The PY3 oil field lies in the Cauvery Basin situated some 10km off the east coast of India in the region of Pondicherry. Historically it has been difficult to define an accurate depth model, which has lead to problems with well prognoses and reserve estimations. The difficulties in depth conversion are the result of rapidly varying overburden stratigraphy due to the field being located beneath a series of stacked shelf edge sequences. This phenomenon causes rapid and unpredictable lateral velocity changes and leads to the situation where time lows convert to depth highs. Numerous depth conversion techniques have been applied ranging from traditional V0, k layer cake interval methods to stacking velocity based procedures. These methods can generate significant depth differences particularly those incorporating stacking velocity information.
As a consequence Pre-Stack Depth Migration was carried out on a number of trial lines in order to derive accurate modeled velocities which were then compared with the original stacking velocity field.
The methodology applied consists of generating initial velocity models using travel time tomography. Travel times are picked pre-stack and are compared with modeled travel times. A least squares inversion technique is used to minimize the error between the picked and modeled travel times by adjusting velocities and depths. This technique is especially suitable for detecting velocity variations that have higher spatial frequencies than the cable length.
After the initial interval velocity depth model is generated using tomography the data is depth migrated. The depth gathers are further analyzed for model completeness and accuracy.
This type of depth model yields a better seismic image but is still prone to depth errors caused by velocity anisotropy. The improved seismic profiles enabled greater definition of the key Tertiary section which gives rise to the rapid lateral velocity variations and also of the reservoir interval itself. The application of the velocity data derived from this process has resulted in a significant reduction in the error margin relating to both the depth model and reserve estimation for the field. The results of this study will also have a considerable influence on the siting of future producer and injector wells.