Open Collections

Internally circulating fluidized bed membrane reactor for high-purity hydrogen production

University of British Columbia

A novel reactor configuration, the internally circulating fluidized bed membrane reactor (ICFBMR), was studied in an experimental program for the steam reforming (SMR) of natural gas to produce hydrogen. This work builds on previous fluidized bed membrane reactor (FBMR) research, in which H[sub 2]-selective membranes were located within a fluidized bed of catalyst to produce a H[sub 2] streamdirectly from the reactor, thereby shifting the chemical equilibrium of the reforming reaction forward. The ICFBMR advances this concept by modifying the reactor geometry in order to induce circulation of catalyst solids up a central core draft assembly, which house vertical planar H[sub 2] membranes, and down an outer annular region. The catalyst solids circulation has a number of benefits, especially when the reactor is applied to autothermal reforming (ATR), where the endothermic reforming heat is supplied by direct addition of air to the reactor. In this case, the circulating solids transfer heat from the upper oxidation zone to the core reforming zone, with very little circulation of the nitrogen entering with the oxidation air. The hydrodynamics of the ICFBMR geometry were studied using a Plexiglas cold model. Dimensionless variables were used in an attempt to match key scaling parameters between the cold model, which used air and fluidized catalytic cracking (FCC) solids, and the hot reformer. Solids circulation was studied as a function of the main and annular gas feed rates for three different membrane panel geometries. It was found that solid membrane panels, which prevented communication between the core flow slots, led to maldistribution of solids and gas. Helium tracer studies confirmed that only a small portion (~10%) of the N[sub 2] in the oxidation air fed to the upper reactor transferred to the reactor core with the returning solids. Solids circulation was found to increase linearly with the main feed rate up to a core superficial gas velocity of ~0.3 m/s, and tended to level off after a superficial gas velocity of ~0.5 m/s. The experimental data were used to find predictive equations for solids circulation that could be used for the hot reformer design. Double-sided planar H[sub 2] membranes (each 83 x 280 x 6 mm) were prepared using 50-μm thick palladium alloy foil using techniques of Membrane Reactor Technologies Ltd. Six membranes were installed in a pilot reactor (diameter 0.135 m, height 2.3 m) and a number of pilot reforming experiments were performed. The reformer was successfully operated up to 650°C and 1,500 kPa with a feed of natural gas and steam, under both SMR (external heat) and ATR (direct air addition). Helium tracer studies were performed on the hot reformer, and internal solids circulation was measured to be 0.21 kg/s at a typical operating condition, closely matching the value predicted from cold model experimentation. Pure H[sub 2] (>99.999%, excluding N[sub 2] ) was produced for the first ~180 hours of testing, after which the H[sub 2] purity from two of the six membranes dropped to ~99.7% for the remaining ~150 hours of hot operation. The highest hydrogen production from the pilot reactor was 1.06 Nm³/h. The highest measured ratio of permeate H[sub 2] to feed natural gas was 1.17 Nm³/Nm³, well below the value required for economic operation (~2.5), highlighting how the reactor performance was limited by the installed membrane area. ATR operation showed that permeate H[sub 2] production is only marginally affected by the rate of air addition. Two types of catalyst powders, a SMR (NiO) catalyst and a novel ATR catalyst, were used in the pilot reformer. Low catalyst activity affected a number of the experimental runs. The ICFBMR reactor was simulated using a commercial process simulator (HYSYS) to study the influence of a number of variables on a reactor producing 30 Nm³/h of H[sub 2]. The simulation ignored reaction kinetics, a reasonable assumption for this reactor configuration as reactor performance is overwhelmingly controlled by membrane performance and reactor geometry, with reactor gases near equilibrium. Simulations indicate that the predicted solids circulation rate is sufficient to maintain the core temperature drop to below 30°C, and that there would be limited reduction in membrane area if the circulation rate were to be increased. The reactor model was incorporated within a simulation for the complete system, leading to a predicted overall energy efficiency of 69%, based on utility consumption and the higher heating value of reactants. An economic evaluation of the ICFBMR system was performed and compared with published data from a conventional small-scale SMR system. Results indicate that the ICFBMR can achieve higher process efficiencies, but that membrane cost and longevity are critical to making the process economically viable.

Text

2011-01-20

For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use.

http://hdl.handle.net/2429/30745

Chemical and Biological Engineering

University of British Columbia

Graduate