Construction and Application of Scenario-Based Data Model for Oilfield Development Plan Extraction

Data is the core element underpinning reservoir prediction and the optimization of development plans in the digital transformation of petroleum exploration and development. However, three critical challenges data silo, underutilized unstructured data value and disconnection between data and application scenario hindering practical application. In this study, a systematic data model constructing approach, including structured data governance - unstructured data governance - data fusion - E&P application model construction is proposed and applied to the development plan. A stepwise structured data governance approach, including E&P application organization, raw data management and data asset catalog construction is first proposed to provide a clear and reliable asset view for the subsequent data model. Then, a combined framework, integrating large model with RAG, is efficient used to increase the ability of unstructured data governance. Above all, an E&P application scenario data model is proposed to increase value association from data to E&P application scenarios. The application of scenario-based data model can be compatible with multi-source heterogeneous data and support dynamic expansion, eliminating the need for manual repeated data adaptation or model reconstruction.