Hydrogen Use in CO2 Capture Technologies

Through international world-class cooperation with UC Berkeley and Sandia National Laboratories in the US, the HYCAP project aims at answering fundamental questions targeted to the development of the next-generation carbon capture and storage technologies (CCS). Specifically, HYCAP focuses on efficient and reliable large-scale energy conversion of hydrogen, which is a key part to several pre-combustion CO2 capture strategies. By the pre-combustion CO2 capture concept, the carbon is removed from the fossil fuel (which may be natural gas or coal) prior to the combustion, and the remaining hydrogen-rich gas is utilized for power production. A main objective of the project has been to develop a high-fidelity numerical design tool for applications to challenging flow configurations relevant to the development of hydrogen-fired gas turbines for power generation. Hydrogen has different thermophysical properties than natural gas, and issues related to autoignition, flame stabilization and flashback must be addressed in order to develop novel gas turbines for hydrogen-rich fuels. The design of new gas turbine combustors is required to ensure efficient, safe, and reliable conversion of hydrogen, while at the same time environmental concerns such as low emissions of NOx are ensured. Part of the project work has involved an NTNU PhD student who has collaborated closely with Prof. J-Y Chen at UC Berkeley. The PhD student was enrolled at NTNU in September 2015 and spent the academic year 2016-17 as a visitor at the Department of Mechanical Engineering at UC Berkeley. The PhD student is expected to graduate from NTNU in December 2018, with a possible delay to February 2019. An MSc student at NTNU has also contributed to the project through his master thesis based on numerical simulations of a hydrogen-rich jet. The numerical simulation tool has been developed in the project to do computations on hydrogen flames specifically, and turbulent reactive flows more generally, and has been applied to challenging flame configurations with relevance to hydrogen-fired gas turbines. The simulation code has also been parallelized for computations on large clusters, i.e., sets of connected computers that work together. Such parallel computing is an important and necessary part of detailed computations of combustion processes. In order that the computations give as accurate as possible representation of the combustion, it is necessary to resolve the reactive flow down to the smallest turbulent scales, which in general requires very high computational powers. Incidentally, the numerical simulation tool developed in HYCAP is built on an innovative model and algorithm that substantially reduces the effective run time compared to other tools that offer the comparative fidelity. In short, HYCAP has contributed to in-depth knowledge and an advanced numerical design tool which has the potential to be important in enabling the pre-combustion CO2 capture concept.

The unique numerical design tool LEM3D developed in HYCAP has the great benefit that it may provide detailed and accurate information on combustion characteristics at a relatively low computational cost. In HYCAP, LEM3D has specifically been developed for applications to flame configurations with relevance to hydrogen-fired gas turbines. The LEM3D code constitutes a valuable test ground and tool for learning and competence building for future PhD and MSc students. In a longer-term perspective, the capabilities of LEM3D have the potential to be integrated into commercial state-of-the-art CFD simulation codes. This should be of great interest to the developers of commercial CFD software and beneficial to the industrial end-users of such. The LEM3D tool has the potential to be of important future value for the enabling of hydrogen-fired gas turbine combustors specifically, and combustion applications and the pre-combustion CO2 capture concept more generally.

Predictive modeling of mixing and reaction in turbulent flow environments would accelerate the development of next-generation carbon capture technologies. The quest for such capability at the needed scale and level of detail confronts the fundamental scie ntific challenge of developing a reduced physical/mathematical representation of turbulence and its interactions with chemical reactions and related sub-processes. Without such a reduced description, the needed fidelity is available only from Direct Numer ical Simulation (DNS), i.e., numerical solution of the exact equation set. DNS is computationally affordable only at scales much smaller than needed, but is useful, e.g., for validation of reduced models, and such use is part of the research proposed here . An ongoing collaborative effort by SINTEF and international partners has established a promising pathway to attaining the needed capability. Conceptual innovations addressing this need have led to the RANS-LEM3D approach for modeling mixing and chemical reactions in turbulence. Initial application of this approach to combustion technology development is being pursued through the ongoing CAMPS project. Complementary to CAMPS, the effort proposed here will use RANS-LEM3D and the underlying LEM method to a ddress technological challenges associated with the combustion of hydrogen within pre-combustion CO2 capture concepts. The challenges are linked to the particular thermo-physical properties of hydrogen compared to conventional hydrocarbons, leading to dra matically different combustion behavior, e.g. with regard to auto-ignition, flame stabilization, flashback, and NOx production. The research proposed here will focus on (1) the interaction between hydrogen combustion phenomenology and membranes used for h ydrogen separation, and (2) hydrogen jets and jet flames in cross-flow, for which DNS predicts significant counter-gradient diffusion effects with potential impacts on the noted technological challenges.