Improved Real–Time Estimation of Return Flow Rate of Drilling Fluids by Model Adaptation for Friction Parameter

Early kick and loss detection is a major concern in any drilling operation. A real-time return flow rate measurement would be an added advantage for the early kick and loss detection for conventional drilling systems. In this paper, the real-time estimation of the return flow rate during drilling is studied, using a Venturi channel. A reduced order, 1D, mathematical equation for the shallow water flow/open flow in the Venturi channel is investigated with various non-Newtonian friction models. The volumetric fluid flow rate is estimated online using an unscented Kalman filter. In order to improve the fluid flow rate estimation, the model is adapted to some of the rheological parameters in real time. A simple and effective fit-for-purpose friction model is developed and compared with the other friction models from literature. The fit-for-purpose model has the least estimation error (RMSE 1.465%) and requires less information as inputs. The proposed method exhibits promising potential to be used as an online return flow rate measurement for early kick and loss detection in conventional drilling.