Comparison of Kalman filter-based approaches for permanent downhole gauge pressure estimation in offshore oil production

Abstract

The permanent downhole gauge (PDG) pressure is the most important variable to describe the dynamics of an offshore oil well. Unfortunately, this measurement is often unavailable due to premature sensor failure and the considerable price for its replacement. An alternative to dealing with the lack of the PDG pressure measurement is its estimation. For this, Kalman filters are common tools, with which two distinct state estimators (extended and unscented Kalman filters) were proposed in the literature. These approaches have disadvantages related to the requirement for system linearization and the high computational cost, bringing the necessity for new techniques. In this work, five different Kalman filter-based approaches are tested and compared for the PDG pressure estimation. Through simulated and industrial data, the advantages and disadvantages of each filter are pointed. The results broadly show that the cubature Kalman filter returns the best estimation in the industrial case study, in which the model was properly adjusted. Meanwhile, the extended filters have the best performance in the simulated scenarios, considering that the model was not properly adjusted. In addition, the estimation using a single measurement that is highly correlated to the PDG pressure is enough for its estimation.