The Reliability of Peak Particle Velocity Analysis Methods

This paper considers the current methods of analysis employed, the validity of assumptions made and the reliability of results and predictions derived; where appropriate, alternative methods are proposed which are more accurate and can be applied with greater confidence. In regard to data collection, new local blasting data has been analysed and conclusions drawn but, due to the restricted amount of space available, reference to data set A only has been made in the text. The problem of mathematical model choice is considered and results for two and three parameter power laws are compared. The very important matter of parameter determination is discussed fully, with major emphasis being placed on the comparative merits of log-log linear, log-log planar and direct power law regression. Supporting data analysis examples are included. Assumptions regarding the normality of ‘best-fit’ regression model residual distributions are questioned and the reliability of associated safety limit determinations is discussed.