A critical review of drilling mud rheological models

Abstract

Drilling mud is a mixture of base fluid and other materials combined in specified amounts for the purpose of cleaning a drilled well. Some of its functions during drilling include but are not limited to the following: cuttings transport, cooling and lubrication of drillstring, consolidating wellbore walls and controlling formation pressure. These functions are executed concurrently. In order to improve the drilling process, a robust mud rheological model is needed to capture the dynamics of the mud flow. This work synthesises an appreciable amount of studies on the modelling of oil well drilling mud rheological properties. While a substantial number of models have impacted the literature, to the best of our knowledge, a critical appraisal of these models has not. Furthermore, there is little or no literature review of the knowledge gained from these models to date. A lot can be gleaned by concentrating on the former; principal of which would be relevant for deepening knowledge on the subject. Hence, an attempt has been made to orient the contents of this review towards critiquing the extant models and charting a new course for future studies. This is what makes this study novel and serves as its major contribution. To enable an understanding of the subject, some consideration of theoretical information has been put in place. The general ‘filter’ for inclusion of models in the current review has been that the model should (1) have been published in peer-reviewed journal papers or conference proceedings and (2) the model is specifically for predicting drilling mud rheological properties (plastic and apparent viscosity, yield point and gel strength) and not crude oil or oil well cement rheology. The review highlights what algorithms were used, their relative successes, limitations, performance as well as input parameters which aided in estimating mud rheological properties. For convenience and ease of reference, these models are organised chronologically in simple tables. Results of the bibliometric survey show that though there are numerous modelling approaches in the literature, artificial neural network was the most popular algorithm among researchers. It is also noticed that fuzzy logic theories and a fusion of traditional sensors and machine learning algorithms are recently gaining traction in their use. Furthermore, it is also observed that while yield point, plastic and apparent viscosity are widely modelled, little attention is given to the gel strength. This review constitutes the first critical compilation on a broad range of models applied to estimating mud rheological properties with the aim of supplying vital elements necessary for an improved understanding of the concept of mud rheology. Despite the authors’ effort at being comprehensive, the list of models in this work is far from exhaustive. Nonetheless, the consolation is that it has captured the majority of mud rheological properties models in the literature. The information contained in this review is intended to be of assistance to researchers and the drilling community in the making of technically correct and cost-effective decisions on mud rheology.