Statistical Characterization of Performance of Biopolymer Drill-In Fluid for Different Rheological Models

Abstract

Appropriate selection of rheological models is important for hydraulic calculations of pressure loss prediction and hole cleaning efficiency of drilling fluids. Power law, Bingham-Plastic Herschel-Bulkley models are the conventional fluid models used in the oilfield. However, there are other models that have been proposed in literature which are under / or not utilized in the petroleum industry. The primary objective of this study is to recommend a rheological model that best-fits the rheological behaviour of xanthan gum based biopolymer drill-in fluids for hydraulic evaluations. Ten rheological models were evaluated in this study. These rheological models have been posed deterministically. Obviously this is unrealistic so these deterministic models are replaced by statistical models by adding an error (disturbance) term and making suitable assumptions about them. Rheological model parameters were estimated by least-square regression method. Models like Sisko and modified Sisko which are not conventional models in oil industry gave a good fit. Modified Sisko model which is a four parameter rheological model was selected as the best-fit model since it produced the least residual mean square. There is 95% certainty that the true best-fit curve lies within the confidence band of this function of interest.