Frédéric Billette

The key practical aspects of 3D tomography: Data picking and model representation

There are a wide variety of approaches to building velocity models with prestack depth migration. The most common pitfalls we see in them are extracting reliable yet detailed information about velocity errors from migrated data and then the inability to produce geologically reasonable velocity models from these error estimates. We extract detailed error estimates along horizons from finely sampled migrated data, edit them statistically, and QC the edits in map view. Tomography converts these error estimates into new velocity models or velocity model updates using a flexible model building system.

Taking advantage of dual‐azimuth analysis for model building and imaging over Mad Dog

We describe the advantages of using dual-azimuth streamer data for complex model building and imaging in the Gulf of Mexico Deep Water. We iteratively image two datasets separately while building a unique common velocity model. Tomography and salt interpretation can take advantage of two illumination patterns that complement each other while each migration is performed along the native shooting direction. The results show improvements compared to the monoazimuth exploration vintage image and provides a step change needed for the field development.

Pushing the limits of resolution at Holstein: A case history from the deepwater Gulf of Mexico

The high costs associated with production in deepwater have led to increased business support for acquisition and processing of high-quality seismic for field development. The relatively thin, stacked reservoirs discovered at Holstein Field, which contain significant recoverable hydrocarbon volumes (350 million boe), posed a challenge for development planning.

July/Ramadan: a 3D imaging project of 6 surveys

Summary We present the July/Ramadan case study located in the Central Gulf of Suez area. For this application, we unified 6 different surveys in order to improve the data coverage around individual acquisition boundaries. A phase matching filter, initially designed for 4D purposes, was estimated by comparing each survey to a reference one. Applying this filter allowed us to continuously image the whole area. The geological regime is quite complex and previous time or post-stack processing did not provide sufficiently accurate images. Multiple reflections is a major problem that was only partially addressed here. The velocity macro model was estimated using migration-based tomography and salt flooding. The non-regular geometry of the unified dataset did not allow us to take advantage of common-azimuth types of migrations. Thus, we used a general Kirchhoff algorithm. Compared to conventional processing, the 3D prestack imaging sequence was able to provide, as expected, great improvements.