Evaluating Key Factors Influencing Scale Formation for Optimized Production in Gas-Lifted Oil Wells

Background and Objective: Scale deposition in oil wells has been recognized to be a major operational problem as well as a major cause of formation damage, both in injection and producing wells. Scales contribute to equipment wear, corrosion, flow restrictions, and undesirable back pressures that cause a drastic decline in oil and gas production. This study presents a data-driven approach to investigate the effect of gas-lift injection rate on the scaling tendencies of oil wells. Materials and Methods: Mathematical models were used to obtain insight into the temperature and pressure profiles of gas-lifted wells as a function of gas injection rates, while the feed-forward neural network was used to develop a classification model to predict the status of the well regarding scale deposition. Univariate histogram/density distribution plots for the different variables to show how each variable influences scale precipitation, while pair plots that show the bivariate relationship between variables in the dataset and their influence on scale formation were also developed. A single-factor ANOVA was performed at a significance level of 0.05. Results: The classification model developed in this study accurately predicted the status of over 98% of the test wells, of which 37 wells were predicted to deposit scales and 68 not to deposit scales. The model only faltered in not adequately predicting two wells of their actual status. The accuracy, precision, sensitivity, and specificity of this model are 0.9813, 0.9855, 0.9855, and 0.9855, respectively. Conclusions: This model can provide benefits in investigating the scaling potential of gas-lifted wells compared to empirical models that apply simplifying assumptions with a high degree of accuracy.