Innovative Data Processing QA/QC for the World's Largest 3D Seismic Survey

Abstract

Abstract With close to 90,000 km2 of onshore and offshore seismic data being acquired and processed in less than 5 years, ADNOC has introduced new innovative QCs to ensure the huge amount of data is properly vetted at each step of the processing sequence. The pre-stack time and depth processing flows are very demanding and amplitudes must be preserved throughout. We present four key diagnostics for QA/QC methods at critical steps in the processing sequence. These selected QA/QC are meant to achieve a resilient seismic data processing outcome amenable to all types of seismic interpretation, either conventional seismic interpretation or advanced seismic reservoir characterization or seismic quantitative interpretation. Seismic bandwidth is the lynchpin of that QA/QC strategy and it will be kept in the spot light to make sure that no processing artifacts erode the acquired seismic bandwidth and that we ultimately preserve the original six octaves of the recorded seismic data. Additionally utilizing the full seismic waveform, either signal or noise, such as surface waves, ground roll, or mudroll, can help compute long wavelength static solutions. The near surface velocity model is a prerequisite for accurate depth imaging not only for the shallow overburden but also at target reservoir levels. We also explore unsupervised machine learning for the automatic detection of amplitude anomalies in ocean bottom nodes (OBN) and to estimate the velocity of the ground roll or mud roll surface wave noise given the huge amount of data we are dealing with.