Statistical methods applied to the analysis of blasting vibrations in a diabase quarry in the city of Campinas, (SP) Brazil

Abstract

Statistical methods of multiple regression analysis, trend surface analysis and principal components analysis were applied to seismographic data recorded during production blasting at a diabase quarry in the urban area of Campinas (SP), Brazil. The purpose of these analyses was to determine the influence of the following variables: distance (D), charge weight per delay (W), and scaled distance (SD) associated with properties of the rock body (orientation, frequency and angle of geological discontinuities; depth of bedrock and thickness of the soil overburden) in the variation of the peak particle velocity (PPV). This approach yielded variables with larger influences (loads) on the variation of ground vibration, as well as behavior and space tendency of this variation. The results showed a better relationship between PPV and D, with D being the most important factor in the attenuation of the ground vibrations. The geological joints and the depth to bedrock have a larger influence than the explosive charges in the variation of the vibration levels, but frequencies appear to be more influenced by the amount of soil overburden.