Prediction of Lost Circulation Using Artificial Neural Networks in the Al Jeribe Formation of Omar Field

Lost circulation is a significant challenge in drilling operations, leading to fluid losses, increased non-productive time, and well instability. This paper develops a predictive model to quantify lost circulation in the Al Jeribe Formation in the Al-Omar field, one of the largest oilfields in the Middle East. Lost circulation is especially prevalent when drilling through the Al Jeribe formation due to the presence of vugs and caves. However, current models for predicting lost circulation often suffer from limited accuracy and efficiency due to the complexity of geological formations and the variability of drilling conditions, leading to unreliable predictions in challenging environments. This research aims to overcome these limitations by developing a more accurate and efficient predictive model tailored to the Al Jeribe Formation, providing valuable insights to mitigate fluid loss and improve drilling efficiency. This paper introduces a novel predictive model for lost circulation in the Al Jeribe Formation, utilizing artificial neural networks (ANNs) trained on extensive field data from over 100 wells. The model incorporates key drilling parameters such as mud weight (MW), yield point (Yp), equivalent circulation density (ECD), rate of penetration (ROP), revolutions per minute (RPM), strokes per minute (SPM), Plastic viscosity (PV), and weight on bit (WOB) as input parameters. The ANN achieved excellent predictive performance, with Training R2 = 0.99 and Testing R2 = 0.99. Error metrics also confirmed strong generalization, with RMSE = 1.70% (training) and 1.40% (testing), and AAPE = 11.0% (training) and 10.2% (testing). In addition, the model identified the most critical parameters influencing lost circulation and provided optimized parameter ranges to mitigate fluid loss during drilling operations. This study focuses on lost circulation prediction in the Al Jeribe Formation, identifying key drilling parameters and providing optimized ranges to reduce losses and improve wellbore stability. It offers insights not covered in previous research, specifically targeting the Al Jeribe Formation. The model predicts lost circulation and suggests practical adjustments to drilling parameter values. The findings are expected to enhance drilling efficiency and minimize downtime in the Al-Omar field. This methodology can also be applied to similar geological formations worldwide to reduce lost circulation in oil fields.