A contribution to the study of the near-field contrail by using a high-fidelity computational method combined to an advanced microphysical model

Abstract

Global emphasis on environmental protection, expected growth in air traffic, and accumulating evidence on aviation’s non-CO2 radiative forcing have made contrails an important component in assessments of aviation climate impacts. The climate influence of cirrus formed from aircraft condensation trails remains one of the most uncertain components of anthropogenic radiative forcing. Observations and models are used to clarify how ice crystals form in the near-field aircraft wake and to estimate global distributions of their microphysical and optical properties. The radiative effect of contrails depends on ice crystal size and number concentration, yet both are difficult to quantify because of assumptions about wake dynamics and simplified treatments of early ice-phase microphysics. The objective of this thesis is to develop and validate a modeling framework that simulates exhaust jet flow behind a realistic turbofan and its early entrainment into the wake-vortex field, while integrating online-coupled microphysical processes to determine the initial properties of ice particles. Dynamic simulations are performed with an in-house CFD code (FludiLES) using large-eddy simulation (LES) with sixth-order spatial and third-order temporal accuracy. A free jet at atmospheric pressure and a realistic Reynolds number (Re≈106) is computed and validated against experimental results. For the microphysics, soot and ambient aerosol particles are tracked with a Lagrangian method. Two-way coupling is applied to vapor mass transfer and energy, whereas momentum is one-way (fluid → particles); particle-to-fluid momentum feedback is negligible, and energy feedback is small. A dual-stream jet (core and bypass) is simulated to quantify bypass ratio effects using three engines—CFM56-5B3/3, Leading Edge Aviation Propulsion (LEAP)-1A/33, and an ultra-high-bypass-ratio (UHBR) configuration—at ambient temperatures of 215, 220, and 225 K. An initial comparison of a turbofan with an equivalent turbojet showed that the bypass stream accelerated near-field mixing and cooling of the core plume, increasing the mean particle radius by up to 30%. Across the engine set, larger bypass ratios further intensified mixing, promoted soot activation and subsequent freezing, and yielded initial contrail properties that were less sensitive to ambient temperature. Subsequently, the microphysical model was extended to include solute effects through the hygroscopicity parameter (κ) and was evaluated in both online-coupled and offline box model configurations. Under realistic LEAP-1A cruise conditions, three scenario groups are examined: (i) κ = 0.0005, 0.005, 0.0142, corresponding to equivalent fuel sulfur contents (FSC) of 50, 410, and 1270 ppm; (ii) soot number emission indices of 1013 to 1015 #/kg-fuel; and (iii) soot core radii of 10–30 nm. Reducing κ from 0.0142 to 0.0005 slightly decreases particle radius but increases the activation fraction by about 20% due to greater vapor availability. Lowering the initial soot number from 1015 to 1013 particles per kilogram of fuel increases the mean particle radius from roughly 0.3 μm to 2.4 μm at 1 s and raises the activated fraction by about 66%. Larger soot cores enhance activation by about 20%, with mean-radius differences reaching approximately 80% near 0.4 s and narrowing to about 10% by 1 s. Three-dimensional LES with online-coupled microphysics is shown to be essential for near-field sensitivity analysis; a zero-dimensional offline box model overpredicts activation and misrepresents κ sensitivity. Simulations are then extended into the vortex phase to compare temporal and spatial LES formulations and to quantify how numerical methodology influences microphysics for soot-only and soot+ambient scenarios. To assess the effects of initialization, the wake-vortex was initialized from jet phase modeling at tj= 0.12s and tj=0.5s. Under cruise conditions representative of a LEAP-1A powered Airbus A320neo, the temporal modeling shows stronger early-stage mixing and faster scalar dissipation in the jet phase than the spatial modeling, yielding approximately 60% more activated ice crystals in the soot-only case and approximately 70% more when ambient aerosols are included. At early vortex initialization, temporally initialized cases sustained slightly higher activated ice numbers, whereas spatial initialization produced larger late-time mean radii; these disparities diminished, particularly when ambient aerosols were present. At late vortex initialization, the trend reversed: spatial initialization yielded higher activated ice numbers, while temporal initialization retained the larger mean radius. In all cases, ambient aerosol increases ice number yet limits per-crystal growth through vapor competition, which moderates sensitivity to jet initialization. Overall, vortex phase entrainment is relatively insensitive to the jet phase modeling choice; however, the vortex start time remains consequential, and absolute microphysical values retain a measurable memory of near-field jet mixing.

Date	21 Jan 2026
Original language	American English
Awarding Institution	École de technologie supérieure
Supervisor	François Garnier (Supervisor)