The seismic method is an important indirect method to investigate the subsurface of the earth. By analyzing how the earth affects the propagation of mechanical waves, the structure of the earth and its seismic properties can be inferred. The seismic vibrator is the most commonly used land source in active-source exploration and monitoring to generate these waves and is the subject of this thesis. The goal of a seismic vibrator is to produce seismic waves with a known signal signature. Commonly sinusoidal signals whose frequency varies over time, called sweeps, are used for this purpose. These signals are typically quite lengthy to compensate for the fact that the instantaneous amplitudes of the vibrator are relatively weak compared to the ones from impulsive sources and the target depths faced. Via the processing step of correlation, the lengthy source signature is collapsed and virtual records are generated as if the vibrator would have released all energy at once. The quality of these virtual records depends on the ability of the vibrator engines to generate the force signature wanted and the ability of the vibrator mechanics and the vibrator-ground interaction to successfully transform the driving force to a seismic wave. In this thesis we investigate the feasibility of driving a vibrator with linear synchronous motors, the influence of drive level on the signals a vibrator generates, the effect of the vibrator-ground coupling, and the possibilities to design more robust source signals.@en

The geometry of the contact between a vibrator and the earth underneath influences the dynamics of the vibrator. Although a vibrator might appear to be well-coupled with the earth on a macroscale, perfect coupling certainly does not occur on the microscale. With the aid of contact mechanical modeling and concepts, it can be shown that this lack of contact at the microscale, or rather the change thereof during a sweep, can have a significant effect on the dynamics of the vibrator-earth system. Modeling of such changing contact predicts that the dynamic behavior varies considerably with the vibrator drive level. The most significant effect predicted by the model is a decrease in the base-plate resonance frequency with an increasing drive level. Extensive drive-level tests carried out in a field experiment confirm this change of resonance behavior with drive level. @en