Research Title: Reservoir Parameters Prediction and Characterization Using Integrated Sequence Stratigraphic, Dynamic Modeling, Neural Networks Approach in the Interpretation of Well Tests on Deep Offshore Reservoirs.
Start date: February 2015
Developing a complete characterization of reservoir parameters involved in subsurface multi-phase flow models is a very challenging task. In most instances, these parameters like compressive strength, reservoir pressure, fluid content/contact; porosity profiles, permeability profiles; relative permeability profile with time, irreducible/residual, fluid saturations, cannot be directly measured and if directly measured, are only available at a small number of well locations. These limited data are then combined with geological interpretation to generate a geostatistical model. Reservoir parameters predicted from the different methods like well logs, cores, well testing, do not provide the same result hence increasing the degree of uncertainty thereby making reservoir prediction less accurate.
Based on the aforementioned, there is need to close the gap between especially seismic, well logs, PLT and well test (all types) outputs by possibly updating/upscaling through machine learning and case based reasoning, provide the bench mark to calibrate one against the other and also create a platform for real time reservoir properties prediction from exploration to production. Although the focus of this research is on the Gulf of Guinea Deepwater, the technology should be applicable to global provinces with minimum tuning and effective for reservoir and production management.
Presentation: Facies Modelling Using Neural Networks Technique in Onshore Niger Delta.