MODELING OF PRIMARY REFORMER TUBE METAL TEMPERATURE (TMT) USING LS-SVM

This report discusses the research done on the chosen topic, which is Modeling of Primary Reformer Tube Metal Temperature (TMT) using LS-SVM. The objective of the project is to develop a modelthat can predict the temperature of the reformer tubes. The scope of the study focused on the modeling of the primary reformer TMT of PETRONAS Ammonia hydrocarbons such as natural gas into its constituents which are carbon dioxide, carbon monoxide and hydrogen. Pressurized feed (300barg) of hydrocarbon and steam is fed into the reformer tubes and heated by the burners at about 800-1000°C to facilitate the hydrocarbon conversion. The temperature of the tubes is an important parameter to determine the life-time of the tubes. Operating the reformer beyond the TMT design limits can cause premature failures on the tubes which lead to production losses and higher downtime. Based on the literature survey, it shows that the mathematical modeling and simulation approaches are used to determine the behavior of the reformer tubes. For this project, empirical model developed by integrating the process variable will be used to predict the reformer tubes temperature. Empirical model is developed based on real-time data obtained from PASB plant. LS-SVM is used in developing the model and Back Propagation Neural Network is used to develop a model that serves as the benchmark for this project.