In this thesis numerical modelling has been used to investigate and describe pseudorandom seismic sources, for the purpose of making these more geophysically competitive with explosive sources. Such research is important, since there is growing public and political pressure to reduce the use of explosive sources (e.g. in environmentally sensitive and populated areas, or due to international security issues.) This thesis focuses on the Vibroseis and Mini-SOSIE surface sources, which are popular in the petroleum and coal seismic industries.
For the vibroseis technique three types of pilot signal have been examined. The common swept-frequency pilot has been taken as a comparative reference. Two pseudo-random alternatives have also been examined namely the full-cycle polarity-reversal and the novel half-cycle polarity-reversal techniques. The half-cycle technique exhibits particular advantages with regard to wavelet shape, harmonic
noise, bandwidth, and wavelet interference. However some implementation issues exist, which will require a modification to the standard vibrator operating procedures.
For Mini SOSIE, software stacking and correlation techniques have been examined. These show advantages over real-time hardware stacking, as they allow prestack processing to attenuate noise. Particular processes demonstrated here include predictive deconvolution (to reduce noise due to non-random impact sequences and, sign-bit and median stacking (to reduce effects of high-amplitude cultural noise).