Publications

Conventional monitoring and identification of problems in continuous gas lift wells are time-consuming, involving data collection and analysis activities. In this study, machine learning (ML) is developed to monitor gas lift systems based on a digitalized two-pen recorder chart in the form of casing pressure, tubing pressure, the pressure difference between casing pressure and tubing pressure, pressure changes of casing pressure, and pressure changes of tubing pressure and troubles. The training data used for the ML algorithm is 70% of the total data and 30% of the total data is test data. Several ML algorithms such as Gradient Boosted Tree(GBT), Decision Tree(DT), Random Forest(RF), k-nearest Neighbor(KNN), and Support Vector Machine(SVM) are used to predict problems in continuous gas lift wells. After obtaining the ML algorithms with an accuracy of more than 90%, the ML algorithms are tested with 52 data to validate the effectiveness of the algorithms. Performance evaluation of the machine learning algorithms was conducted by assessing the accuracy, precision, recall, and F1-measure values. Performance evaluation results indicate that RF can be used as the least inaccurate machine learning algorithm for predicting well problems in continuous gas lift wells with a 97.9% accuracy rate. By further implementation of this study, effective monitoring and identification of problems in the continuous gas lift wells operations can be obtained.