Simulation and modelling of hydrogen production by sorption enhanced steam methane reforming in fixed bed reactors
MetadataVis full innførsel
An increased demand for hydrogen as energy-carrier and as fuel for clean power generation is expected during the 21st century. The Kyoto-protocol states that the world has to decrease it’s CO2-emissions to the atmosphere. A concept combining hydrogen production and sequestration of CO2 is the sorption enhanced steam methane reforming (SE-SMR) process. This is an alternative to the traditional steam methane reforming (SMR) for production of hydrogen. SE-SMR is a concept that has received increased attention in recent years. The process utilizes a solid CO2-acceptor to capture CO2 in the reforming reactor and thereby change the normal thermodynamic limitations of steam methane reforming. The work in this thesis has focused on simulation of hydrogen production by sorption enhanced steam methane reforming in a fixed bed reactor. A robust transient one dimensional model has been formulated and implemented for the simulations of the reforming reactor. Three main models have been formulated, one pseudo-homogeneous model and two heterogeneous model that account for intraparticle mass and heat transfer. The two heterogeneous models are different in the way the solid materials are placed in the reactor. The 1-particle model considers one type of pellet in the reactor consisting of both catalytic and sorbent material, while the 2-particle model considers two separate pellet types with catalytic and sorbent material. Kinetic models for all major reactions must be formulated to simulate the sorption enhanced steam methane reforming reactor. The steam methane reforming reactions have been extensively studied earlier, and the kinetic model of Xiu and Froment was used in the simulations. Different solid synthetic materials for the high temperature CO2 capture have been studied, and kinetic models for capture of CO2 on these materials have been formulated in this thesis. Two of the materials, nanocrystalline lithium zirconate and sodium zirconate have been synthesized at NTNU, while the lithium silicate was obtained from Toshiba. The materials synthesized at NTNU showed quite similar kinetic properties, and the capture rate of CO2 was described by a first order rate reaction with respect to fractional conversion of the solid. However, while the shape of the rate expression was similar for the two zirconates, the reaction rates did differ substantially. The lithium zirconate had the slowest capture rate of the materials; with a kinetic constant about 100 times lower than the one for sodium zirconate, which showed the fastest kinetics. The capture rate on lithium silicate was found to be between the two other materials. The reactor simulations of SE-SMR show that it is possible to produce hydrogen with purity above 80 % on a dry basis in a fixed bed reactor with the investigated sorbents. The reactor performance are highly dependent on the capture kinetics and while hydrogen with over 80% purity could be produced with a superficial velocity of 2 m/s and a steam to carbon ratio of 3 with Na2ZrO3 as sorbent, a superficial velocity of 0.3 m/s and a steam to carbon ratio of 5 was necessary for Li2ZrO3. The performance of Li4SiO4 is between these two materials. In all reactor simulations it was found that there will be large temperature gradients in the reactor even if the total reaction is not very endothermic. If a fixed bed reactor is operated without external heating, the temperature close to the inlet will decrease dramatically, while the outlet temperature will increase. This means that temperature control could be necessary during hydrogen production, and that heating/ cooling not only has to be supplied during regeneration. The results from the reactor simulations have been incorporated in computer simulations of the whole process of producing pure hydrogen by SE-SMR. Sorption enhanced reforming, regeneration of sorbent, heating and cooling of the reactor bed, CO2 compression, H2 purification by presure swing adsorption, and heat integration of the process are the main parts of the hydrogen production process. The data on regeneration has been limited and the process sizing and performance is mainly based on the reforming being the limiting factor. The efficiency is very dependent on the amount of heat that must be supplied to the reactor in the regeneration step and the CO2-sorption kinetics of the sorbent. A thermal efficiency of 0.71 was calculated for an SE-SMR process with Na2ZrO3 as sorbent with a gas velocity of 2 m/s, a 10 m long reactor, a pressure of 10 bar, a reforming temperature of 848 K and a steam to carbon ratio of 3. The heat for regeneration was in this case supplied by combustion of methane in pure oxygen and yielded high CO2 removal (≈100%). The thermal efficiency is comparable and better than for autothermal reforming with CO2 removal by an amine process. The temperature of regeneration used for Li2ZrO3 as sorbent were only 52°C higher than the reforming temperature, while it was 325°C higher for Na2ZrO3. When Li2ZrO3 replaced Na2ZrO3 as CO2-acceptor, the lower temperature increase for regeneration, which lead to less heat supplied, did make up for some of the disadvantages of Li2ZrO3, but in total the slower kinetics make it a less promising sorbent for SE-SMR. With all other parameters equal the thermal efficiency fell from 0.71 to 0.67 when using Li2ZrO3 as acceptor instead of Na2ZrO3. At the same time the cross-section area of the reactor had to be increased almost 10 times to get the throughput that was necessary to have equal production of hydrogen. Producing hydrogen with a total lower heating value of 700 MW the necessary reactor cross-sectional area was about 20 m2 for the simulation with Na2ZrO3 as sorbent, while the necessary crossectional area for Li2ZrO3 were 173 m2. The possibility of not producing pure hydrogen, but a mixture of hydrogen and methane by SE-SMR with Na2ZrO3 as sorbent has also been investigated. With the low conversion of methane, the CO2-capture simulated for this process was only 62 % when the methane content in the product is calculated as CO2-equivalents. The thermal efficiency of this process was about 0.83, compared to the 0.71 for the case with pure hydrogen as product. If the kinetics of CO2-sorption could be increased, giving higher hydrogen content in the product, an increased CO2 removal can be reached without lowering the thermal efficiency.