This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper OTC 31295, “Automated Reservoir Management Work Flows To Identify Candidates and Rank Opportunities for Production Enhancement and Cost Optimization in a Giant Field Offshore Abu Dhabi,” by Carlos Mata, SPE, Luigi Saputelli, SPE, and Dorzhi Badmaev, SPE, ADNOC, et al. The paper has not been peer reviewed. Copyright 2021 Offshore Technology Conference. Reproduced by permission.


An intensive effort is required to identify the right candidates and to ensure the technical and economic success of well interventions, infill-drilling locations, and sidetrack locations. An accelerated method for selection of existing wells for workovers and sidetracks would be of great value. A work flow is developed to automate tasks such as data gathering and validation. The solution allows integrated assessment of production-enhancement opportunities through consistent analytic computations and machine-learning techniques including Bayesian networks and time-series forecasting models.

Automated Data Integration

The advisory system starts with data integration and work-flow automation steps. The system was implemented in the commercial software PetroVisor. The software integrates data points as signals that are tied to entities such as wells. Different relationships with multiple levels can be specified between the integrated entities. The software allows the integration of any legacy data source all along the value chain to topside facility equipment.

The advisory system aggregates data from selected entities on different hierarchy levels and incorporates data from more than 400 wells with 700 perforations and 5 decades of production history.

The selected field contains two reservoirs, each with three sublayers. Production began with primary depletion. Waterflooding, and then crestal gas injection, began in the early 2000s. Reservoir properties and dynamic data were exported from the latest history-matched reservoir simulation model.

The well trajectories and the exported fluid contacts allow the generation of Voronoi grids for each reservoir and layer. The literature refers to Voronoi grids as perpendicular bisection (PEBI) grids. Each gridblock is locally orthogonal. Unstructured PEBI grids are favorable when complex well geometries and geological features are present. Voronoi cells were generated automatically for the layers in the software, where grid-cell boundaries are created and constrained by encountering faults and fluid contacts. Grid nodes represent the wells. An example of a Voronoi cell is shown in Fig. 1.

Voronoi grids should be recomputed over time because of the dependency of actual fluid contacts and active wells. It is recommended to regenerate them whenever a recently history-matched model is available. After generation of grids, exported reservoir and fluid properties can be assigned to each node. The spatial properties can be upscaled locally within each cell, allowing prompt evaluation of original oil in place (OOIP) and remaining oil in place of grid cells.

This content is only available via PDF.
You can access this article if you purchase or spend a download.