Production Optimization and Value Maximization of a Giant Ultra-Sour Gas Carbonate Reservoir, Onshore Abu Dhabi

This paper describes how the reservoir team at ADNOC Sour Gas developed the ability to dynamically adjust and manage their production strategy based on plant product output and market requirements, driving profitability and maximizing value of the sour gas assets of the UAE. The reservoir team developed and successfully implemented an extensive data acquisition program, enabling adequate characterization of a giant ultra-sour gas carbonate reservoir in the Late Jurassic Arab Formation in the western area of Abu Dhabi. The field is located in the southern part of UAE, in the Liwa province, and covers an area of 57 km2. It consists of four main reservoir zones: Arab A, Arab B, Arab C, and Arab D. Current development is focused on the central part of the field with most of the wells dedicated to Arab C. Future development plans will focus on the southern and northern areas of the field. Early during the appraisal stage, the data suggested the existence of an areal gradient in composition across the reservoir. As such, a clear understanding of this areal distribution in addition to the usual reservoir gas composition, properties and behavior was essential in optimizing field production and maximizing value. Over the course of field development, reservoir fluids from different well locations were sampled and analyzed. Various issues were encountered during this process including H2S stripping in down hole samples, contamination from stimulation fluids and quality assurance and quality control concerns in lab measurements. Resolving these issues allowed a coherent understanding of the compositional variation in the Arab Formations. To properly model the compositional variation, an innovative methodology was implemented by the team to initialize the dynamic model. The methodology consisted of two major steps. Firstly, PVT data was analyzed and correlations between H2S and other components were developed. Secondly, through PETREL, compositional maps were created. Ultimately, each grid block was assigned a unique composition honoring the areal variation in composition across each reservoir zone. In addition, empirical correlations between fluid components and plant product streams were developed through material balance analysis. Using product models, these correlations were input into the dynamic model which allowed estimated plant products to be output directly from simulation runs. Simulation forecasts of estimated plant products were later verified by actual plant yields, giving confidence in the methodology implemented. Further, this method allowed a quick turnaround in production planning and optimization thereby reducing the reliance on a fully-fledged plant simulator for short term gains and quick wins.