Machine Learning-Based Regression Model for Detection of Petroleum Engineering Problems

The digitalization of the oil field is happening quickly in the oil and gas sector. The assets known as unconventional petroleum are plentiful and dispersed; they include tight oil and gas, coal bed methane and oil, heavy oil, and oil sand. Drilling presents several risks, expenses, and operational issues in the petroleum production business because of people and quantitative mistakes, among other types of faults. A more effective method is needed to overcome the obstacles, get insights, and make wise judgments quickly. Therefore, this paper proposes a machine learning-based regression model for the detection of petroleum engineering problems and is implemented by Python Jupiter software. The drilling data and volcano field data are collected in this work. In these datasets, the data preparation process is used to handle the missing values and for data validation purposes. Following that, data normalization is employed to improve integrity. Next, the data splitting process handles the data set and divides it for the purpose of the regression process. Machine learning (ML) algorithms based on Optimize Gradient Boosting Regression (RBR) are used to solve petroleum engineering challenges. Finally, on the outcome side, the R2 score and Root Mean Square Error (RMSE) values are achieved.