Field and Numerical Modelling Investigations on the Stability of Underground Strata during Longwall Workings

Understanding the behaviour of underground workings is essential for the success of any mining 12 method. The longwall mining method is one of the predominant underground methods to extract coal. 13 Since 1978, in India, 22 underground coal mines of different collieries have been implemented the 14 mechanized longwall method. SCCL is one of that colliery has mixed working experiences with 15 longwall method in their mines. The longwall faces in GDK-10A, JK-5, and VK-7 of SCCL had 16 produced good results, but the faces in GDK-7, GDK-9, GDK-11A, and PVK-5 had suffered due to the 17 geological disturbances and unavailability of real-time information about the strata behaviour. By 18 addressing the previous experiences of longwall workings, Singareni Collieries Company Limited 19 (SCCL) has implemented a high capacity (2 × 1152T) powered support system in Adriyala Longwall 20 Project (ALP) at a depth of 375m. In this study, extensive field monitoring with different strata 21 monitoring instruments was conducted in ALP to analyze the gate roads convergence, stress variation 22 on longwall and chain pillars at different stages of extraction (i.e., 8m, 25m, 35m, and 45m) and the 23 pressure variation on the powered support systems. It was observed from the convergence in the gate roads was increasing with the advance of the longwall face and the area of 25 exposure. The pressure of the legs on the dip side was less than the pressure of the legs on the rise side, 26 which implies a stable roof condition over the longwall face. To better understand the behaviour of 27 ALP workings, a numerical modelling study with FLAC 7.0 has been conducted with actual physio- 28 mechanical properties. The computed numerical modelling results have been remarkably well in 29 consistent with the field monitoring results. The stability of chain pillars has been estimated at every 30 stage of extraction by the Factor of Safety (FoS) criterion and it was found that the pillars could be 31 ensured stability in longwall workings. 32

convergence in the gate roads was increasing with the advance of the longwall face and the area of 25 exposure. The pressure of the legs on the dip side was less than the pressure of the legs on the rise side, 26 which implies a stable roof condition over the longwall face. To better understand the behaviour of 27 ALP workings, a numerical modelling study with FLAC 7.0 has been conducted with actual physio-28 mechanical properties. The computed numerical modelling results have been remarkably well in 29 consistent with the field monitoring results. The stability of chain pillars has been estimated at every 30 stage of extraction by the Factor of Safety (FoS) criterion and it was found that the pillars could be 31 ensured stability in longwall workings. 32 in the longwall panel-1 of ALP. The details of the panel can be seen in Table 1, and the section of the 74 borehole representing the layers of the seam-1 can be seen in Figure 2. 75

Field instrumentation results and the assessment of longwall workings behaviour 77
The gate roads, longwall face, and the powered support system of the longwall panel-1 have been 78 monitored with different strata instruments. Convergence indicators, Telltale extensometers, load cells, 79 PMC-R (Programmable Mining Controller-Roof Support), and stress cells were used in the monitoring 80 study. Figure 3 shows the instruments locations in the longwall panel-1. Round the clock the 81 monitoring was done to observe the periodic weighting phenomenon, induced stress distribution and 82 convergence in the gate roadways. Trigger Action Response Plan (TARP) was prepared for the 83 development to act with the necessary secondary support system to ensure the roof's stability in case of 84 any trigger like more convergence or geological disturbance occurs. 85

a) Roof weighting phenomenon 86
Local falls, main falls, and periodic weightings were observed in the panel during retreating of the 87 longwall face. The first local fall occurred after a retreat of 10.75m. The main fall of area 12,215m 2 88 occurred after a retreat of 39.9m from the mine boundary. Heavy sounds were observed during the main 89 fall in the goaf. A detailed note of such events has given in table 2. 90 Main fall 39.9 12,215 -Sounds were heard in the goaf.

b) Vertical stress variation on longwall and chain pillars 92
Stress cells were installed at every 100m intervals in the longwall pillar and every 200m intervals in the 93 chain pillars. The in-situ stress at a depth of 375m is 9.35MPa. On longwall pillar, the stress begins 94 drop though after reaching a maximum value of 1010.35 KPa (9.35MPa is being in-situ) at a distance 95 of 13.2 m from the measuring station. Nevertheless, in the case of chain pillars stress readings, it was 96 observed that the stress rises steadily before becoming almost constant at 43.2 m from the face with 97 maximum stress of 580.41 KPa (9.35MPa is being in-situ). The variation of induced stress on the 98 longwall pillar and chain pillar can be seen in Figures 4 and 5, respectively. 99

c) Convergence of the gate roadways 100
The roof was monitored with convergence indicators installed for every 25m interval in the gate roads, 101 two-point telltale extensometers were installed for every 50m interval, i.e., at the junctions of cut-102 throughs, and the monitoring was done on a daily basis.
From the observations of the instruments, it was reported that the cumulative convergence remains 106 fairly constant and negligible at a distance of 25 m and 50 m from the measuring stations but shows a 107 gradual increase as the face retreats nearer to the measuring stations, as shown in the Figures 6 and 7. 108 A minimum of 1mm convergence to a maximum 12mm convergence was noticed in the MG. The 109 maximum convergence observed in the MG is 12mm after 75m of longwall face extraction, as shown 110 in Figure 8. 111

Tail Gate-1 (TG-1) convergence 112
The convergence instruments installed along the TG-1 showed a similar convergence pattern as MG, 113 as discussed in the earlier section. The maximum convergence observed was 8mm at 75m of longwall 114 face extraction, as shown in Figure 11. The variation of convergence at 25m, 50m and 75m of longwall 115 face extraction can be seen in Figures 9 to 11. 116

Tail Gate-2 (TG-2) convergence 117
The convergence readings vary from 1 mm to 6 mm at all convergence stations, as shown in Figures  118 from 12 to 14. The convergence readings recorded at TG-2 are considerably less than those recorded at 119 MG. The maximum convergence of 6 mm was recorded after 75 m longwall face extraction. 120

d) Dip and Rise side pressure observations on Powered supports 121
The pressures of the chock shields are continuously monitored with PMC-R (Programmable Mining 122 Controller-Roof Support) and from the control center at the surface. All the data will be transferred to 123 the surface to monitor all the chock shields' pressures across the face, and it can store data for each 124 second. Total 146 powered supports installed along the 250m length of the face were divided into three 125 sections as Top section (111 to 146), Middle section (38 to 110), and Bottom section (1 to 37). Among 126 7 distribution of pressure on these chock shields can be seen in Table 3. From the data obtained from the 129 PMC-R, it was observed that the load developing on the powered supports is not symmetrical along 130 the longwall face. The Middle section of the chock shields will take a greater load than the other two 131 sections. The variation of pressures developing on the powered supports can be seen in Figures 15 to  132 16. 133

Development of numerical models with FLAC 7.0 135
Evaluating the rock mass behaviour in underground workings can possible by the Numerical 136 modelling technique. In this study, 2D numerical models were prepared by considering the geo-mining conditions of the seam-1 of ALP by using FLAC 7.0 software. FLAC 7.0 of Itasca Consulting Group Inc. [10] is relatively used for solving problems related to tunneling and geotechnical engineering. To 139 understand the behaviour of longwall workings, the side section of longwall panel-1 was considered 140 for modelling. The length of the model is 300m along X-direction and 100m height in Y-direction as 141 shown in Figure 17. 3m thickness of coal is left as the immediate roof for the workings and the 142 truncated load is applied according to the 375m depth of workings (i.e., 375m-(100m-143 (3.5m+25m))=303.5m, truncated load is 7.58MPa)). In the first phase, 8m width of coal has been 144 extracted for installation of the powered support system of capacity Five different numerical models were developed by considering the lithology and physico-mechanical 159 properties shown in Figure 18. These models were developed to evaluate the displacement profile and boundary, as shown in Figures 19, 20, 21, and 22, respectively. After attaining the initial equilibrium 162 position to the model, the coal part's 8m width has been removed from mine boundary by assigning 163 the null model to it as shown in Figure 19.

Convergence in the gate roadways 166
After the in-situ model has been simulated, 8m width of coal has been removed to install the support 167 system at the longwall face. After 8m of extraction, no or significantly less convergence in the roof 168 has been noticed. The value of convergence being observed is 1 mm. The same scenario has been 169 noticed up to the width of 25m. The maximum convergence in the roof strata has been observed at 170 40m coal extraction, i.e., 10mm. The progressive advance of longwall face with powered support 171 system has resulted in less development of convergence in the roof strata. The variation of convergence 172 in the roof strata at different stages of extraction can be seen in Figures 24 to 27. 173

Vertical stress observations 174
The in-situ stress could be developed on the longwall pillar at a depth of 375m is 9.35MPa. At the 175 first stage of extraction, i.e., 8m of coal face advance, the maximum vertical stress developed on the 176 longwall pillar is 10.35 MPa (9.35MPa is in-situ stress). As the face advances further, the vertical 177 stress developing on the longwall pillar is also increasing, as shown in Figures  3.0, and 2.6, respectively. By analyzing these results, it could be ensured that the chain pillars will be 203 stable to protect the underground workings. The FoS of chain pillars at different extraction stages can 204 be seen in Figure 32. 205

Validation of field and numerical modelling results 206
It is always good to validate the numerical modelling results with field instrumentation results. 207 Convergence noticed in the gate roads and the vertical stress developed on the longwall pillar results 208 were validated. Figure 33 shows the comparison of field and numerical modelling results for 209 convergence of gate roads at different extraction stages. From the validation results, it was observed 210 that the results obtained from both the studies were good in agreement and are encouraging. 211 The abutment stresses obtained from the field study were compared with the data obtained from the 212 numerical modeling results. They are plotted in graphs to study their nature and interrelation. induced vertical stress of 13.85MPa (9.35MPa being the in-situ stress) was observed on the chain pillars 234 at a distance of 40m from the mine boundary. Based on the Factor of Safety (FoS) criterion, it was 235 assessed that the chain pillars could provide stability to longwall workings. 236