the effect of dynamic data adjustments in production system simulation models on oil production forecasting applied to reservoir simulation models

Simulation models of integrated reservoir and production systems are required for a robust production forecast. traditionally, reservoir and production system models are calibrated against dynamic data to establish future boundary conditions. herein, we propose probabilistic data assimilation for production system models to improve the quality of production forecasts. we used a benchmark case through a reference model, which represents the real field, and a simulation model for (1) sensitivity analysis of production system parameters; (2) adjustment of production system parameters, based on dynamic production history data, to minimize the gap between data and model using an optimization method; and (3) comparison of production forecast in the simulation model, coupled to historymatched and nonmatched production systems, and a reference model. sensitivity analysis of production system parameters indicated a significant impact of the pressure gradient adjustment parameter. but we verified that there were no unique correlations (multiphase flow and fluid) and absolute roughness in the production tubing that fit overall production history, affecting production forecast. comparing production curves of simulations, coupled with historymatched and nonmatched production system models to the reference model, we show that adequately adjusted models are closer to the real model. it is mainly the case for systems with higher capacity, where production is more dependent on the responses of the production system. the probabilistic calibration approach of production systems before integrating reservoir models to adjust production systems simulation models is simple to perform. it can improve the quality of the forecast of the field.