Models for Prediction of Surface Vibrations from Pile Driving Records

Abstract

This study compares high strain dynamic testing measurements taken near the top of a driven pile to peak particle velocities on the ground surface and sound levels detected in the air some distance from the pile during driving. Based on a sample of installation records from 16 piles driven at the Marquette Interchange Project in Milwaukee, Wisconsin, a series of peak particle velocity plots versus distance, energy and scaled distance were created using traditional horizontal distance and rated hammer energy. These plots were modified using the seismic distance, the diesel hammer potential energy from the calculated stroke, and the energy transferred to the pile top. Incorporating these measurements tended to reduce some of the scatter in the data. More importantly, it was also discovered that components of peak particle velocity in the ground can be well correlated to the total pile resistance measured by dynamic testing. A plot of total resistance versus depth often independently yields the same shape curve as a plot of at least one component of peak particle velocity versus depth. A simple mathematical attenuation model is proposed as an initial step toward utilizing this relationship to predict at least one component of ground motions. Measured peak overpressure (noise) in the air correlated less directly to the quantities measured on the pile, but a conservative and simple mathematical model can still be proposed based on the dynamic testing-measured velocity near the pile top and idealized sound generation and attenuation theories.