Eventos Anais de eventos
COBEM 2021
26th International Congress of Mechanical Engineering
Reservoir characterization using ES-MDA method combining pressure and temperature data
Submission Author:
Vinicius Mattoso Reis Da Silva , RJ
Co-Authors:
Vinicius Mattoso Reis Da Silva, Danmer Maza, Abelardo Barreto Jr, Marcio CARVALHO
Presenter: Vinicius Mattoso Reis Da Silva
doi://10.26678/ABCM.COBEM2021.COB2021-0255
Abstract
Reservoir characterization is an important tool for reservoir management. Formation test is fundamental because it is the only source of dynamic data during the exploitation step. It typically consists of measuring the pressure and temperature responses at the well, during production, injection, or static conditions. Usually, only pressure data is post-processed to estimate reservoir characteristics and improve the reservoir model. Most analyses of formation tests assume isothermal flow. In the past, an isothermal flow hypothesis was considered and accepted, because the resolution of temperature gauges was insufficient to detect small variation during the flow. However, with the technical evolution of temperature sensors, the quality of the measured data has improved considerably and it is becoming common to have temperature sensors permanently in the wells. Recent studies show that considering isothermal flow and only post-processing pressure data can lead to misinterpretation due to the neglected thermal effect, leading to errors in the reservoir properties estimation and consequently inefficient reservoir management. This problem becomes more critical in high transmissibility reservoirs, such as in the pre-salt reservoirs in Brazil. In this work, a synthetic reservoir was studied using an in-house flow simulator that considers Joule-Thomson heating and cooling, adiabatic fluid expansion/compression, conduction, and convection effects in the thermal energy balance equation to provide temperature and pressure data. After that, the inverse problem was solved using an ensemble-based method to characterize the reservoir using the pressure and temperature data artificially created by introducing noise in the prediction of the direct problem. To estimate the reservoir properties and evaluate the uncertainties into the variables in the analysis, the ensemble smoother with multiple data assimilation (ES-MDA) is applied creating many models to produce a confidence interval to the parameters. Results show that the ES-MDA method applied with the coupled pressure and temperature transient data provides better reservoir characterization and uncertainty quantification comparatively with only pressure data.
Keywords
thermal model, Reservoir Characterization, Ensemble method, Inverse problem, ES-MDA
DOWNLOAD PDF VIEW PRESENTATION
