, 87, 321–329. [CrossRef] 62. Prabu, V.; Jayanti, S. Simulation of cavity formation in underground coal gasification using bore hole combustion experiments. Energy 2011, 36, 5854–5864. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).