Deterministic tools to estimate induction time for methane hydrate formation in the presence of Luvicap 55 W solutions

Abstract

The formation of gas hydrates in offshore pipelines is a severe flow assurance issue. The hydrates may form quickly in pipelines without any warning. Thus, effective remediation approaches may require to be performed for days or months. Injecting chemicals such as kinetic hydrate inhibitors (KHIs) can be implemented to prevent or manage gas hydrate formation in offshore facilities. KHIs with polymeric structures and various functional groups postpone hydrate nucleation and growth with a subcooling of 8–10 °C. In this paper, a new approach is introduced to predict the induction time of methane hydrate formation in the presence of Luvicap 55 W solutions (as a KHI) with a broad range of concentrations. In the intelligent models, including least squares support vector machine (LSSVM), adaptive network-based fuzzy inference system (ANFIS), and gene expression programming (GEP), are employed in this study, where 440 experimental data are collected. In the intelligent modeling, 85% of data are utilized for the training step and 15% for the testing step. The smart tools relate the induction time parameter as a target to input parameters such as the molecular weight of solution, mass fraction of KHI, temperature, pressure, and subcooling. The key statistical parameters, including average relative error percent (ARE %), average absolute relative error (AARE %), and coefficient of determination (R2), are calculated to evaluate the performance of the deterministic models. The values of the coefficient of determination (R2) for the developed GEP model are 0.9582 and 0.9726 in the training and testing steps, respectively. The GEP model exhibits the best performance for induction time estimation compared with the LSSVM and ANFIS models, though the run/computational time of this model (e.g., 18 min) is considerably greater than that of other deterministic approaches. Using the ANFIS model, the most influential parameters affecting the induction time are the system pressure and temperature with direct relationship. In addition, other input parameters, including the molecular weight of the solution, mass fraction of the Luvicap 55 W, and subcooling, exhibit indirect relationship with the induction time.