Abstract
Effectively preventing pipe sticking in oil and gas wells is a key aspect of safety and productivity in the oil and gas industry. This paper presents a new torque and drag model developed using machine learning techniques to improve accuracy and predictive capabilities. The model is based on a comprehensive analysis of many factors, including geological characteristics of the well, drilling parameters, fluid parameters, hydraulic conditions, and production equipment parameters. This research explores the application of a chained regression model combining Multilayer Perceptron (MLP) and XGBoost for predicting multiple physical properties in the oil and gas industry. The study aims to enhance prediction accuracy in hydraulics, torque, and drag by leveraging the strengths of both models. The model will help improve the monitoring and interpretation of drilling data streams. Besides the model will help detect signs of other interfering problems in the downhole wellbore conditions and take the necessary measures in advance to prevent them. As a result, it will help optimize costs, minimize non-productive time, and improve safety.