Integrated Reservoir Description From Seismic, Well Log, to Production Data

Abstract

Abstract The increasing need to improve production from developed oil and gas fields requires more detailed and more accurate reservoir description. Our answer is a multidisciplinary approach integrating geophysics, geology, geostatistics, and engineering. For a field with seismic data, well logs, and production data, an integrated reservoir model can be more accurate than the conventional approach of passing the model from geophysicists to geologists then to engineers. In the development stage of a field, a geological model would have been built. The model would have been tested against production data. Now If we are going to build a more detailed model, the existing reservoir model is a valuable input. When dealing with seismic data, time permitting, known information of structure and stratigraphy may be used. When converting time data to depth, well data may be used together with seismic information to build a 3D velocity model. Geostatistics may be used with velocity logs and seismic data to build a more accurate and detailed velocity model. Without a reasonable velocity model for time depth conversion, it is difficult to integrate well data and seismic data. A lithology model or porosity model may be built with well data, structure information, and seismic attribute as soft data. Seismic attributes are calibrated to well data. Geostatistics provides a way to properly weight between well data and seismic data. The reservoir model may be compared or calibrated with production data. Then we may go back and modify the velocity model or horizon picks, or we may modify the conversion from seismic attribute to soft data used in the geostatistical gridding. By iteration, our reservoir model will become more accurate because it honors geophysical, geological, and engineering data. Common data links and tools like geostatistics make the integration easier. Field applications show that an integrated iterative approach to detailed reservoir modeling can provide a better model and help optimize production.