Production Forecast based on Reservoir Geological Model Validation

Vera Magdolna Schultz

Research output: ThesisDoctoral Thesis

47 Downloads (Pure)


It is a common petroleum industrial practice to build geo-cellular models of the hydrocarbon storing geological structures, which describe their physical properties in space and characterize their original hydrocarbon content. Based on the generated static models, and the known historical production data from the wells, the numerical model of the phase movements and reservoir pressures over the entire operation history is created. The goal of this is to localize the actual (today¿s) oil and gas content of the reservoir and to perform production forecasts for planned future operating scenarios. The lack of information about the reservoir in combination with the conventional numerical simulation methods does not make it possible to accurately describe the past behavior of the wells, therefore the characterization of the remaining reserves and forecasts are not reliable either. A reservoir model, which cannot provide this essential information is useless. As part of the model-building procedure, static properties of the geocellular model are modified and its effect on the modeled dynamic response of the wells is observed. It is often applied to reduce the discrepancy between the measured and modeled well production rates. Operative decisions today are often made based on ¿history-matched¿ (HM) models. This dissertation refers to studies proving that the trial-and-error method cannot achieve a perfect three-phase match of all wells, neither can assure the model¿s resemblance to the real reservoir, or increase the reliability of the forecast. With these limitations, the history matching is applicable to perform sensitivity runs, but it cannot be regarded as a solid basis of decision making. This dissertation presents an alternative, complete reservoir model building approach called the geological ¿model validation¿ (MV approach). It states, that the real system¿s dynamic responses must be over the entire life cycle correctly reproduced. To achieve this, the wells must be given access to the nearest reservoir zones with movable phase content: which is the fundamental difference from the classical HM concept. That where the model wells produce from (perforations, near well area, or from further) show the local and global quality of the static model, indicate any necessary revision and assists optional local model refinements. The method is able to overcome geological modeling limitations of mainly large complex heterogeneous structures, which would normally hinder proper conclusions on the remaining hydrocarbon content and evaluation of future development scenarios. As part of the Ph.D. project, the difference between the history matching attempts and the here introduced concept¿s results are illustrated on the largest Hungarian hydrocarbon occurrence, the Algy¿-2 reservoir. The author analyzes the history matching results of the field operator MOL. The proposed model validation workflow is presented on sectors. Its reservoir-scale implementation is demonstrated over a 20-year long production period. In contrast with the HM approach, it was possible to detect the remaining amount of (inaccessible) movable oil. In this work, the different possibilities of how to get economically viable access to them could not be investigated, because the business strategy of the operator MOL is not known. The Model Validation results already clearly show that the proposed approach is the correct way to go forward with the planned revitalization of the Algy¿-2, and probably it will be applicable to the other Upper-Pannonian oil-bearing reservoirs of the Carpathian Basin too.
Translated title of the contributionProduktionsprognose basierend auf Reservoir Geologische Modellvalidierung
Original languageEnglish
Awarding Institution
  • Montanuniversität
  • Spath, Jeffrey, Co-Supervisor (external), External person
  • Hofstätter, Herbert, Assessor A (internal)
  • Amro, Moh'd M., Assessor B (external), External person
  • Heinemann, Zoltan, Supervisor (internal)
Publication statusPublished - 2022

Bibliographical note

no embargo


  • reservoir modeling
  • geological model validation
  • well model
  • digital twin

Cite this