Abstract: Determination of the blasting attenuation characteristics is of most importance for blasting design. The linear and nonlinear regression analysis methods to evaluate the parameters in the S formula and revised S formula are investigated. Based on the investigation, the concept for the residual sum of squares and nonlinear residual sum of squares (standard residual sum of squares) are proposed. The algorithm of nonlinear residual sum of squares is also introduced in detail. In accordance with the engineering case, the standard residual sum of squares is used to evaluate the deviation of vibration velocity acquired by the site monitoring and the regression analysis results by S and revised formula. It is shown that the accuracy of the nonlinear regression analysis method is significantly better than the linear regression analysis method. It is also reported the residual sum of squares and the standard residual sum of squares obtain by S formulas and revised S formula are almost similar; for this case, the two blasting attenuation formulas are available.

Key words: blasting, attenuation parameters, nonlinear regression analysis, residual sum of squares