Mitigating Stuck Pipe by Predicting Filter Cake Thickness Using Artificail Intellgence

Abstract The primary objective of this paper is to mitigate stuck pipe incidents by accurately predicting filter cake thickness using advanced Artificial Intelligence (AI) techniques. A thin, impermeable filter cake is crucial for minimizing formation damage and preventing costly operational delays caused by stuck pipe issues. This study explores the application of AI models to estimate filter cake thickness using readily available drilling fluid properties. This work employs three AI-based techniques—Artificial Neural Networks (ANN), Fuzzy Logic (FL), and Functional Networks (FN)—to predict the filter cake thickness. Data for model training and testing were obtained from experiments conducted using real-field oil-based mud samples with CaCO3 as a weighting agent. Input parameters include apparent viscosity, plastic viscosity, yield point, and drilling fluid density. The models were evaluated using different data splits for training and testing, and their accuracy was assessed using the correlation coefficient (R2) and the arctangent absolute percentage error (AAPE). The results demonstrated strong prediction capabilities across all AI models, with R2 values exceeding 0.84 and AAPE below 5%. Among the models, the Functional Network (FN) outperformed others, achieving an R2 of 0.91 for training and 0.92 for testing, with corresponding AAPE values of 2.8% and 3.9%, respectively. These findings indicate that AI-driven models can provide reliable estimates of filter cake thickness, enabling proactive measures to mitigate stuck pipe incidents. This approach offers significant cost-saving potential by preventing unplanned downtime and optimizing drilling operations. This paper presents a novel application of AI techniques, specifically Functional Networks, for predicting filter cake thickness in real-field scenarios. By utilizing common drilling fluid properties, this study provides a cost-effective, data-driven approach to mitigating stuck pipe risks, offering new insights and practical solutions for the petroleum industry.