A Study on Optimal Design for Gas-to-Liquid Process by Utilizing Micro-channel Fischer-Tropsch Reactor

Abstract

For several decades, a gas-to-liquid (GTL) process has been identified as a promising technology for converting abundant natural gas (NG) to clean synthetic fuel. Throughout the GTL process, NG is firstly converted to synthesis gas, mainly composed of hydrogen and carbon monoxide by a syngas reforming process. Syngas is then chemically converted into the liquid fuel by Fischer-Tropsch (FT) reaction process, wherein several carbon atoms in the syngas are oligomerized to form a long chain hydrocarbon product. This hydrocarbon product is also known as FT synfuel, and it contains many hydrocarbon species of carbon number ranging from 1 (methane) to more than 30 (FT wax). FT synfuel has high research octane number (RON) and cetane number (CN) and almost nitrogen and sulfur free. So it is taking a premium position in the fuel market. FT synthesis reaction is known as an strongly exothermic reaction: A large amount of heat, ca. 165kJ per mol of converted CO is generated during the reaction. This heat must be removed to prevent runaway situation and achieve safe isothermal operation of the reactor. Various types of reactors, such as fixed bed, slurry bubble column, circulating fluidized bed, and fixed fluidized bed FT reactors, have been developed and applied to the GTL industries. In order to utilize the FT reaction in off-shore platform, however, more compact, but highly productive, FT reactor is essential. Recently, the concept of micro channel FT reactor has evolved because it can effectively handle large amount of heat based on the high heat exchange surface area per unit volume. It was reported that the heat removal rate in the micro-channel reactor was around 15 times higher than that in a conventional fixed bed reactor. Moreover, micro-channel reactors are good for scale-up perspectives because the wax production can be easily increased by simply adding modularized reactors. It has been reported that the performance of the pilot scale micro-channel reactor was consistent with that of a single micro-channel reactor due to the achievement in isothermal operation. This thesis has addressed the optimal design for GTL processes based on the micro-channel Fischer-Tropsch reactor technology: Modeling of the micro-channel FT reactor, optimization of the GTL process based on the developed reactor model, and dynamic simulation and optimal operating procedures for the micro-channel FT reaction system. The reactor models were validated against the real operation data. A distributed parameter model for micro-channel FT reactor was developed by using a new method, in which all the process and cooling channels are decomposed into a number of unit cells. Each neighboring process and cooling channel unit cells are coupled to set up material and energy balance equations, including heat-transfer equations for the entire reactor domain, which are then solved simultaneously. The model results were compared with the experimental data for a pilot-scale reactor described in the literature, and were found to be in good agreement. Several case studies were performed to see the effect of variables such as catalyst loading ratio, coolant flow rate, and channel layout on design of a cross-current type reactor with state-of-the-art Fischer–Tropsch catalyst. The cell-coupling model was then modified to consider more realistic type of flow configurations and flow distribution effect. Cell domain was re-defined for each flow configuration, and the realistic flow distribution effect was incorporated into the model by using results obtained from computational fluid dynamics (CFD). Several case studies were conducted to see the effect of flow configurations, flow distribution, and catalyst loading zones. It was observed that the geometry of cross-co-cross current was found to give the best performance among the designs considered. Optimal GTL process was suggested by conducting steady-state simulations where the developed micro-channel FT reactor model was implemented in the form of regressed artificial neural network (ANN) model. First, steady state model for a conventional GTL process was developed. Then, an optimization problem was formulated by defining objective function as the net profit. Design variables for this problem were the pressure and temperature of the FT reactor, split ratio for purge, and recycle flowrate to the FT reactor. Nelder-Mead algorithm was used to solve this derivative-free optimization problem. It can be said that by utilizing the reaction heat of the FT reactor, the reboiler duty for the CO2 separation was reduced, and the overall efficiency was increased. Optimal solution showed better economic performances over the base case design. A dynamic model for the FT reactor was developed. A partial differential equation for the 3D cell-coupling model was formulated and solved to obtain time dependent temperature profile in the entire reactor domain. Several case studies were performed to analyze dynamic behavior of the micro-channel reactor. Separate dynamic simulations were also conducted to suggest optimal start-up and shut-down procedure for the FT reactor system. Several scenarios were generated to analyze the thermal and hydrodynamic behvaior of the reactor. Optimal operating strategies for both start-up and shut-down of the reactor could be obtained. This work could contribute to desigining optimal GTL process, especially using a large-scale micro-channel Fischer-Tropsch reactor containing more than 1,000 process channels. The developed reactor model, steady-state model, and dynamic model could be utilized for designing and operation of the GTL system.