Hydrogen Production by Integration of Fluidized Bed Heat Exchanger in Steam Reforming of Natural Gas

Licentiatavhandling, 2019

The possibility to integrate fluidized beds as heat sources for steam reforming in hydrogen production plants is examined in this work. Sites with large-scale production and consumption of hydrogen such as refineries are considered as especially interesting. The proposed processes include the traditional catalytic steam reforming of natural gas (SMR), shift reactors and purification steps, the main difference being the reformer furnace. In the conventional plant a gas-fired furnace is utilized where the heat necessary for the endothermic steam reforming reaction is generated and transferred to the steam reforming tubes by mainly radiation. The alternative which is investigated in this work is to immerse the reformer tubes in fluidized bed heat exchangers. Based on the high heat transfer which can be expected from a bubbling fluidized bed to an immersed tube surface, the potential benefits are significant compared to conventional technology mainly as a result of improved heat transfer to the reformer tube. The proposed processes are compared from both a thermodynamic, economic and environmental point of view with conventional steam methane reforming. The work combines process simulations of hydrogen production plants using Aspen Plus with lab-scale experiments on fluidized bed combustion and fluidized bed heat transfer.

The first proposed process uses oxygen carrier particles in a single fluidized bed heat exchanger where the fuel is converted in the bed, a novel process facilitated by the use of oxygen carrier particles. The process is estimated to reduce the supplementary fuel consumption of natural gas which also reduces the CO2 emissions from the process by approximately 12% and the levelized hydrogen production cost is approximately 7% lower. A lab-scale experiment was performed where methane, the component in the fuel gas which is considered to be the most difficult to convert in the bed, was burnt at moderate furnace temperatures (600-800˚C) with an inert bed material as silica sand and an oxygen carrier, ilmenite. The experiments showed that the methane can be converted in the dense bed even at moderate temperatures and the use of an oxygen carrier increased the fuel conversion in the bed.

The second proposed process uses two interconnected fluidized bed reactors, with an oxygen carrier circulating between them. In one of the reactors the oxygen carrier is oxidized with air and in the other the oxygen carrier is reduced by the fuel. The flue gas stream obtained from the fuel reactor is not diluted with nitrogen and the CO2 produced can easily be captured. The supplementary fuel consumption increases only slightly compared to the first proposed process. The energy penalty to enable CO2 is small and the supplementary fuel consumption is still significantly lower than for the conventional process.

The third proposed process uses biomass instead of natural gas as supplementary fuel, which enables the possibility of achieving net negative emissions from the hydrogen production process. By using a higher steam-to-carbon ratio and a higher temperature at the outlet of the reformer it is possible to increase the hydrogen yield in the process and at the same time achieve significant negative emissions from the process.

One of the key reasons for using fluidized beds as a heat source is that the heat transfer from bed to tube is expected to be high in relation to the conventional furnace. Although much work has been done to investigate bed-to-tube heat transfer most of these experiments have been conducted at bed temperatures lower than 400˚C which is below the operating temperature sought in these processes. Lab-scale experiments are therefore performed to verify if i) high bed-to-tube heat transfer coefficients can be expected in the targeted systems ii) well-known heat transfer correlations determined at lower temperatures also can used to predict the heat transfer at higher bed temperatures. The estimated heat transfer coefficients were high, 768-1858 W/(m2K) at 400-950˚C bed temperature, for the tested bed materials of relevant particle sizes. Three of the heat transfer correlations studied showed good accuracy to predict the bed-to-tube heat transfer coefficient in the experimental unit. The heat transfer correlations could therefore be used to predict the bed-to-tube heat transfer coefficient for the three proposed process configurations. Both experimental campaigns supported the claim that fluidized bed heat exchangers are suitable for use in the proposed SMR application. The proposed processes display high thermal efficiency, a cost competitive levelized production cost and the possibility to achieve significant reductions in CO2 emissions in the industry of hydrogen production.