They say you never forget your first love.
I like to hug engines as much as the next Convergent Science employee, but I may live out my entire engineering career as a scramjet guy at heart. In foundational and open-literature research, scramjets are almost synonymous with hydrogen. It's easy to understand why—hydrogen offers a number of advantages for such work. Hydrogen burns ferociously fast, and it auto-ignites at scramjet-ambient temperatures. It is extremely energy-dense, nearly double the energy per unit mass of any hydrocarbon. Of course, hydrogen vaporizes exceptionally readily. Finally, for us computationalists, it is completely reasonable to describe the entire hydrogen-air combustion mechanism—no need to worry about mechanism design and reduction, or develop a reduced-order combustion model.
My simulations in graduate school focused on studying hydrogen-air mixing, but several of my colleagues ran successful combusting simulation campaigns. Their technique could be roughly summarized as "make sure you have enough memory available, then turn combustion on". This isn't to suggest a lack of sophistication, but it indicated that we were already accounting for all of the other important physical processes, or that these processes overwhelmed other effects that we were choosing to neglect.
If my system is like that, then every system is like that, right? So when I looked at the 2023 CONVERGE CFD Conference agenda, why did I see so many presentations that would talk about the challenges of hydrogen?
Though you don't forget your first love, you can teach a dog new tricks. That bit about scramjet processes overwhelming other effects—that bit is important. I was surprised to learn, or re-learn, about many of the challenges specific to an internal combustion engine or gas turbine that we simply didn't have to worry about in the scramjet community. What makes hydrogen great for my old team makes it tricky for my current one.
The inside of an IC engine or GT combustor is a violent place, but it's not violent enough to overwhelm some of hydrogen's peculiarities. The very high flamespeed I used to love, in combination with a very thin flame front, forces us to run a simulation with a very fine grid and accordant small time-step. Some sort of flame thickening strategy, or a reduced-order combustion model, is an enabling capability for an engineering-relevant calculation. It gets worse. Because hydrogen is very light compared to air, interesting fluid-mechanical instabilities and intermolecular diffusion effects become important. Flame fronts self-wrinkle rapidly, increasing reaction rates further.
Hydrogen's extreme affinity for combustion presents other problems. Hydrogen has a very wide flammability range, which increases the importance of understanding and predicting engine knock, cycle-to-cycle variability, and flashback. It introduces new challenges, like flame arrestment–a hydrogen flame front can propagate through narrow, sub-millimeter-scale gaps between components. In a geometrically complex configuration like an IC engine or a gas turbine, we must become more exacting with our crevice modeling, valve events, etc.
Finally, we must relax the assumption I made about a small chemical mechanism. There is great engineering interest in modeling hydrogen combustion in mixed-fuel systems, using hydrogen's violent nature to initiate or sustain combustion with a fuel like methane or ammonia. This reintroduces our interest in mechanism selection and reduction for simulating detailed chemistry, as well as development and refinement of reduced-order combustion models.
If you want to learn more about how engineers are taking on the challenges of hydrogen modeling, check out the publicly-available presentations from the 2023 CONVERGE CFD Conference on our website. You can also visit our hydrogen modeling webpage.