Prediction and optimization of milling parameters during Work-over operation in Hassi Messouad region using ML

Abstract

The petroleum industry depends on effective well intervention techniques to sustain and boost the productivity of oil and gas wells, this work explores the optimization of milling parameters using machine learning. It examines well intervention and fishing operations, detailing the parameters and tools used in milling. Using MATLAB, calculations and simulations were performed for wells MD 290 and OMG 612, focusing on the recording of the Rate of Penetration (ROP) and operational parameters. The study applies machine learning techniques to predict and optimize milling parameters for the 7" liner milling of 62 wells in the Hassi Messaoud region. The optimized parameters achieved are a flow rate (Q) of 1219.91500[l/m], a rotational speed (RPM) of 95.64057[rpm], a weight on bit (WOB) of 7.57244[tones], and a yield (Y) of 46.09066, leading to an improved Rate of Penetration (ROP) and reduced costs of milling operations.