TY - THES

T1 - Development of Integrated Production Model of WIDE Northern Oil Fields

AU - Kronberger, Peter

N1 - embargoed until 02-09-2021

PY - 2016

Y1 - 2016

N2 - The three fields Aldorf, Bockstedt and Dueste are producing since almost 65 years into one processing plant in Barnstorf, where the oil production is measured and – based on irregular field measurements – back-allocated to the individual wells. This allocation was perceived to be an error prone activity, as deviations of 30% were observed in 2014. Therefore, in order to enable cross-checking the data from PDMS and improve the current data situation, the development of an integrated network for the back-allocation of the data began. The integrated model considers not only the irregular liquid rate and water cut measurements, but also important pump characteristic parameters. The mainly used software were MBAL, GAP, PROSPER and RESOLVE from Petroleum Experts as well as an Eclipse reservoir model. At first the whole pipeline network was built in GAP, and PROSPER models were created for each individual well. The reservoir representation for Dueste was provided by an Eclipse model. The Bockstedt reservoir is represented by two MBAL models, because of its complexity. Both had to be properly adjusted and finally connected to the GAP systems. The Aldorf field had no representative MBAL or Eclipse model. Therefore, it was re-analyzed, set up as an MBAL model and finally history matched. For a faster data transfer of results and input parameters, various Microsoft Excel macros with an OpenServer connection were created. Ultimately, all modelled tools were put together in RESOLVE and properly adjusted to each other to achieve consistency. Relative permeability curves, designed with a Brooks Corey model for each production well, were used as tuning parameters to get satisfying fits of the rates. Furthermore, mainly near wellbore permeability changes and skin adjustments were performed to match the injection rates. The fully integrated model showed good results for the back-allocated liquid rates per field after a verification simulation. The achieved deviations between the model and the PDMS by the end of 2015 were 3.0% for Aldorf, 2.7% for Bockstedt and for Dueste 2.4%. Compared to the sum per field of the well testing units, the differences in Aldorf were -16.2%, -9.3% in Bockstedt and -6.5% in Dueste. Furthermore, three short-term forecast scenarios until the end of 2020 were created to show the predictive functionality of the model. The verification runs showed the over-estimating of the field measurements quite clearly. It was found out that on a field and compartment level more precise matches were achieved than on the individual well basis and that the relative permeability curves are very sensitive as the average water cut over all three fields is with 96% already at an extremely high stage. Keeping in mind that two fields are based on multi-tank material balance, the results achieved provide a reasonable cross-checking tool for the PDMS and for creating various prediction scenarios for sensitivity studies.

AB - The three fields Aldorf, Bockstedt and Dueste are producing since almost 65 years into one processing plant in Barnstorf, where the oil production is measured and – based on irregular field measurements – back-allocated to the individual wells. This allocation was perceived to be an error prone activity, as deviations of 30% were observed in 2014. Therefore, in order to enable cross-checking the data from PDMS and improve the current data situation, the development of an integrated network for the back-allocation of the data began. The integrated model considers not only the irregular liquid rate and water cut measurements, but also important pump characteristic parameters. The mainly used software were MBAL, GAP, PROSPER and RESOLVE from Petroleum Experts as well as an Eclipse reservoir model. At first the whole pipeline network was built in GAP, and PROSPER models were created for each individual well. The reservoir representation for Dueste was provided by an Eclipse model. The Bockstedt reservoir is represented by two MBAL models, because of its complexity. Both had to be properly adjusted and finally connected to the GAP systems. The Aldorf field had no representative MBAL or Eclipse model. Therefore, it was re-analyzed, set up as an MBAL model and finally history matched. For a faster data transfer of results and input parameters, various Microsoft Excel macros with an OpenServer connection were created. Ultimately, all modelled tools were put together in RESOLVE and properly adjusted to each other to achieve consistency. Relative permeability curves, designed with a Brooks Corey model for each production well, were used as tuning parameters to get satisfying fits of the rates. Furthermore, mainly near wellbore permeability changes and skin adjustments were performed to match the injection rates. The fully integrated model showed good results for the back-allocated liquid rates per field after a verification simulation. The achieved deviations between the model and the PDMS by the end of 2015 were 3.0% for Aldorf, 2.7% for Bockstedt and for Dueste 2.4%. Compared to the sum per field of the well testing units, the differences in Aldorf were -16.2%, -9.3% in Bockstedt and -6.5% in Dueste. Furthermore, three short-term forecast scenarios until the end of 2020 were created to show the predictive functionality of the model. The verification runs showed the over-estimating of the field measurements quite clearly. It was found out that on a field and compartment level more precise matches were achieved than on the individual well basis and that the relative permeability curves are very sensitive as the average water cut over all three fields is with 96% already at an extremely high stage. Keeping in mind that two fields are based on multi-tank material balance, the results achieved provide a reasonable cross-checking tool for the PDMS and for creating various prediction scenarios for sensitivity studies.

KW - full field model

KW - integrated production model

KW - MBAL

KW - GAP

KW - PROSPER

KW - RESOLVE

KW - OpenServer

KW - integriertes Produktionsmodell

KW - MBAL

KW - GAP

KW - PROSPER

KW - RESOLVE

KW - OpenServer

M3 - Master's Thesis

ER -