Improved method of fire hazard assessment taking into account the effect of the primary temperature of a coal seam on the desorption ratio of gas indicators

The methods of fire risk assessment in its early phase employed so far used to take into account only the temperature of the heating of coal. The research works reported so far in this field have been conducted at ambient temperatures, without considering the rock mass primary temperature occurring in a particular coal seam and neglecting the depth of coal seams. The method for the determination of the composition of gases emitted from coal samples, employed in the study presented here, takes into account the effect of rock mass pressure, and therefore the depth of the coal seam, as well as the temperature typical for the rock mass in a particular coal mine. In the paper the results of the experimental research on the effects of the primary temperature of a coal seam on the content of gases emitted from coal, performed with the use of a specially designed experimental stand are presented. The method may be useful in developing fire hazard predictions for longwalls with residual coal in goafs. The relations between the self-ignition characteristics as well as selected physical properties of coal samples and carbon monoxide emission were also given. The research results proved that the accurate fire hazard assessment requires considering other than just self-heating of coal causes of carbon monoxide emission, including the rock mass primary temperature, which is expected to improve the existing methods of fire risk assessment in coal mines and post-mining areas.

the low temperature conditions, is considered to be the predominant source of heat provided to the process of self-heating of coal in underground mines and the main gas emission driver 11-13 . In the attempts of the description of the oxidation process mechanism reported in the literature various parameters are applied in the development of kinetic models, covering coal sample weight loss, heat release, oxygen demand, as well as kinds and amounts of gaseous and solid products 14 . The effect of an air inflow to coal beds is also studied as one of the factors facilitating the development of fires in underground mines 15,16 .
When it comes to analyzing in details the sources of carbon monoxide emission it has already been reported that the two main are: coal crushing and oxidation at ambient temperature 17 . It has also been proven in studies on the impact of moisture, primary temperature, as well as the content of seam gas and reactive pyrite on coal self-heating rates at low temperatures using adiabatic tests, that gases characteristic for a mine atmosphere may be also emitted as a result of natural desorption 18 .
In the study on gas release in the process of coal production higher initial amounts and rates of gas desorption were observed for fragmented coal than for the body of coal 19 . Carbon monoxide is also naturally present within the seams of coal (irrespective of coal self-ignition process) and is detected during the course of mining activities at the levels of up to several hundred ppm 20,21 . Fracturing of coal in the process of coal extraction may also itself release gases from coal structure with the values of carbon monoxide concentrations depending on the coal rank, the mode and depths of the exploitation as well as the type of the overlying rocks. Pressure and temperature also have an influence on coal sorption properties [22][23][24][25] .
Therefore, it should be taken into account that the simply mechanical factors may also affect (increase) the values of gas concentrations in a coal mine atmosphere and consequently the correctness of the assessment of coal self-heating status, performed on the basis of carbon monoxide concentrations detected in a mine atmosphere and determining the specified, legally required fire-fighting activities 26 .
In underground mining, the zone of full roof failure includes roof strata of the thickness equal to 1-2 times the thicknesses of the seam. Peng 27 gave the principles of the calculation of the vertical roof failure with respect to rock relaxation coefficient. The increasing amount of rocks fills the post-extraction space, which finally meets the roof strata with no transformation into the roof failure state. The rock debris is bent and squeezed and in consequence cracked and delaminated by the roof strata over the full failure zone, creating the high roof failure zone of the height of 1-1.5 times of the seam thickness. The full and the high roof failure zones form the roof strata of the thickness of 3-3.5 times of the seam thickness, which enables estimation of the force effect of the roof strata on the residual coal. Thus, intensified extraction of coal creates conditions facilitating desorption of carbon monoxide from longer walls, by affecting the range of the stress relief in the surrounding bed, and therefore also the volume of the degassed bed. This makes the applicability of carbon dioxide as a coal self-heating indicator arguable 28 , while the enhanced release of carbon monoxide from the body of coal is considered to be useful in these terms 29 .
Crushing of coal, taking place in the goafs of longwall increases the surface area of coal bed exposed and thus intensifies gas desorption. The depth of coal extraction affects the primary temperature of the rock mass (e.g. in Polish coal mines it varies from approximately 20 to 50 °C), and therefore the rise in carbon monoxide content in a mine atmosphere cannot be considered solely as a certain prove of coal self-heating.
In the light of the above, a new, more precise method of fire hazards assessment, based on the analysis of carbon monoxide concentrations related both to the primary temperature of the rock mass and the depth of a seam is needed, which was also proven experimentally in the study presented in this paper on the basis of the research results acquired with the use of the unique experimental set-up. This test stand was designed to enable simulation of the coal crushing process behind the roof support and under the conditions of the primary temperatures in the workings. Thus, the innovativeness of the approach adopted in the study presented covers two main methodological aspects: taking into account the effect of rock mass pressure (depth of the coal seam) and the temperature, typical for the rock mass in a particular coal mine, and the research installation specially designed and constructed, and applicable in simulation of the process of coal crushing behind the roof support, and under the conditions of the primary temperatures in the workings. The research results presented in the paper are therefore considered to be contributing to the improvement of the fire hazard assessment in coal mines.

Materials and methods
Materials. In order to determine the amount of carbon monoxide emitted at primary temperature of the rock mass, coal was sampled from a freshly uncovered longwalls, at the largest possible mass, and transported to the laboratory in metal, airtight vessels. The physical and chemical characteristics of coal samples are given in Table 1.
Methods. The research method consists in degassing the samples of crushed coal at four temperatures. For the purpose of the study presented in this paper, the following temperatures were predetermined: 20, 30, 40 and 50 °C, reflecting the primary temperature ranges occurring in Polish mines.
The crushing of coal takes place inside a research vessel without external air access. In the course of the experiment, the concentration of carbon monoxide in gaseous samples is determined. The test stand is presented in Fig. 1. The main element of the installation is a hermetic vessel into which samples of 1200 g of coal in grain size 10-30 mm are put, tightly locked and crushed with a crushing punch and the pressing force of 20 MPa. Next the desorption of gasses to the free space of the vessel is allowed to proceed for 24 h and then the tested gaseous mixture is driven away from the vessel by inflating a rubber balloon inside it to avoid diluting the tested mixture with atmospheric air. Therefore, the design of the device allows for the natural crushing of coal, taking into account the pressure of the rock mass. The second important feature of the device is its capability to simulate the primary temperature of the rock mass. A large mass of coal allows eliminating the inhomogeneity occurring www.nature.com/scientificreports/ in coal seams. These features makes the experimental conditions adopted better reflect the the conditions of a coal mine. Shimadzu GC 2014 gas chromatograph (GC) with a TCD detector, a molecular sieve column and a Porapak column for determination of concentrations of O 2 , N 2 , CH 4 , CO and CO 2 is employed.
The content of dry and ash free coal in a sample, z csw , is calculated as 30,31 : where W c -total moisture content, % w/w, A a -ash content, % w/w. Then the mass of dry ash free coal, m csw, is: where m w -coal sample mass, g.
The volumes of coal sample, V w , and free space in the vessel are 1533 cm 3 and 10,350 cm 3 , respectively.
The volume of CO in tested gaseous mixture, V CO , is calculated as: m csw = m w· z csw 100 , g, www.nature.com/scientificreports/ where V 0 -degassed mixture volume, cm 3 , S co -carbon monoxide content in degassed mixture, % v/v. Carbon monoxide content in a sample is: The self-ignition index 32 is also determined.

Results and discussion
The analyses of selected gaseous components release were conducted for 15 bituminous coal samples exposed to 20 MPa, simulating the rock mass pressure at the depth of 800 m and at the temperature of 20, 30, 40 and 50 °C, simulating the rock mass primary temperature in the studied coal seams. The obtained results demonstrated that carbon monoxide desorption increases rapidly with the rise in the rock mass primary temperature and the amount of carbon dioxide released varies for various samples (Fig. 2). Figure 3 presents these values arranged according to the increasing rock mass primary temperature. In general, the contents of carbon monoxide at low primary temperatures (26-36 °C) were low and increased above 41 °C. The exception was reported for samples 4, 5, 9 and 15, which showed high concentrations of carbon monoxide despite the fact that the rock mass temperature occurring in these seams was low and ranged from 21 to 23 °C.
The correlation between the content of carbon monoxide desorbed from a coal sample and the respective moisture content and the self-ignition index value was also observed (Fig. 4).
As it can be seen from Fig. 4, significant amounts of carbon monoxide were desorbed from previously mentioned samples 4, 5, 9 and 15, characterized by the highest values of the self-ignition index (self-ignition groups IV and V) ( Table 1). Interestingly, samples 12, 13 and 14, of relatively low values of the self-ignition index also demonstrated high concentrations of carbon monoxide. In Table 2 the amounts of carbon monoxide recalculated to mass dry ash free coal substance are presented.
The values of the total moisture content in coal samples tested (Table 1) varied from 1.2%w/w for sample 8 to 19.1%w/w for sample 4. It has been previously reported that the sorption capacity of coal may decrease with high ash content 33 . In the study presented in this paper, however, the ash content was low, in the range from 1.6%w/w (sample 3) to 10.1%w/w (sample 5) and therefore no direct impact of this parameter on carbon monoxide desorption was observed.
The oxidation heat depends on the mass of the coal which undergoes the oxidation process. The content of dry ash free coal substance in the examined samples, z csw , was from 72.6%w/w for sample 4 to 96.4%w/w for sample 3 (Table 1). Figure 5 shows volumes of carbon monoxide released for temperature conditions characteristic of particular seams and recalculated to the mass of dry ash free coal substance.
The higher the self-ignition group, the more intensive the emission of carbon monoxide was still observed after the above recalculation, and taking into account the contents of ash and moisture in coal samples tested. Another correlation, not related to the coal self-ignition group, but to the rock mass primary temperature, was also reported for samples 12, 13 and 14. www.nature.com/scientificreports/ Research studies on gas adsorption/desorption on/from bituminous coals reported in the literature most commonly refer to the combination of carbon dioxide and methane 34 . Based on the knowledge on the structure of bituminous coals, and especially on the nature of coal surface as well as the studies on sorption processes, the authors presented interpretations of the research results with reference to the changes occurring during coal self-heating. The conditions of the experiments adopted in these studies, however, poorly reproduce the conditions of a coal mine, to mention only the mass and granulation of coal and vacuum degassing of the samples.
Natural desorption of gases from bituminous coal results in methane release to the atmosphere with smaller amounts of ethane, propane and carbon dioxide [35][36][37] . Extensive studies in this field were performed by Xu et al. 38,39 and Chen et al. 40 . Exposing of coal to air results in parallel processes of coal oxidation and desorption of carbon monoxide and dioxide 41 . The product of early coal oxidation, detected in the highest concentrations and at low temperatures, both under the underground or laboratory conditions is carbon monoxide. Trace amounts of other  www.nature.com/scientificreports/ gases make them difficult to be detected at this stage 42 . In case of natural desorption, the rise in carbon monoxide content takes place. Fan et al. 43 on the basis of the experimental study on the sorption of the primary carbon monoxide in coal seams concluded that its adsorption mechanism is similar to the adsorption of methane. The applicability of the method originally dedicated to measuring the amount of methane, also to carbon monoxide, was demonstrated. The results of the study also confirmed that the primary carbon monoxide is present in Xichuan colliery, and that the amount of this gas changes with the location, reaching the value of 487.79 × 10 -6 m 3 CO /Mg and is not related to the process of coal self-heating. Three explanations for the occurrence of carbon monoxide were given: (i) release in coal extraction, (ii) release from rock strata cracks caused by coal extraction, and (iii) oxidation of residual coal 43 .
Zhang et al. 44 measured the carbon monoxide emission under non-isothermal conditions using an apparatus of their own design and construction, and coal type, grain size and oxygen content in coal as variables adopted in the experimental procedure. They found that the application of coal of larger particle diameters resulted in decreasing the emission of carbon monoxide, particularly at higher temperatures.
Xu et al. 45 also studied the process of coal self-heating at low temperatures. In their experiment, the concentrations of gaseous products as well as the temperature inside the research vessel were measured when the coal Table 2. The content of carbon monoxide recalculated to mass of dry ash free coal substance. www.nature.com/scientificreports/ samples were subjected to heating. Based on data analysis, four stages of the oxidation process were identified: (i) preliminary heating without oxidation reaction or volatilization, (ii) volatilization without oxidation, (iii) unevenly accelerated oxidation with parallel vaporization, and (iv) late heating stage, when the entire volume of oxygen is consumed in the oxidation reaction. Tang 17 also confirmed the correlation between the grain size of coal and carbon monoxide release intensity: the finer the coal sample was ground (1-3 mm) with a ball grinder, the higher the amount of carbon monoxide was produced. The concentration of carbon monoxide released was also lower for coal samples of higher moisture content. Xie et al. 46 found that the presence of certain oxidation products may facilitate identification of particular phases and the intensity of the coal self-heating process. The sequence in which the products occur varied depending on coal types. However, the mechanism of these phenomena, and in particularly the effects of various factors on the initiation and the development of the self-heating process, has been reported to be still unrecognized.
In the research works presented in this paper the share of grains above 0.7 mm constituted above 70% of the coal sample; therefore, the analyses performed with porosimeters would not reflect the coal mine conditions. The experimental stand employed in the study allowed desorbed gas sampling at ambient pressure and with no air-derived diluting effect.
On the basis of the research results presented in this paper, it may be concluded that coal samples of the highest moisture content (samples 4, 5, 9 and 15, Table 1) released the highest amounts of carbon monoxide. However, a multi-dimensional approach is needed to accurately estimate the emissions of carbon monoxide and other gases; the coal self-ignition group or self-ignition index, as well as the rock mass primary temperature occurring at the given seam depth have also to be taken into account. A complex analysis of these dependencies is presented in Fig. 6.
As it has been previously mentioned the reported amounts of carbon monoxide desorbed were high for samples originating from high coal self-ignition groups (IV and V): samples 4, 5, 9 and 15, acquired respectively from seams 212, 215, 206 and 207, at shallow depths, where the rock mass primary temperature only slightly exceeded 20 °C.
Another strong correlation was observed for the primary temperature: samples 12, 13 and 14, although of I and II self-ignition groups, also showed high amounts of carbon monoxide desorbed, which resulted from relatively high rock mass primary temperature values, of 41, 43 and 48 °C, respectively. Under such primary temperatures, coals theoretically not very prone to self-ignition, when taking into account their self-ignition groups, release under the mine conditions the amounts of carbon monoxide comparable to coals particularly prone to the self-ignition process.
To sum up, on the basis of the experimental results presented in this paper, the coal samples studied may be divided into the following groups: -Coals emitting less than 0.001 m 3 CO /Mg csw ; with the rock mass primary temperature below 40 °C and of the self-ignition groups of II and III (samples 1, 2, 3, 6, 7, 10 and 11). www.nature.com/scientificreports/ -Coals emitting over 0.001 m 3 CO /Mg csw ; with the rock mass primary temperature above 40 °C and of the selfignition groups of I and II (samples 12, 13 and 14). -Coals emitting over 0.003 m 3 CO /Mg csw ; with the rock mass primary temperature below 25 °C and of the selfignition groups of IV and V (samples 4, 5, 9 and 15).
The primary temperature of the rock mass surrounding the longwall affects significantly the amounts of carbon monoxide released and therefore testing of coal at laboratory temperature leads to erroneous conclusions. The same coal at the laboratory temperature will emit much lower concentrations of carbon monoxide than at the primary temperature of the rock mass. The higher is the temperature of the rock mass, the greater the error in the correct assessment of fire risk is. Since the concentrations of carbon monoxide and other gases are commonly employed as the indicators of coal self-heating progress, the desorption of gases and the factors affecting its intensity need to be taken into account, especially when coal extraction takes place at increasingly deeper depths.

Conclusions
On the basis of the research results presented in this paper the following conclusions may be drawn: 1. The following correlations between the amount of carbon monoxide released and the physical and chemical characteristics of coal as well as the rock mass primary temperature was observed: -Coal samples 4, 5, 9 and 15, of the highest self-ignition groups (IV and V), were characterized by the highest rate of carbon monoxide desorption, despite the fact that they were acquired from shallow coal seams of relatively low rock mass temperature, of 21-23 °C. -Coal samples 12, 13 and 14, of the lowest self-ignition groups (I and II), but acquired from deeper coal seams, of the rock mass temperature above 40 °C, also desorbed significant amounts of carbon monoxide. -The remaining coal samples, of self-ignition groups II and III, and taken from coal seams of the moderate rock mass temperatures (below 40 °C), emitted considerably smaller amounts of carbon monoxide.
2. When the depth of the coal bed (and the rock mass primary temperature) is taken into account in the method of fire hazard assessment, then the previously employed scheme of coal classification changes: coals of low self-ignition groups, under the conditions of high rock mass primary temperature, emit carbon monoxide in the amounts similar to these characteristic for coals of high self-ignition groups. 3. The rise in the rock mass primary temperature results in intensified natural desorption of gases commonly applied as indicators of the coal self-heating process development. This, in particularly, may result in detection of carbon monoxide concentrations similar to the ones characteristic for coal self-heating. 4. The accurate fire hazard assessment requires considering other than self-heating of coal causes of carbon monoxide emission, including the rock mass primary temperature. 5. The methodological approach proposed in this paper is expected to improve the existing methods of fire risk assessment in coal mines and post-mining areas.

Data availability
All data generated or analyzed during this study are included in this published article.