Knowledge-based and data-driven fuzzy modeling for rockburst prediction

Since rockburst is a violent expulsion of rock in high geostress condition, this causes considerable damages to underground structures, equipments and most importantly presents serious menaces to workers' safety. Rockburst has been associated with thousands of accidents and casualties recently in China. Due to this importance, this research was intended to predict rockburst intensity based on fuzzy inference system (FIS) and adaptive neuro-fuzzy inference systems (ANFIS), and field measurements data. A total of 174 rockburst events were compiled from various published research works. Five different models were investigated. The maximum tangential stress, the uniaxial compressive strength, the uniaxial tensile strength of the surrounding rock and the elastic strain energy index were considered as the inputs while the actual rockburst intensity was the output. In some models, the inputs were extended to the stress coefficient and the rock brittleness coefficient. The results obtained from the study conclude that the knowledge-based F1S model shows lowest performance with 45.8%, 13.2%, 16.5% and 66.52% of the variance account for (VAF), root-mean square error (RMSE), mean absolute percentage error (MAPE) and the percentage of the successful prediction (PSP) indices, while the ANFIS model indicates the best performance with 92%, 1.71%, 0.94% and 95.6% of VAR, RMSE, MAPE and PSP indices, respectively. These results suggest that the developed models in the present study can be used for the rockburst prediction, and this may help to reduce the casualties sourced from the rockbursts. (C) 2013 Elsevier Ltd. All rights reserved.