Revisiting the fast desorption method – Initial gas release from pulverised coal

Gas content measurement in a coal seam is critical to design gas drainage systems and assess mine ventilation requirements, directly affecting coal production rates and thus being required to be measured for outburst management compliance. Gas content is a basis for outburst threshold parameters such as desorption rate index (DRI) and threshold limit value (TLV), primarily influencing the probability and size of a gas outburst. The Australian Standard (AS) 3980-2016 covers two direct methods (slow and fast desorption), being involved with crushing subsamples to quickly release the adsorbed gas into a gas phase to determine the gas content (Q₃) of coal.

This project has tackled two issues in Q₃ measurements by revisiting the fast desorption method. First, along with potential errors associated with gas content measurements, the crushing efficiency and gas release rate are largely affected by crushers and crushing conditions used in each gas testing laboratory. This can lead to a great discrepancy in measuring DRI and Q₃. Second, at gas outburst sites, pulverised coal particles (< 100 µm in size) can often be found, instantly releasing a large amount of gas at the very beginning of an outburst. This suggests that the initial gas release from coal particles is of great significance to understand the occurrence of an outburst. As DRI values are based on the first 30 seconds of gas released at Q₃ testing and related to the total gas content, they are not directly reflecting the very initial gas release for < 10 seconds.

To tackle the first issue, pulverised coal samples after Q₃ testing were obtained from two gas testing laboratories and analysed in terms of particle size distributions (PSDs) and gas adsorption isotherms. The gas content data of those samples from four underground coal mines were also collected.  The two laboratories use different types of crushers and crushing conditions. The material portion passing through the 212 µm mesh was found to be 82.8, 50.3, 55.1 and 49.4 % for pulverised coal samples from mines, all of which did not meet the recommended particle sizes (95 % materials passing through a 212 µm) stated in AS 3980-2016. Also, pulverised samples from Lab 2 had some large particles (> 1 cm), generally containing bigger particles than the ones from Lab 1. As gas diffusion is a function of particle size, the crushing conditions can significantly influence the initial rate of gas released at Q₃ testing, consequently affecting the DRI values. From adsorption isotherm measurements of CO₂ and CH₄, samples from mine A were found to have more diffusional restrictions than the others due to the presence of distinctive peaks at micropore sizes between 4 and 5 Å.

A separate set of gas content data obtained from mine A over 6 years indicated that for given total gas contents (Q_m), lost gas (Q₁) values were found to be very different between CO₂ and CH₄ rich samples whereas Q₂ values were similar. Especially for over 8 m³/t of Q_m, the Q₁ values of CO₂ rich samples were much greater than those of CH₄ rich samples and potentially related to Q₃ values. The Q₃ values of 95 samples in this project were generally a linear function of Q_m, regardless of gas compositions.

For the second issue, a Q₃ measurement system was established at a CSIRO laboratory, using a conventional water displacement method with graduated cylinders and separately allowing direct measurements of pressure, temperature and the volume of gas released. Four types of crushers were tested while monitoring changes in pressure and temperature. Double pucks were found to be more effective crushing than other types in terms of PSDs with a minimum temperature increase (by 0.95 K for 30 seconds of crushing). This insignificant temperature rise means that the gas pressure rise while crushing is mainly due to the increased volume of gas released. Due to unavoidable gas loss during the sample delivery from mine sites to the CSIRO laboratory, the Q₃ values reported by mines were not used. Instead, Q₃ testing for as-received 15 sub-core samples from mine A was separately carried out with both graduated cylinders and volume meters at the CSIRO laboratory. The rates of initial gas released over the first 5-10 (R_5-10), 10 (R₁₀) and 30 (R₃₀) seconds were overall found to be linearly proportional to the measured Q₃ values. Also, for samples with similar Q₃ values, R_5-10 values were greater than R30 values and the difference was increased with Q₃ values. This indicates that the relationship between DRI values and outburst TLVs can be weakened with an increase in Q₃ (or Q_m) values as R_5-10 can be more relevant to outburst occurrence. If the rates of gas released were measured onsite, it is expected that the difference between R_5-10 and R₃₀ would be greater, stressing the importance of initial gas release at Q₃ testing. Onsite measurements are recommended to understand this observation, relating to the total gas content.

Along with the conventional water displacement method which requires slight vacuum at the start, this project used a volume meter to directly measure the volume of gas released at ambient pressure. The Q₃ values were well correlated with but slightly lower than the ones with graduated cylinders due to the initial pressure difference. Unlike the water displacement method, a second rise in the volume of gas released while crushing was found for some samples whereas other samples with similar Q₃ values exhibited a gradual increase. This indicates that other factors such as coal characteristics are involved with the second rise. The early indication was that the micropore size distributions (MPSDs) seem to make a difference in the gas desorption behaviour. The reason(s) behind the second rise needs to be further investigated.