The propagation of two-dimensional cellular gaseous detonation bounded by an inert layer is examined via computational simulations. The analysis is based on the high-order integration of the reactive Euler equations with a one-step irreversible reaction. To assess whether the cellular instabilities have a significant influence on a detonation yielding confinement, we achieved numerical simulations for several mixtures from very stable to mildly unstable. The cell regularity was controlled through the value of the activation energy, while keeping constant the ideal Zel’dovich - von Neumann - Döring (ZND) half-reaction length. For stable detonations, the detonation velocity deficit and structure are in accordance with the generalized ZND model, which incorporates the losses due to the front curvature. The deviation with this laminar solution is clear as the activation energy is more significant, increasing the flow field complexity, the variations of the detonation velocity, and the transverse wave strength. The chemical length scale gets thicker, as well as the hydrodynamic thickness. The sonic location is delayed due to the presence of hydrodynamic fluctuations, for which the intensity is increased with the activation energy as well as with the losses to a lesser extent. The flow field has been studied through numerical soot foils, detonation velocities, and 2D detonation front profiles, which are consistent with experimental findings. The velocity deficit increases with the cell irregularity. Moreover, the relation between the detonation limits obtained numerically and in detonation experiments with losses is discussed.

The paper presents an investigation on water drop breakup in the 'catastrophic' mode at Weber numbers above 10 5. Experimental data have been obtained on a detonation shock tube operated at a Mach number between 4.2 and 4.6. Displacement and deformation of the mist cloud generated around the droplet were observed with a shadowgraph system, and Schlieren imagery was used to visualise the bow and wake shocks around the droplet. We observe that all measured quantities scale with the initial drop diameter. In order to analyse the experimental results and relate these observables to the breakup process, numerical hydrodynamic simulations have been performed. Once validated by direct comparison with our experimental observations, the simulation results are used to see "through" the mist and infer the droplet evolution with dimensionless time T. According to our results, the breakup mechanism can be divided into 3 steps. First (T < 1), most of the liquid mass remains in one main drop, whose shape flattens due to the hydrodynamic forces; then (1 < T < 2) fragmentation begins along the outer rim of the liquid drop and splits the corresponding mass into two parts; the first spreads out radially in small fragments while the remains finally (2 < T < 3.5) take the shape of a filament aligned with the flow. Our results are consistent with a complete breakup time T b = 5.5.