Title: Optimising mud management: adaptive moment estimation-based ANN for predicting rheological and filtration properties of KCl-PHPA-Polyol drilling fluids

Authors: Raunak Gupta; Uttam K. Bhui

Addresses: Drilling Services, Ankleshwar Asset, ONGC, Ankleshwar-393010, Bharuch, Gujarat, India; Department of Petroleum Engineering, School of Energy Technology, Pandit Deendayal Energy University, Raisan, Gandhinagar, Gujarat 382426, India ' Department of Petroleum Engineering, School of Energy Technology, Pandit Deendayal Energy University, Raisan, Gandhinagar, Gujarat 382426, India

Abstract: Field results highlight the importance of real-time monitoring of drilling-fluid properties to prevent operational issues by detecting changes in fluid rheology and filtration behaviour. Traditional laboratory methods, while precise, are time-consuming, error-prone, and do not reflect the rapidly changing conditions encountered during drilling. Thus, there is a pressing need for real-time predictive models that can adapt to on-site data for immediate adjustments. This study introduces a novel method using a deep neural network to predict the rheological and filtration properties of KCl-PHPA-polyol mud, focusing on historically underemphasised parameters like pH, alongside density and Marsh funnel viscosity (MFV). Utilising 2,000 field data points, the artificial neural network (ANN) model demonstrated robust performance, achieving R² values between 0.738 and 0.910, and MAPE from 4.54% to 9.87%. This model significantly advances traditional methods by enhancing the interpretability and utility of ANN, improving operational efficiency through accurate prediction of key rheological and filtration attributes. [Received: June 27, 2024; Accepted: July 31, 2024]

Keywords: adaptive moment estimation; artificial neural network; ANN; artificial neural network; rheological and filtration; KCL-PHPA-Polyol; drilling fluids.

DOI: 10.1504/IJOGCT.2025.148040

International Journal of Oil, Gas and Coal Technology, 2025 Vol.38 No.2, pp.199 - 221

Received: 10 Jun 2024
Accepted: 31 Jul 2024
Published online: 15 Aug 2025 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article