Back to search

University of Minnesota

Multi-Fidelity Frameworks for Plasma-Assisted Reacting Flows: From Plasma Chemistry Reduction to Phenomenological Discharge Modeling

Abstract

dc:description.abstract

Non-equilibrium plasma discharges are increasingly used to assist combustion and reforming systems that are difficult to sustain by conventional means. Ammonia, an alternative carbon-free fuel, ignites slowly and produces substantial NOx emissions when burned in air. Methane and other small alkanes also pose challenges as they resist efficient conversion into higher-value products at moderate temperatures. Certain synthetic fuels such as low-cetane synthetic aviation fuels also face practical challenges due to their unreliable ignition characteristics in compression-ignition engines while using conventional ignition assistance techniques. Non-equilibrium plasma addresses each of these limitations by producing reactive radicals directly through electron-impact dissociation and exciting vibrational and electronic states that subsequently relax into gas heating, thereby depositing energy on timescales far faster than conventional thermal heating. Investigating the application of non-equilibrium plasmas numerically is a monumental challenge due to the coupled multi-physics nature of the problem, involving fluid flow, gas and plasma chemical kinetics, and electrostatics. This thesis develops a multi-fidelity set of computational frameworks for plasma-assisted reacting systems, spanning detailed chemical-kinetic analysis and mechanism reduction at one end and fast, practical-scale phenomenological modeling at the other, with higher-fidelity plasma-fluid simulations used to inform and improve the fidelity of the phenomenological models. A coupled plasma-combustion mechanism routinely spans hundreds of species and reactions, including excited and charged states which absorb and release energy through ionizations, dissociations, and excitation--de-excitation cyclic reaction channels. Conventional path-flux and global-pathway analysis methods trace only simple paths with no repeated species, so they cannot represent these cycles even when they dominate the system's reactivity. Plasma-based Global Pathway Analysis (PGPA) extends global pathway analysis with a dedicated cycle-identification algorithm and three metrics that rank both pathways and cycles by their contribution to reactivity: element-flux dominance, gas heating, and radical production. Applied to a zero-dimensional nanosecond-pulsed ammonia-air combustion, PGPA attributes the difference in ignition behavior between a low and a high reduced electric field to a shift from vibrational to electronic excitation-relaxation cycles. It also traces the lower peak NOx observed under plasma assistance to an early, plasma-driven reforming pathway that consumes fuel-bound nitrogen before it can form NO. The same framework identifies the electron-impact and electronically excited nitrogen pathways responsible for the enhanced C₂ species production observed during plasma-assisted methane reforming relative to thermal reforming at much higher total energy input. Subsequent analysis shows the same mechanism generalizing to a representative multi-alkane flare-gas mixture, where the resulting reforming shortens ignition delay by 24-40 % and reduces the equilibrium methane fraction by 5-10 % relative to the unreformed mixture. Despite the utility of zero-dimensional calculations and analyses, it overlooks the effects of space-charges and the corresponding plasma transport. Simply employing the detailed plasma kinetics in spatio-temporal problems such as two- or three-dimensional calculations is computationally challenging. Plasma-based Global Pathway Selection (PGPS) extends PGPA from an analysis tool into a mechanism-reduction algorithm. It retains the pathways, cycles, and individual reactions that PGPA identifies as dominant, while additionally enforcing that any retained ionization channel is paired with a corresponding neutralization channel so the reduced mechanism remains charge-conserving. Applied to a detailed, 80-species ammonia plasma-combustion mechanism, PGPS produces a family of reduced mechanisms, down to 42 species, that reproduce the detailed mechanism's zero-dimensional ignition delay, temperature, and major-species histories to within roughly 10 % at every level of reduction. A two-dimensional axisymmetric fluid-plasma model, with a new electron-temperature formulation and a mixture-specific photoionization treatment, shows that this zero-dimensional accuracy does not carry over to streamer propagation. The most aggressively reduced mechanisms predict the streamer connection 29 % faster than the detailed mechanism despite their close zero-dimensional agreement. This discrepancy is traced to the loss of specific nitrogen vibrational excitation channels and a weakened ion-electron recombination pathway. A larger, 64 species reduced mechanism that retains the nitrogen vibrational excitation channels is found to closely match the detailed mechanism's streamer propagation. In spite of their accuracy, detailed kinetics and fluid-based models are infeasible for the investigation of practical systems due to their computational complexity. In that regard, phenomenological models are of greater utility owing to their numerical simplicity. The phenomenological nanosecond-discharge model of Castela et al., which represents a plasma pulse through spatially and temporally uniform source terms for ultra-fast heating, dissociation, and vibrational excitation, is applied to two practical problems. In a compression-ignition engine, plasma ignition of a low-cetane synthetic aviation fuel is shown to depend on the igniter intersecting the fuel spray to benefit from repeating low-energy pulses, and to succeed at high-altitude conditions where a conventional spark misfires outright. This model is then extended to a four-way energy partition that tracks methane dissociation and vibrational excitation, and when applied to methane reforming, shows a hot core that dissociates fuel while allowing recombination products to survive only in a cooler outer shell. Also, splitting a fixed total energy across multiple lower-energy pulses changes which relaxation process dominates the heating. Both applications, however, share the same underlying simplification of spatially and temporally fixed energy partitioning, an assumption that is inaccurate for high-frequency pulsing. Two extensions address this limitation directly. The first replaces the fixed, error-function plasma power density used by the original model with a spatiotemporally resolved one, mapped from two-dimensional axisymmetric fluid-based plasma simulations of air discharge. Validated against experimental gas-heating and radical-production data, the resulting model reaches the same temperature and O-radical increase in roughly 19 nanoseconds, against roughly 50 nanoseconds for the original model. It also reproduces the post-discharge pressure-wave and heated-channel radii without needing to artificially downscale the deposited energy, unlike the original model. Across three discharge gap distances, the improved model correctly reproduces streamer initiation at the electrodes, propagation toward the gap center, and connection, in contrast to the original model's spatially uniform, temporally invariant channel. The second extension replaces the fixed energy-partition fractions with a two-dimensional manifold tabulated against the reduced electric field and a combustion progress variable. This enables the model to respond locally to the evolving thermochemical state rather than remaining fixed at its fresh-mixture value. Applied to repetitively pulsed methane-air discharges, the manifold-based model captures how the dominant radical-production pathway shifts from oxygen dissociation toward the dissociation of combustion products as later pulses discharge into partially burnt gas. Together, the pathway-analysis and mechanism-reduction methods and the phenomenological modeling framework developed in this thesis help enable the detailed chemical kinetics and transport of non-equilibrium plasmas to device-scale geometries, with zero- and two-dimensional detailed plasma simulations providing the link that improves the fidelity of the fastest, most practical models.

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Johnson, Praise Noah

Subjects

dc:subject × 6

Rights

Language dc:language.iso
en

Identifiers

dc:identifier.*
Dc Identifier Citation
Johnson, Praise Noah. (2026). Multi-Fidelity Frameworks for Plasma-Assisted Reacting Flows: From Plasma Chemistry Reduction to Phenomenological Discharge Modeling. Retrieved from the University Digital Conservancy, https://hdl.handle.net/11299/283445.

Chain of custody

source
Harvested from
University of Minnesota
Base URL
conservancy.umn.edu/server/oai/request
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
citation

Johnson, Praise Noah. Multi-Fidelity Frameworks for Plasma-Assisted Reacting Flows: From Plasma Chemistry Reduction to Phenomenological Discharge Modeling. 2026. https://hdl.handle.net/11299/283445