Blast induced ground vibrations generated by explosives in tunnel construction may cause structural damages. Therefore, the aim in blasting must be to suppress vibration effects and mitigate possible hazards on structures. Commonly used criteria for vibration - induced building damage are summarized in DIN 4150. It is generally assumed to be in terms of peak particle velocity. But the psychological character of human response to vibration involves highly subjective attitudes about what kind of environment is allowed even if no structural damage has occurred. The purpose of this article is to evaluate and predict blast induced ground vibrations using an adaptive network fuzzy inference system (ANFIS). Vibrations caused by blasting in tunnel tubes were monitored on an old residential building, which is located almost directly above the tunnel tube. Many researchers found out that the artificial neural network and a neuro-fuzzy technique have better prediction capability than conventional vibration predictors. 48 blast vibrations were recorded. Among these, 41 blast vibrations data sets were used for the training of the ANFIS model, whereas 7 remaining blast vibration data sets were used for the validation and comparison by ANFIS. The model is based on three parameters (maximum charge per delay, frequency and distance between a blast face to a monitoring point) which were used to predict the peak particle velocity.
展开▼
