Factors impacting gas content measurements using gas desorption by drilling underground boreholes

Abstract

Accurate determination of the gas content in coalbeds is important for safe mining. Currently, gas desorption by drilling underground boreholes is the most commonly used gas determination method. However, this method is not very accurate and needs to be improved. In this study, we established a laboratory protocol based on coal adsorption studies to analyse factors affecting the measurement accuracy. The results showed that exposure time, sampling method, sample weight, particle size and gas loss estimate significantly affected the gas content measurement using gas desorption by drilling underground boreholes. Longer exposure time and increased particle size resulted in higher relative errors. Sampling by coring is more accurate than sampling by drilling. The higher the sample weight is for samples weighing less than 240 g, the larger the error of the in situ measurements of desorbed gas, residual gas and lost gas is. The error tends to stabilize for heavier samples. The gas losses at different exposure times calculated using the commonly used Barrer model, power function method and negative exponent method were compared. The gas loss error within 0–12 min, computed with the Barrer model, and after 12 min, computed with the power function method, is minimal. The modified formula of gas loss was obtained using the combination of a fitting analysis and the relationship between gas loss and exposure time. Subsequently, the optimal procedure for in situ gas content measurements using gas desorption by drilling underground boreholes was determined. The gas content errors for anthracite, gas coal, lean coal and long flame coal, which were measured using gas desorption by drilling underground boreholes and corrected using the gas loss formula, decreased significantly to less than 10%, thus, meeting the engineering accuracy norm.

Keywords

Gas desorption method by drilling underground boreholes desorbed gas residual gas lost gas gas content error correction

Introduction

Energy released by coal and gas outbursts can not only destroy underground roadway, ventilation facilities and production equipment, resulting in huge economic losses to the mine and great difficulties in resuming production, but is also a serious safety threat for underground workers (Díaz Aguado and González Nicieza, 2007; Yang et al., 2014). Therefore, timely forecasts and coal and gas outburst prevention are very important. Gas content is one of the major parameters used to study coalbed gas. It reveals the internal energy or the risk of gas and coal outbursts of the coalbed and therefore is an important forecast indicator and one of two key parameters for assessing outburst prevention (Cao et al., 2001; Wold et al., 2008; Zhang and Lowndes, 2010). The measurement of gas pressure is infeasible in the majority of coal mines in China. Thus, the gas content is frequently used as an evaluation index for outburst prevention. The test accuracy therefore directly affects mine safety. The gas content determines the reliability of outburst risk prediction and the effectiveness and economy of gas control measures, thus, it plays an important role (Wang et al., 2013).

A large number of methods have been used to measure gas contents. These methods can be classified as intrageological exploration methods and underground measurement methods, according to their application range, and as direct or indirect methods, according to their own characteristics (Yee et al., 1972). Direct methods can be further divided into coal coring, gas desorption by drilling geological boreholes and gas desorption by drilling underground boreholes (GDDUB), according to their differences in sampling tool and compensation mode. Table 1 lists three commonly used gas content determination methods, and their similarities and differences in criteria standard, scope of use and sampling methods, etc. are compared. GDDUB is the most commonly used method among them. GDDUB is mainly divided into three parts: desorption of underground gas, calculation of gas loss and determination of residual gas (Moore, 2012). For the desorption of underground gas, boreholes are drilled to the deep coal seam, the high pressure air blower is used to collect broken cinder into a sampling tank at the borehole orifice, the sampling tank is rapidly closed and connected to the desorption metre, and in situ desorption is performed. The loss of gas content is calculated according to the desorption law. The residual gas content is determined on a sample, which was crushed in a sealed coal-crushing machine. The total gas content in the coal seam is the sum of the three parts. For coal coring, the entire coal sample is obtained from the deep coal seam and placed in the coal tank to determine the gas content (Waechter et al., 2004a, 2004b).

Table 1.

Comparison of different methods for determination of gas content.

Comparison contents	Geological exploration method	GDDUB	Direct method
Standards	GB/T23249-2009 (GB/T23249-2009, 2009)	GB/T23250-2009 (GB/T23250-2009, 2009)	ASTM D7569 (ASTM D7569, 2015)
Range of application	Coalfield exploration stage	Mine shaft construction and production stage	Ground coalbed methane exploration stage
Determination of standard	Dry ash-free basis	Air dried basis, dry ash-free basis	Dry ash-free basis
Sampling method	Sampling drilling rod	Drilling or coring methods	Wireline drill pip
Sample state	Cylindrical or block, more complete	The shape of particles or powder, crushing	Cylindrical or block, more complete

GDDUB: gas desorption by drilling underground boreholes.

However, field results showed that these methods need much data and are affected by several uncertainty factors in practice. Moreover, each step of the determination process might introduce errors to the final result (Creedy, 1986). According to statistics, the error of gas content, determined using GDDUB, reaches values as high as 50% (Chen et al., 2010). Therefore, it is necessary to systematically study factors causing errors in the application of GDDUB and to propose reasonable measures to reduce errors to improve the accuracy and reliability of forecasting coalbed gas emission and outburst accidents and ensure safe coal mining (Liang et al., 2012).

Errors of the gas content measured using GDDUB are mainly due to estimates of in situ gas desorption and gas loss. Therefore, we focused on these two parts in this paper. First, we conducted laboratory experiments to determine the effects of exposure time, sampling methods, particle size and sample weight on the gas content determination. Then, we calculated the gas losses at different exposure times using three different gas desorption models (i.e. Barrer mathematical model (Barrer, 1951), power function method and negative exponent function method) and analysed their calculation errors. Lastly, we conducted a fitting analysis on the experimental results, obtained the correction formula, taking the theoretical relationship of gas loss and exposure time into consideration, and optimized the process for in situ gas content measurements using GDDUB. The research results provide a reliable basis to reduce errors in gas content measurement and prevention of coal and gas outbursts.

Sample preparation and methods

Samples

Samples were collected from the Linhuan No. 9, Taoyuan No. 10, Yangzhuang No. 5, Yima No. 2₃ and Qincheng No. 2 coal mines of the Huaibei coal mining area in the Anhui Province, China. Samples were crushed and partitioned into eight particle sizes, i.e. <0.075, 0.075–0.15, 0.15–0.25, 0.25–0.5, 0.5–1.0, 1.0–2.0, 2.0–4.0, 4.0–6.7 mm, using 200-, 100-, 80-, 60-, 35-, 18-, 10-, 5- and 3-mesh standard sieves, respectively. An industrial analysis and gas adsorption experiments were performed to determine sample moisture (M air dried[ad]), ash (A_ad), and volatile matter (V_ad), Langmuir's volume, Langmuir's pressure (Table 2).

Table 2.

Industrial analysis and adsorption values.

Coal sample	Ro,max (%)	M_ad (%)	A_ad (%)	V_ad (%)	a (cm³/g)	b (MP_a^-1)
Linhuan No. 9	0.78	0.68	12.52	24.41	19.19	1.13
Yangzhuang No. 5	1.73	1.58	45.43	10.15	17.60	1.22
Taoyuan No. 10	0.75	1.38	9.63	36.22	22.88	1.11
Yima No. 2₃	0.56	3.07	22.45	30.1	25.63	1.08
Qincheng No. 2	2.78	2.24	16.92	8.7	49.58	0.95

A = ash; a = Langmuir's volume; ad = air dried; b = Langmuir's pressure; M = moisture; Ro, max = maximum reflectance of vitrinite; V = volatile matter.

Experimental scheme

The experimental equipment mainly included a high-pressure tank, sealing piston, vacuum pump, gas cylinders, press machine and computers. The high-pressure tank was self-designed and developed. At a height of 40 mm from the inner bottom of the tank, four small holes need drilled with a diameter of 16 mm were drilled and distributed uniformly along the perimeter with an internal tapping of 16 mm×1.5 mm. At the centre of the bottom disk of the tank, two small holes with a diameter of 8 mm were drilled along its radial direction to the centre of the bottom disk, where a small hole with diameter of 12 mm and internal tapping of size 12 mm × 1.5 mm was drilled. This hole was designed to install both the gas injection and exhaust pipes of the simulation borehole. The sealing piston was connected to the press machine using eight screws, two Y-type gaskets and one O-ring. The device withstood seal gas pressures greater than 10 MPa. Figure 1 shows the device design.

Effects of sampling method: We examined the effects of two sampling methods on the gas content measurements, drilling and coring. The experimental drilling procedure is as follows. (1) Place 1 kg of Linhuan No. 9 coal sample with 0.5–1 mm particle size into the high-pressure tank and turn on the press machine to drive the sealing piston that seals the adsorption tanks for briquette compression; (2) connect the vacuum pipeline to the gas injection and exhaust the pipe valves of the tank and turn on the vacuum for 6 h of desorption; (3) connect the gas injection pipe to the gas cylinder to inject the high-pressure gas for 12 h to reach adsorption equilibrium and record gas pressure using a gas collection system at room temperature and atmospheric pressure; (4) lift the sealing piston using the press to expose the coal sample in the tank to air and simultaneously record the exposure time; (5) after 3 min of exposure, break coal samples, rapidly pack them into the sample tank, close the tank tightly, make sure no gas is leaking and measure the amounts of gas desorption and residual gas (Laxminarayana and Crosdale, 1999); and (6) repeat all steps nine times to test the coal samples exposed for 5, 7, 10, 12, 15, 20, 30, 50 and 70 min, and record the results.

Figure 1.

Schematic of coal sample adsorption experimental setup. 1 – press pump; 2 – press; 3 – PC; 4 – data collector; 5 – sealing piston; 6 – Y-type gasket; 7 – O-type seal; 8 – high pressure tank; 9 – pressure transmitter; 10 – coal samples; 11 – gas injection and exhaust pipes; 12 – precision pressure gauge; 13 – gas cylinders.

The experimental coring procedure is as follows. Steps 1–3 are the same as those of the drilling method; (4) use the coring pipe for sampling at an exposure time of 3 min, rapidly place the intact coal sample into the sample tank, tightly close the tank, make sure no gas is leaking and measure the amounts of gas desorption and residual gas; and (5) repeat all steps nine times to test the coal samples exposed for 5, 7, 10, 12, 15, 20, 30, 50 and 70 min, and record the results.

Experiment on the effects of sample weight: One kilogram of Linhuan No. 9 coal with a particle size of 0.5–1 mm was used for briquette compression and gas adsorption equilibrium. At an exposure time of 3 min, samples weighing 30, 60, 90, 120, 150, 180, 240, 300, 350 and 500 g were drilled, and gas desorption and residual gas were determined in situ.

Experiments on the effects of sample size: Linhuan No. 9 coal which was split into eight samples with a particle size (<0.075, 0.075–0.15, 0.15–0.25, 0.25–0.5, 0.5–1.0, 1.0–2.0, 2.0–4.0 and 4.0–6.7 mm) was used for briquette compression and gas adsorption equilibrium. After 3 min of exposure, samples were rapidly packed into the tank, and gas desorption and residual gas were quantified in situ.

Determination of actual gas content

To calculate the error in gas content measured using the direct method, it is necessary to first determine the actual gas content, i.e. to determine the gas adsorption constants a and b in the laboratory (Table 1). The actual gas content was calculated indirectly as the sum of free gas content and adsorbed gas content (Xu andWang, 2011), that is

X = X_{x} + X_{y}

(1)

where X_a and X_f are the adsorbed gas content and free gas content, respectively, in m³/t.

The adsorbed gas content of coal X_a can be calculated using the Langmuir's equation considering the impact of moisture, ash, temperature and combustible material percentage, that is

X_{x} = \frac{abp}{1 + bp} \times e n (t_{0} - t) \times \frac{100 - A_{ad} - M_{ad}}{100} \times \frac{1}{1 + 0.31 M_{ad}}

(2)

where a is the Langmuir's volume, m³(t r); b is the Langmuir's pressure, MPa⁻¹; p is the coalbed gas pressure, MPa; A_ad is the ash of coal, %; M_ad is the moisture of coal, %; t is the coalbed temperature, ℃; t₀ is the temperature of the laboratory at the time of adsorption constant measurement, ℃; and n is the coefficient, n = 0.02/(0.993 + 0.07p).

The free gas content of coal X_f can be calculated using the equation of state of gas

X_{y} = \frac{Vp T_{0}}{T p_{0} ξ}

(3)

where V is the volume per unit coal mass, m³/t; P₀ is the pressure under standard conditions, P₀ = 0.101325 MPa; P is the gas pressure, MPa; T₀ is the absolute temperature under standard conditions, T₀ = 273.2 Kt; T is the absolute temperature of the gas, K; and ξ is the compression coefficient of the gas.

Bringing the actual pressure and temperature of the gas and the experimentally derived adsorption constants a and b in equations (1)–(3), the actual gas contents of the five coal samples at different pressures were calculated. The gas contents were used for comparison with data from other methods.

Experimental results

The exposure time, sampling methods, sample weight, sample particle size and other factors directly affected the measured gas content. The experimental results were analysed to study the characteristics of these effects on the measured gas content and to find a method for error reduction.

Effect of exposure time on the measured gas content

The main error of the gas content measurement using GDDUB is introduced by the calculation of gas loss, which is closely related to the exposure time. The longer the exposure time is, the greater the error of estimated gas loss. Thus, the relationships of desorbed, lost and residual gas contents of coal at different exposure times and the exposure time were measured. Figures 2 and 3 show the experimental results from drilling and coring methods, respectively.

Figure 2.

Relationships of desorbed, residual and lost gas contents measured using the drilling method to the exposure time.

Figure 3.

Relationships of desorbed, residual and lost gas contents measured using the coring method to the exposure time.

There is no significant difference in the time curves of the two sampling methods (Figures 2 and 3). In other words, the relationships of desorbed, residual and lost gas contents and the exposure time are similar for different sampling methods.

Within 0–20 min of exposure time, the measured lost gas content increased significantly with increasing exposure time. Within 20–70 min, the lost gas content increased with increasing exposure time at a slower rate and eventually reached a stable state. The change in the slope of the fitted curve declined, indicating that the rate of gas desorption also decreased over time.

Within 1 h of on-site desorption, the desorbed gas content decreased with increasing exposure time. The decline rate was significantly greater at exposure times less than 30 min. The change in the slope of the fitted curve declined, indicating that the rate of gas desorption also decreased over time.

The amount of residual gas decreased gradually with increasing exposure time and eventually reached equilibrium.

Effect of sampling methods on the measured gas content

The heat produced from friction-induced collision could result in changes in temperature at the time of sampling. The temperature difference due to different sampling processes affects the determination of desorption gas content and gas content loss. The desorption rate is accelerated, and the loss of gas content increases when the temperature rises (Yue et al., 2015). In order to avoid the effects of temperature change on experimental results, we placed the coal tank in a water bath at a constant temperature (20 ℃) to determine the gas content of samples collected using different methods.

Figure 4 shows the relationship between the relative error of gas content and the exposure time measured using drilling and coring methods.

Figure 4.

Relationships of the relative error of gas content measured using the drilling and coring methods to the exposure time.

The relative error of gas content measurements using both drilling and coring methods increases with increasing exposure time. The results determined using both methods differed little within 15 min, while the difference in the results of both methods increased as the exposure time increased, with the coring method being more accurate. The gas contents determined using the coring method were more reliable in practical underground applications with longer sampling times (Yang et al., 2010).

The error rate increased slower within 15 min and then increased quickly within 15–70 min using the drilling method. Therefore, the linear fittings of errors before and after 15 min have the following fitting equations using 15 min as the demarcation point

W = (0.129 t + 16.31) \times 100 %, R 2 = 0.991

(4)

W = (0.23 t + 15.81) \times 100 %, R 2 = 0.963

(5)

where W is the relative error, %; t is the exposure time, min; and R² is the correlation coefficient. Note that the correlation coefficients all are greater than 0.96, showing a higher fitting degree. From the error formula one has

W = (1 - \frac{Q_{1}}{Q_{2}}) \times 100 %

(6)

Q_{2} = \frac{Q_{1}}{1 - W}

(7)

Bringing equations (4) and (5) into equation (7), one obtains

t < 15 min, Q_{2} = \frac{Q_{1}}{1 - (0.129 t + 16.31) \times 100 %}

(8)

t > 15 min, Q_{2} = \frac{Q_{1}}{(1 - 0.23 t + 15.81) \times 100 %}

(9)

where W is the relative error, %; Q₁ is the tested gas content, cm³/g; and Q₂ is the actual gas content, cm³/g.

The linear fitting of the error curve measured using the coring method has the following equation

W = (0.147 t + 13.249) \times 100 %, R 2 = 0.986

(10)

where the correlation coefficient equals 0.986, indicating a higher degree of fitting. From the error formula, the actual gas content can be deduced

Q_{2} = \frac{Q_{1}}{(1 - 0.145 t + 13.33) \times 100 %}

(11)

Once the gas content is measured, the more accurate actual gas content can be derived from the corresponding relative error formula at the selected exposure time.

Effect of sampling weight on the measured gas content

The weight of the samples from the field has an effect on the measured gas content. Samples with different weights were experimentally tested at an exposure time of 1 h in order to study the effects of sample weight; gas content; and desorbed, lost and residual gas contents (Figure 5).

Figure 5.

Relationship of desorbed, residual and lost gas contents to the sample weight.

The desorbed, residual and lost gas contents measured using GDDUB increased with increasing sample weight. The rate of increase notably reduced and eventually stabilized once the sample weight reached 240 g. Figure 6 shows the relationship between gas content (and error) and the sample weight determined using GDDUB.

Figure 6.

Relationships of gas content and error determined using GDDUB to the sample weight. GDDUB: gas desorption by drilling underground boreholes.

The coalbed gas content determined using GDDUB increased with increasing sample weight. The rise in gas content with increasing sample weight was greater for samples weighing 0–240 g and became smaller and stabilized for samples weighing 240–500 g. Similarly, the relative error of the measured gas content decreased with increasing sample weight and eventually stabilized.

Linear fitting of errors for samples weighing 0–240 g has the following equation

W = - 20.46 \ln (m + 1) + 126.07, R 2 = 0.992

(12)

The correlation coefficient was 0.992, suggesting a high degree of fit. By bringing equation (12) into equation (7), the resulting equation is

As 0 < m < 240 g, Q_{2} = \frac{Q_{1}}{1 + [20.46 \ln (m + 1) - 126.07] \times 100 %}

(13)

The analysis shows that a sample weight of 240 g was most suitable for the underground determination of the coalbed gas content. If the sample weight was less than 240 g, equation (13) could be used for correction of the gas content.

Effect of sample size on measured gas content

The microporous structure of coal is one of the main factors affecting the gas content. The micropores have a significant internal surface area, which adsorbs free gas. When coal is broken, the micropore structure is destroyed and exposed to the local atmosphere. The residual gas content depends on the degree of breakage (Jiang et al., 2013; Wei et al., 1997). For large sample particle sizes the centre is far away from the coal surface, thus, gas emitted from the micropores also passes a longer distance. When only considering coal particle size, the coal sample with a larger particle size has therefore a slower adsorption process, smaller gas desorption rate and greater residual gas content (Wu et al., 2011; Zhou et al., 2012).

In order to study the effect of coal particle size on the measured gas content, the desorbed, lost and residual gas contents of coals with different particle sizes were experimentally measured at an exposure time of 3 min and an on-site desorption of 1 h (Figure 7).

Figure 7.

Relationship of desorbed, lost and residual gas contents to the coal particle size.

The on-site desorbed gas content and the lost gas content of the Linhuan No. 9 coal sample measured using GDDUB declined with increasing coal particle size, while the residual gas content increased with increasing particle size (Figure 7).

The relative error of the gas content reduced with increasing coal particle size (Figure 8). For samples with sizes smaller than 1 mm, the rate of increase in error increased with decreasing particle size. However, for samples with particle sizes larger than 1 mm, the rate of increase in error decreased with decreasing particle size and was significantly smaller (Xu et al., 2015).

Figure 8.

Relationship of the relative error of gas content determined using GDDUB to the coal particle size. GDDUB: gas desorption by drilling underground boreholes.

Based on our analysis, the smaller the particle size of coal samples, the larger the initial desorption rate, lost gas content, residual gas content and the relative error and vice versa. Using the drilling method for measurement of the coalbed gas content, samples with particle sizes should be larger than 1 mm because of severe coal breakage.

Selection and verification of models for calculating gas content under multiple factors

The above-mentioned experiments showed that the rate of gas desorption from coal samples was very high in the initial exposure period, while in situ sampling by drilling requires a longer time. In other words, the collected coal sample will inevitably be exposed to air for a period of time before it is sealed into the tank. Gas is rapidly desorbed and escapes into the air during this period, making it difficult to be collected (Li and Yang, 2012). Therefore, the accurate and effective determination of the gas loss is critical for the accurate determination of the original gas content of the target coalbed (Diamond and Schatzel, 1998).

Comparison of lost gas content estimated using different calculation models

It is difficult to directly measure the gas using current technologies. Thus, it is often estimated or indirectly obtained. Because the estimation will introduce various errors (Yan, 2010), it is necessary to compare the gas loss deduced from different calculation methods and use these methods to reduce the errors.

In order to obtain a more accurate estimation formula, under the same sampling method (the drilling method), sample weight (240 g), sample particle size (1–2 mm) and different exposure time (3, 5, 10, 15, 30 and 60 min), we compared and analysed the gas losses estimated using the three most common methods, namely the Barrer method (√t method), the power function method and the negative exponential function method.

The experimental data were processed using the Barrer model, the power function method and the negative exponential function method and their corresponding formulas described earlier to calculate gas loss

Q_{1}

. The actual gas loss can be obtained subtracting the gas desorption content and the residual gas content from the calculated gas loss

Q_{2}

. The error can be derived from inserting the actual and calculated gas losses into equation (7). The results are shown in Table 3 and Figure 9.

Table 3.

Experimental data of adsorbed gas amount determined at different exposure times.

Name		Barrer model	Power function method	Negative exponent function method
Formula		$V = a \sqrt{t_{0} + t} - b$	$q_{t} = q_{0} \times (1 + t) - n$	$r = r_{0} \times e - kt$
3 min	Calculated loss	80.04 ml	60.99 ml	47.91 ml
	Error	8%	12%	18
	Actual loss	92.50 ml
5 min	Calculated loss	106.85 ml	64.84 ml	54.88 ml
	Error	13%	16%	23%
	Actual loss	114.90 ml
10 min	Calculated loss	146.28 ml	103.66 ml	85.06 ml
	Error	24%	26%	31%
	Actual loss	185.36 ml
15 min	Calculated loss	153.08 ml	110.28 ml	102.46 ml
	Error	38%	31%	37%
	Actual loss	263.32 ml
30 min	Calculated loss	163.723 ml	343.97 ml	170.25 ml
	Error	55%	34%	48%
	Actual loss	410.36 ml
60 min	Calculated loss	171.95 ml	439.17 ml	279.98 ml
	Error	58%	38%	50%
	Actual loss	579.50 ml

Figure 9.

Adsorption curve of coal samples at different exposure times.

Figure 9 shows that within 0–12 min, the error of gas loss estimated using the Barrer model is the smallest. After 12 min, the error of gas loss estimated using the power function is the smallest. The fitting curve of the gas loss with time finds the following:

The error formula of the Barrer model is

W = 0.187 ln (t) - 0.1465

(14)

The error formula of the negative exponential function is

W = 0.116 ln (t) + 0.0521

(15)

The error formula of the power function is

W = 0.09 ln (t) + 0.0334

(16)

By bringing equations (14) to (16) into equation (7) as well as the three equations in Table 3, one can obtain the following formula:

The correction formula for the Barrer model is

Q'_{2} = \frac{b}{1.1465 - 0.187 \ln (t)}

(17)

The correction formula for the negative exponential function is

Q'_{2} = \frac{\frac{r_{0}}{k} (1 - e - kt)}{0.948 - 0.116 ln (t)}

(18)

The correction formula for the power function is

V' = \frac{q_{0} [\frac{(1 + t) 1 - n - 1}{1 - n}]}{0.967 - 0.09 \ln (t)}

(19)

where

Q_{2}^{'}

is the corrected actual gas content, cm³/g; b is the undetermined constant; r₀ is the rate of gas desorption at t = 0, ml/s; k is a constant; t is the exposure time of the coal samples, min; V′ is the amount of corrected gas loss, cm³; q₀ is the rate of gas desorption at t = 0, cm³/min; and n is the attenuation coefficient of the gas desorption rate, 0 < n < 1.

When calculating gas loss, the Barrer model combining equation (17) should be used for samples with exposure times less than 12 min, and the power function method combining equation (19) should be used for samples with exposure times greater than 12 min.

Verification

To verify our results, we conducted experiments in the laboratory using four different coal samples, including anthracite, gas coal, lean coal and long flame coal. The measured gas content was corrected using the weight correction of equation (13) and the loss correction of equation (17). The results are listed in Table 4. The errors were calculated comparing the gas content obtained from the proposed method and that from the conventional method. The relative error rate was used as index to verify the reliability and accuracy of each method. Figure 10 shows the histogram of the gas contents of different coal samples.

Table 4.

Gas content of different coal samples.

Computing result	Yangzhuang No. 5	Taoyuan No. 10	Yima No. 2₃	Qincheng No. 2
Actual gas content (cm³/g)	5.38	6.87	7.71	15.70
Measured gas content (cm³/g)	3.29	4.07	4.54	9.63
Corrected gas content (cm³/g)	5.07	6.21	7.00	14.75
Relative error (%)	38.75	40.74	41.14	38.64
Corrected relative error (%)	5.77	9.60	9.27	6.04

Figure 10.

The relative error in measuring gas content error of different Ro,max.

Coals with a higher degree of metamorphism have higher maximum vitrinite reflectances (Ro, max; %) (GB/T23249-2009, 2009). Meanwhile, the relative errors before and after correction are reduced with increasing Ro, max, which suggests that a higher degree of metamorphism has less impact on the measured coal gas content. Table 4 shows that the error of the gas content determined using GDDUB alone was as large as 40%, while the error of the gas content obtained using the proposed method was smaller than 10%, indicating that the proposed method was more accurate and could meet the actual accuracy requirements. Overall, GDDUB along with the correction formula is very reliable for measurements of the coalbed gas content.

Conclusions

In this work, we established a laboratory protocol, which was used to study the effects of exposure time, sampling method, sample weight, sample particle size and gas loss estimation on the coalbed gas content measured using GDDUB. The following main conclusions are made:

With increasing exposure time, (a) the lost gas content increased, while its rate decreased and stabilized after a certain period of time; (b) the desorbed gas content decreased, and the residual gas content slowly reduced and eventually stabilized; and (c) the relative error of the gas content increased. The measured gas content can be corrected in practice using the fitting equations (8), (9) and (11).

When the exposure time is within 15 min, the gas contents measured using the drilling method and the coring method differ slightly. With increasing exposure time, the gas content measured using the coring method was more accurate than that using the drilling method. The coring method is more reliable for underground in situ measurements of gas content with longer sampling times.

The measured gas content increased with increasing sample weight. The gas content of the samples weighing 0–240 g increases with increasing sample weight, while the gas content of samples weighing 240–500 g tends to be stable. Samples with weights of 240 g are more appropriate for underground in situ measurements. If the sample weight is below 240 g due to practical field constraints, equation (13) can be applied to improve the accuracy of the gas content measurement.

The relative error of the gas content declined with increasing sample particle size. Using the drilling method for measurement of the coalbed gas content, samples with particle sizes should be larger than 1 mm.

When estimating the gas loss, the Barrer model along with its correction formula should be used to determine the gas loss of samples exposed for less than 12 min, while the power function along with its correction formula should be used to compute the gas loss of samples exposed for more than 12 min.

Footnotes

Acknowledgements

We acknowledge the editors for insightful suggestions on this work. And our deepest gratitude goes to the reviewer Dr Tim A Moore and the anonymous reviewers for their careful work and thoughtful suggestions that have helped improve this paper substantially.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by the National Natural Science Foundation of China (No. 51304189), the Program for Innovative Research Team in University of Ministry of Education of China (IRT13098). This work is also funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions.

References

ASTM D7569 (2015) Standard practice for determination of gas content of coal—direct desorption method. Available at: https://www-astm-org.web.bisu.edu.cn (accessed February 2015).

Barrer

(1951) Diffusion in Solid, Oxford: Cambridge University Press.

Cao

Glick

(2001) Coal and gas outbursts in footwalls of reverse faults. International Journal of Coal Geology 48: 47–63.

Chen

Zhou

(2010) New technology of directly measuring gas content of stress-releasing and sealed coal core. In: Liu

Hua

Yuan

(eds) International Mining Forum 2010: Mine Safety and Efficient Exploitation Facing Challenge of the 21st Century, Segovia: Scientific Research Publishing, pp. 38–42.

Creedy

(1986) Methods for the evaluation of seam gas content from measurements on coal samples. Mining Science and Technology 3(2): 141–160.

Diamond

Schatzel

(1998) Measuring the gas content of coal: A review. International Journal of Coal Geology 35(1): 311–331. .

Díaz Aguado

González Nicieza

(2007) Control and prevention of gas outbursts in coal mines, Riosa–Olloniego coalfield, Spain. International Journal of Coal Geology 69: 253–266.

GB/T23249-2009 (2009) Coalbed Gas Content Measurement Methods in Geological Exploration Period, Beijing: Standards Press of China.

GB/T23250-2009 (2009) The Direct Method of Determining Coalbed Gas Content in the Mine, Beijing: Standards Press of China.

10.

Jiang

Cheng

(2013) Research on effective sampling time in direct measurement of gas content in Huaibei coal seams. Journal of Mining and Safety Engineering 1: 143–148.

11.

Laxminarayana

Crosdale

(1999) Role of coal type and rank on methane sorption characteristics of Bowen Basin, Australia coals. International Journal of Coal Geology 40: 309–325.

12.

Yang

(2012) The comparative analysis of gas loss quantity calculation method. Safety in Coal Mines 9: 166–168.

13.

Liang

Sun

(2012) Technical evaluation system of co-extraction of coal and gas. International Journal of Mining Science and Technology 22: 891–894.

14.

Moore

(2012) Coalbed methane: A review. International Journal of Coal Geology 101: 36–81.

15.

Waechter NB, Hampton III GL and Shipps JC (2004a) Overview of coal and shale gas measurements: Field and laboratory procedures. In: Proceeding of the 2004 international coalbed methane symposium, 17pp. Tuscaloosa: The University of Alabama.

16.

Waechter NB, Hampton GL III, Shipps JC, et al. (2004b) Comparison of lost gas projections in coalbed methane, Rocky Mountain Section AAPG Poster Session, Denver, Co, p. unpaginated.

17.

Wang

Cheng

Yuan

(2013) Gas outburst disasters and the mining technology of key protective seam in coal seam group in the Huainan coalfield. Natural Hazards 67: 763–782.

18.

Wei

Jia

Yang

(1997) Research into the deduced content loss of methane by making use of the direct coal-bed methane content measurement. Safety in Coal Mines 7: 9–11.

19.

Wold

Connell

Choi

(2008) The role of spatial variability in coal seam parameters on gas outburst behaviour during coal mining. International Journal of Coal Geology 75: 1–14.

20.

Cheng

Feng

(2011) Determination of gas pressure and gas content based on residual gas quantity of coal sample. Journal of Mining and Safety Engineering 2: 315–318.

21.

Tang

Zhao

(2015) A new laboratory method for accurate measurement of the methane diffusion coefficient and its influencing factors in the coal matrix. Fuel 158: 239–247.

22.

Wang

(2011) The development of the coal and gas outburst prediction positional sampler. Procedia Engineering 26: 1495–1501.

23.

Yan AH (2010) Study on multisource data analysis and prediction of gas content. PhD Thesis, China University of Mining & Technology.

24.

Yang

Tang

(2014) Effects of low molecular weight compounds in coal on the characteristics of its spontaneous combustion. Canadian Journal of Chemical Engineering 93: 648–657.

25.

Yang

Qin

Wang

(2010) Desorption-diffusion model and lost gas quantity estimation of coalbed methane from coal core under drilling fluid medium. Science China Earth Sciences 53(4): 626–632.

26.

Yee

Seidle

Hanson

(1972) Gas sorption on coal and measurement of gas content. Fuel 51: 1–66.

27.

Yue

Wang

Tang

(2015) Physical simulation of temperature influence on methane sorption and kinetics in Coal (II): Temperature evolvement during methane adsorption in coal measurement and modeling. Energy and Fuels 19(10): 6355–6362.

28.

Zhang

Lowndes

(2010) The application of a coupled artificial neural network and fault tree analysis model to predict coal and gas outbursts. International Journal of Coal Geology 84: 141–152.

29.

Zhou

Luo

(2012) Discussion of gas loss amount when underground drilling cuttings desorbed gas. Safety in Coal Mines 4: 109–111.