Higher Precision Automated Managed Pressure Drilling Control Achieved Through the Addition of a Thermal Model

For a robust and efficient automated Managed Pressure Drilling (MPD) operation, the choke controller requires an accurate hydraulics model that can be run at minimum computational expense. Integration of a suitable thermal model would improve the accuracy of the hydraulics model used by the choke controller. The use of existing thermal models, however, comes with additional computational costs that are a hurdle when aiming to achieve real-time control at sufficiently fast time-scales. In this paper, a quasi-steady thermal model is integrated with an automated MPD control approach that uses a reduced Drift-Flux Model (RDFM) to describe the hydraulics of multiphase flow in real-time. This integrated modeling approach provides the dynamic temperature profile along a well without increasing the computational expense. The energy equation is solved using the finite-difference method (FDM) in an explicit scheme, with all the temperature-dependent parameters updated in accordance with the calculated temperature profile in each computation step. The RDFM is also reformulated to account for the heat transfer between the gas and the surroundings. This modified model is then incorporated into an automated observer algorithm to estimate parameters, e.g. volume of gas expansion (dependent on temperature), which are used by the controller for choke opening manipulation. Validations are first conducted in a simulation environment for the scenario with a dynamic temperature profile along the well. The results indicate that the proposed modeling approach offers significant improvement compared to approaches which do not consider thermodynamics. A good agreement of the temperature results is observed between the proposed approach and existing models as well as commercial software. Case studies are also conducted for two scenarios to demonstrate the utility of the proposed integrated thermal and hydraulics model. Simulation results indicate that the proposed modeling approach can generate more accurate estimations of unmeasurable variables, which leads to a better performance of the choke manipulation. It should be noted that when employing the modified RDFM with a finite difference scheme, the computational cost is minimized. On a standard laptop computer, the computational time to simulate an entire well is of the order of 70ms for 1s sensor data sampling. Therefore, the proposed thermal and hydraulics model provides an enabling tool for a faster and more precise control of MPD systems.