Prediction of temperature and CO concentration fields based on BPNN in low-temperature coal oxidation
Introduction
Coal spontaneous combustion has tremendously affected the safety production and utilization of coal [1,2]. The resulting combustion led to severe environmental pollution, heavy economic losses and terrible casualties [3]. Coals with long storage time are particularly prone to spontaneous combustion, especially such as residual coals left in the gob [4,5], a large amount of standby coals that are stored in coal mines [6]and long-distance transferring of coals by ship, etc [7]. The abovementioned situations have brought great challenges to the safe storage and transportation of coal.
To prevent and control coal spontaneous combustion, it is important to understand the coal spontaneous combustion mechanism [8]. There are many factors affecting the coal spontaneous combustion [[9], [10], [11], [12], [13]], including internal factors such as degree of coalification, coal rock composition, coal porosity, moisture content and particle size of coal, and external factors such as environmental temperature and air flow infiltration. The oxidation process is attributed to a combination of various factors, and it has highly complex non-linear characteristics. It is also very difficult to investigate how these factors affect coal spontaneous combustion directly. Surprisingly, the continuous variation of temperature and the gas concentration can be used to describe the process of coal [14,15]. When the coal pile starts to combust spontaneously, the coal bodies consume oxygen and release reaction heat and a large amount of gas. Owing to the different oxidation heat release intensity and heat storage conditions at different points in the coal pile, the heating rate, oxygen consumption rate and gas release rate of coal body are different, which changes temperature field and gas concentration field in the coal pile. Therefore, analyzing the changes of temperature and gas concentration in the oxidation process allows us to have better understanding of the occurrence and development of coal spontaneous combustion [[16], [17], [18]].
Numerous methods have been proposed to monitor and predict the temperature and gas indexes [19]. Back propagation neural network (BPNN) was used for coal spontaneous combustion modeling, which gives accurate predictions. Xie et al. established BPNN on the basis of the gas monitoring system with beam tube [20]. The temperature of gob can be forecast by using CO and CO2 concentration values, which significantly improves the prediction accuracy of spontaneous combustion in gob. Also, a random forest (RF) method, a model ensemble method that combines several decision trees was proposed [21]. The RF model accurately predicted the temperature of coal spontaneous combustion when it was applied to the in-situ data. Moreover, Artificial Bee Colony (ABC) optimization algorithm model is highly important and useful for intelligent coal management systems which can improve decision making and reduce risk, especially for short-term forecasts [22]. Furthermore, a new method, DSC Inflection Point (DSCIP), was proposed to determine the coal auto-ignition temperature [23]. The results demonstrated that the heat flux curve of coal spontaneous combustion can be well fitted using Gaussian mixture model. In addition, R/S analysis method was employed to analyze the chaotic characteristics of coal gas indexes from gas drainage pipeline, this kind of prediction method can reduce the misinformation of the coexisting disaster [24]. Many advanced methods can precisely predict the temperature and gas indexes to prevent coal spontaneous combustion [[25], [26], [27]]. However, these researches just monitored and forecast different points inside the coal pile, they did not analyze the whole coal pile.
In this study, we constructed an experiment system to implement a process of low temperature oxidation of coal and collect data points, with controlling six factors as variables, including activation energy, void fraction, moisture content, air flow rate, stacking time and location of measuring point. The BPNN was established based on 72 sets of data acquired from the experimental system and B-R algorithm was selected appropriately. Moreover, the temperature and gas concentration fields were predicted by BPNN to analyze how various factors affect coal spontaneous combustion and how to prevent coal spontaneous combustion by controlling these factors.
Section snippets
Coal samples
Two typical coals, Yujialiang (YJL) and Shangwan (SW) in China, were selected as the experimental samples. The depth of the samples was 200 and 200−300 m, respectively. The results of industrial analysis, elemental analysis, activation energy [28,29] and spontaneous combustion tendency of coal samples are shown in Table 1.
Experiment devices
When spontaneous combustion occurs in a coal pile, it can be divided into three areas: asphyxiation zone, oxidation zone and cooling zone, and the schematic diagram is shown
Modeling principle
BPNN can realize highly complicated non-linear mapping and can be used for complicated pattern recognition and prediction. Its basic idea is to adjust and modify the connection weights and thresholds of the network through the reverse propagation of network output errors, so as to minimize the mean square error of the output. The learning process includes forward calculation of input information and back propagation of error. The specific algorithm is as follows [32,33]:
(1) Forward calculation
Conclusions
In this study, an experiment system was constructed to investigate the changing of temperature and gas concentration under various factors during low temperature coal oxidation. The BPNN was proposed to simulate coal low temperature oxidation with B-R algorithm. We utilized 72 sets of experimental data to train BPNN model, and applied the trained BPNN to generate 40 random sets of data. The agreement between the measured and predicted values validate our proposed BPNN model for coal oxidation
Fundings
This work is supported by the National Nature Science Foundation Funded Project of China [grant numbers: 51674256; 51974308; 51974312]; and the Major Science and Technology Project of Shanxi Province [grant number: 20181102017].
Author contributions
Ruizhi Chu and Shi Yu contributed significantly to the conception of the study;
Jianqiao Zhao performed the data analyses and wrote the whole manuscript;
Deguang Yang and Jiaxin Wu helped perform the analysis with constructive discussions;
Xianliang Meng and Xiao Li contributed to manuscript preparation;
Guoguang Wu and Zhenyong Miao contributed reagents/materials/analysis tools.
CRediT authorship contribution statement
Jianqiao Zhao: Formal analysis, Writing - original draft. Deguang Yang: Formal analysis. Jiaxin Wu: Formal analysis. Xianliang Meng: Visualization. Xiao Li: Visualization. Guoguang Wu: Resources. Zhenyong Miao: Resources. Ruizhi Chu: Conceptualization. Shi Yu: Conceptualization.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
References (46)
- et al.
Multifaceted processes controlling the distribution of hazardous compounds in the spontaneous combustion of coal and the effect of these compounds on human health
Environ. Res.
(2018) - et al.
Controlling factors of symbiotic disaster between coal gas and spontaneous combustion in longwall mining gobs
Fuel
(2016) - et al.
Multi-physics coupling model of coal spontaneous combustion in longwall gob area based on moving coordinates
Fuel
(2017) - et al.
Modeling and simulation of coal spontaneous combustion in a gob of shallow buried coal seams
Process Saf. Environ.
(2019) - et al.
CFD simulations of the effect of wind on the spontaneous heating of coal stockpiles
Fuel
(2014) - et al.
The effect of moisture content and air-drying on spontaneous combustion characteristics of two Turkish lignites
Fuel
(2003) - et al.
Comparative analysis of thermokinetic behavior and gaseous products between first and second coal spontaneous combustion
Fuel
(2018) - et al.
Prediction of coal spontaneous combustion in goaf Based on the BP neural network
- et al.
A random forest approach for predicting coal spontaneous combustion
Fuel
(2018) - et al.
Intelligent management of coal stockpiles using improved grey spontaneous combustion forecasting models
Energy
(2017)
Risk evaluation of coal spontaneous combustion on the basis of auto-ignition temperature
Fuel
Prediction of gas and coal spontaneous combustion coexisting disaster through the chaotic characteristic analysis of gas indexes in goaf gas extraction
Process Saf. Environ.
A method for evaluating the spontaneous combustion of coal by monitoring various gases
Process Saf. Environ.
Unique spatial methane distribution caused by spontaneous coal combustion in coal mine goafs: an experimental study
Process Saf. Environ.
Synergistic mechanism of CO2 and active functional groups during low temperature oxidation of lignite
Fuel.
Kinetic and thermodynamic studies on the mechanism of low-temperature oxidation of coal: a case study of Shendong coal (China)
Int. J. Coal Geol.
Study on the relationship between microscopic functional group and coal mass changes during low-temperature oxidation of coal
Int. J. Coal Geol.
Reaction pathway of coal oxidation at low temperatures: a model of cyclic chain reactions and kinetic characteristics
Combust. Flame
Risk evaluation of coal spontaneous combustion on the basis of auto-ignition temperature
Fuel.
Reaction pathway of coal oxidation at low temperatures: a model of cyclic chain reactions and kinetic characteristics
Combust. Flame
Effect of oxygen concentration on low-temperature exothermic oxidation of pulverized coal
Thermochim. Acta
The heat of wetting and its effect on coal spontaneous combustion
Thermochim. Acta
A study of spontaneous combustion characteristics of a turkish lignite: particle size, moisture of coal, humidity of air
Combust. Flame
Cited by (33)
Flue gas enhanced water leaching: Prediction and condition optimization of demineralization effect from coal via machine learning algorithm
2023, Process Safety and Environmental Protection