Rockburst is a dynamic rock mass failure occurring during underground mining under unfavorable stress conditions. The rockburst phenomenon concerns openings in different rocks and is generally correlated with high stress in the rock mass. As a result of rockburst, underground excavations lose their functionality, the infrastructure is damaged, and the working conditions become unsafe. Assessing rockburst hazards in underground excavations becomes particularly important with the increasing mining depth and the mining-induced stresses. Nowadays, rockburst risk prediction is based mainly on various indicators. However, some attempts have been made to apply machine learning algorithms for this purpose. For this article, we employed an extensive range of machine learning algorithms, e.g., an artificial neural network, decision tree, random forest, and gradient boosting, to estimate the rockburst risk in galleries in one of the deep hard coal mines in the Upper Silesian Coal Basin, Poland. With the use of these algorithms, we proposed rockburst risk prediction models. Neural network and decision tree models were most effective in assessing whether a rockburst occurred in an analyzed case, taking into account the average value of the recall parameter. In three randomly selected datasets, the artificial neural network models were able to identify all of the rockbursts.

The paper describes a case study of the safety hydraulic system damage in the working of a longwall in a Polish coal mine. The safety elements are a component of the powered roof supports which secure the shield against damage during rock burst incidents. The damage event, which occurred in the hydraulic system during the mining process, caused the uncontrolled lowering of the powered roof support height during the mining process. The uncontrolled lowering of a shield may cause the danger of the loss of the stability along the longwall working in the form of a rock burst and collapses and may represent a serious and immediate danger to the safety and health of employees. Based on the results of the computational fluid dynamics methods (CFD) analysis of the safety elements in the hydraulic system of longwall 2-leg shield, the causes of damage were diagnosed and presented. The CFD and the strength analysis by the finite element method (FEM) were used for numerical modeling. The diagrams and maps of changes of parameters having an impact on the damage mechanism in safety elements of the hydraulic leg were developed based on the results of model tests. The forecasted values of stress distributions in the safety system of the hydraulic leg have made it possible to identify the reasons of the damage causes, verified by real observations.

Coal mining in many Polish mines in the Upper Silesian Coal Basin is accompanied by high levels of induced seismicity. In mining plants, the methods of shock monitoring are improved, allowing for more accurate localization of the occurring phenomena and determining their seismic energy. Equally important is the development of ways of forecasting seismic hazards that may occur while implementing mine design projects. These methods, depending on the length of time for which the forecasts are made, can be divided into: longterm, medium-term, short-term and so-called alarm. Long-term forecasts are particularly useful for the design of seam exploitations. The paper presents a method of predicting changes in energy expenditure of shock using a properly trained artificial neural network. This method allows to make long-term forecasts at the stage of the mine's exploitation design, thus enabling the mining work plans to be reviewed to minimize the potential for tremors. The information given at the input of the neural network is indicative of the specific energy changes of the elastic deformation occurring in the selected, thick, resistant rock layers (tremor-prone layers). Energy changes, taking place in one or more tremor-prone layers are considered. These indicators describe only the specific energy changes of the elastic deformation accumulating in the rock as a consequence of the mining operation, but does not determine the amount of energy released during the destruction of a given volume of rock. In this process, the potential energy of elastic strain transforms into other, non-measurable energy types, including the seismic energy of recorded tremors. In this way, potential energy changes affect the observed induced seismicity. The parameters used are characterized by increases (declines) of specific energy with separation to occur before the hypothetical destruction of the rock and after it. Additional input information is an index characterizing the rate of tectonic faults. This parameter was not included in previous research by authors. At the output of the artificial neural network, the values of the energy density of the mining tremors [J/m³] are obtained. An example of the predicted change in seismicity induced for a highly threatened region is presented. Relatively good predicted and observed energy expenditure of tremors was obtained. The presented method can complement existing methods (analytical and geophysical) forecasting seismic hazard. This method can be used primarily in those areas where the seismic level is determined by the configuration of the edges and residues in the operating seam, as well as in adjacent seams, and to a lesser extent, the geological structure of the rock The method is local, it means that the artificial neural network prediction can only be performed for the region from which the data have been used for its originated learning. The developed method cannot be used in areas where mining is just beginning and it is not possible to predict the level of seismicity induced in areas where no mining tremors have been recorded so far.