Enhancing the safety of remnant pillar recovery in lignite by numerical modelling

CIM Bulletin, Vol. 97, No. 1082, 2004

C.O. Aksoy, H. Kose, E. Yalcin, K. Heasley and C. Mark

In Turkey, the growing demand for lignite has resulted in increased lignite production and an emphasis on improving underground recovery. This has resulted in gradual improvements in mining methods and recoveries, and the extraction of remnant pillars in mines prior to their abandonment to maximize the resource recovery. Extracting these safety pillars presents a special hazard to mine employees and the stability of the mine. Often subjected to high stresses, the behaviour of these pillars is not always predictable. Special care must be taken to ensure that infrastructure does not collapse prematurely, and that working conditions remain safe. Otherwise the extraction must be halted, and the remaining coal lost. This investigation reports on a preliminary study of the stability of remnant pillars during the final extraction phase of the Hustas mine in Soma, Turkey. During the operating life of the mine, pillars were established to support the main mine drift and a shaft. At the end of the life of the mine, these pillars were to be extracted to maximize the resource recovery. The extraction of safety pillars protecting the mine main drift and shaft were simulated using the laminated displacement discontinuity model LAMODEL. This model represents the overburden as frictionless layers of variable thickness, and has been shown to more accurately predict subsidence than some other models. The model also allows yield zones around entries in coal to be simulated, as well as a more realistic representation of pillar behaviour and of gob materials. The model itself was constructed based on plans of the mine layout and extraction sequence. Rock properties were obtained from physical properties testing results, and the thickness of the overburden laminations was determined by calibrating the initial model runs, varying the layer thickness until subsidence values observed at other parts of the property were obtained. The properties of coal gob elements and their behaviour in yield were based on previous work. The model was run to simulate the extraction of the main drift and shaft pillars in a series of stages. The model results indicate close agreement between modelled subsidence and actual subsidence, which lent a degree of credence to other results obtained. Modelled stresses at the seam and at the level of the main drift some 35 m above closely matched underground experience as the pillars were extracted. Generally speaking, stress abutments conformed to accepted longwall theory, showing a pressure abutment about 10 m in front of the face. Convergence results indicated that the main drift would collapse 10 m to 15 m behind the face, which it duly did. Stresses on the gob (collapsed material) did not exceed the cover stress, also as predicted by conventional longwall theory. The study has shown that properly calibrated models can be useful in assessing mine safety during remnant pillar extraction; the results confirmed previous studies. The modelling work allowed the mine operators to put in place safeguards that allowed the safe extraction of the remnant pillars.

Mots Clés: Safety pillars, Underground recovery, Numerical modelling, LAMODEL