Production optimization by exploiting new technologies for look-ahead geosteering and completion design while-drilling

In this project we want to develop a new methodology for updating the reservoir model while drilling to optimize both the well placement and the completion design. By use of look-ahead deep reading resistivity measurements right-time updating of the geological model will be performed by exploiting a novel joint inversion of the seismic data. The new reservoir geometry and saturation distribution from this inversion is subsequently used to update the reservoir flow model and to modify the well trajectory while drilling. Before pulling out of the well the completion design will be updated based on the updated reservoir model. The tight time-constraints imposed by the drilling and completion operation is an important driver for the development of new optimization methods in this project. In 2019 the activity will center around integrating the optimization framework from NTNU with the model update methodology from SLB. In 2020 we explore the interplay between model updates and the optimization methodology. The goal here is to create a more efficient optimization process by leveraging the information contained in the model update. Up to this point in the project we have evaluated these two steps separately and the integration has been done on the software level. The main activity in 2021 has been: 1. Writing of two papers on the completion optimization. 2. Further development of the prediction methods for reservoir property updating and improvement of this using statistical prediction and machine learning. A prototype is planed used in a shadow operation in early 2022. development of the prediction methods for reservoir property updating and improvement of this using statistical prediction and machine learning will proceed to commercial engineering.

Prosjektet har ført til følgende virkninger og effekter for prosjektpartnerne: For Schlumberger har prosjektet ført til utviklingen av et nytt og forbedret produkt for modelloppdatering basert på teknologi utviklet i prosjektet. Kommersiell utvikling av produktet med intern finansierng er vedtatt startet i 2022. For NTNU har prosjektet ført til oppdatering av optimaliseringsrammeverket FieldOpt slik at det nå og kan optimalisere brønnkomplettering. Dette arbeidet er og publisert i fagfellevurderte publikasjoner. 3 masterstudenter har og skrevet oppgaven innenfor rammen av prosjektet (2 fra NTNU og 1 fra Politechnico di Torino). Prosjektet har og bygget kompetanse innenfor området optimalisering og forskerer fra NTNU som har deltatt på prosjektet har gått videre til stillinger i industrien (Kongsberg Maritim, Vysus Group and Sintef).

In this proposal we will improve well placement and well completion design by exploiting geological look-ahead from deep reading resistivity (DDR) and 3D seismic measurements in a novel optimization framework. The DDR measurements will be used to perform a real time seismic re-characterization of the reservoir model using automated interpretation tools. The innovative step is to include in the real time well optimization methodology not only well placement, but also well completion design and to perform the optimization in a real time operating environment. The joint inversion of DDR and 3D seismic measurements in real-time will provide an updated well-calibrated 3D model of the geology landscape ahead of the drill-bit and thereby enable the operator to geosteer the well optimally in the updated landscape with respect to fluid - and lithology - boundaries in this landscape . Hence, an updated local reservoir model - that is honoring the new LWD measurements - is established in a cylinder around the well ? bore limited by the penetration depth of the DDR measurements, which is typically 200 feet. There are critical time-constraints on updating this reservoir model so that the updated model may take advantage to modify the well ? trajectory. The updated geological model will typically also have implications for the associated reservoir flow model that has been used to design the optimal completion string in this production well. Hence, the completion needs to be redesigned. In the following we will demonstrate that it is crucial for the accumulated production from typical wells to update the completion design honoring the new LWD information. However this redesign has to be carried out under severe time-constraints, since the updated completion-string needs to be in place downhole immediately after pulling out the drill-string.