Prediction of methanol loss in liquid hydrocarbon phase during natural gas hydrate inhibition using rigorous models

• The methanol loss in liquid phase of alkane (saturated) hydrocarbons is determined through employing the machine learning approaches.
• The predictive tools are developed based on 200 data points.
• The models exhibit high capability to accurately predict methanol loss in condensates during gas hydrate inhibition operation.
• A very good agreement is attained between the models outputs and observed data.

Methanol is the most widely used natural gas hydrate inhibitor and it is only effective as a hydrate inhibitor in the aqueous phase. Methanol is not regenerated in natural gas inhibition process due to its intermittent application in most cases. However, a significant cost is associated with the process because of methanol loss while utilizing this inhibitor. In this work, several intelligent models along with a new mathematical correlation are presented in terms of methanol concentration in aqueous phase and temperature to precisely forecast the methanol loss in the saturated hydrocarbons phase. An excellent match was noticed between the calculated results and literature data.