Abstract In mature petroleum provinces, such as Abu Dhabi, reserves addition is commonly accomplished by addressing deeper, more difficult reservoirs that have been less intensively explored than the main producing units. ADNOC has unlocked significant reserves by reassessing the Lower Jurassic Izhara and Hamlah formations - long known but underexplored. Using advanced Machine Learning and high-quality data acquisition, the offshore team successfully calibrated these reservoirs, setting a benchmark for the field. This paper describes how these reservoirs were reassessed and tested, unlocking huge additional reserves in accordance with ADNOC aspiration to increase hydrocarbon production. A development well was deepened to appraise Izhara-Hamlah formations to confirm the hydrocarbon-potential. Good quality logs were acquired, and a core was cut in the Izhara-Hamlah reservoir. Moreover, formation pressures were acquired using a formation testing logging tool alongside fluid mobilities and samples. The integration of all these data has led to better characterisation of the reservoirs in terms of fluid and petrophysical characteristics, accordingly a testing program was set and successfully completed. The newly acquired logs were used as a reference to calibrate existing logs in other wells in the field. A deep machine learning algorithm was applied to predict absent data using a high-resolution self-organizing map (SOM) with the new well set as a learning input. Multiple input curve combinations were reiterated to achieve the best results and new curves were generated. The characterisation of the Izhara and Hamlah formations reveals that the Izhara A1 reservoir has a thickness of 110 ft with porosity ranging from 7-18%, while the Hamlah has a thickness of 75 ft with porosities ranging from 5-10%. Both intervals were perforated and stimulated by standard matrix acid without fracture stimulation and flowed light, sweet oil with minor H2S at commercial rates. For volumetric calculations, a geological model was built using contacts defined from the combination of fluid samples, GWD and well test data. The newly acquired and calibrated logs were used to generate a porosities grid and a saturation height function from analogue reservoir was used to generate a water saturation grid. Uncertainty analyses were undertaken to define the volumetric ranges and the most uncertain parameters of the Izhara and Hamlah formations. This discovery from deep reservoirs in ADNOC offshore is a play opener in the UAE. By completing this operation, the full potential of Izhara-Hamlah in one offshore field has been unlocked. Therefore, similar operations can be planned in nearby fields and AI technologies and machine learning will help to unlock further opportunities. This marks a significant milestone for the UAE energy sector, leveraging on advanced drilling technologies and machine learning to unlock significant unexploited resources. This paper describes how these reservoirs were reassessed and tested, unlocking huge additional reserves in accordance with ADNOC aspiration to increase hydrocarbon production.
Al-Kaabi et al. (Mon,) studied this question.