From graph theory and geometric probabilities to a representative width for three-dimensional detonation cells

TL;DR

This paper introduces a stochastic graph-theoretic model to predict the representative width of three-dimensional cellular detonation fronts in reactive gases, aligning well with experimental data for certain fuel mixtures.

Contribution

The model combines graph theory, geometric probabilities, and ZND calculations to accurately predict detonation cell widths, accounting for irregular cellular patterns in 3D reactive gas fronts.

Findings

Model predicts cell widths with high accuracy for H2, C3H8, C2H4 mixtures.

Overestimates for CH4:O2 suggest additional ignition mechanisms like turbulence.

Model is computationally efficient and easily integrated with ZND profiles.

Abstract

We present a model for predicting a representative width for the three-dimensional irregular patterns observed on the front views of cellular detonation fronts in reactive gases. Its physical premise is that the cellular combustion process produces the same burnt mass per unit of time as the average planar steady Zel'dovich-von Neuman-D\"oring (ZND) process. The transverse waves of irregular cells are described as a stochastic system subject to a stationary ergodic process, considering that the distributions of the patterns should have identical temporal and statistical average properties. Graph theory then defines an ideal cell whose grouping is equivalent to the actual 3D cellular front, geometric probabilities determine the mean burned fraction that parameterizes the model, and ZND calculations close the problem with the time-position relationship of a fluid element in the ZND reaction zone. The model is limited to detonation reaction zones whose only ignition mechanism is adiabatic shock compression, such as those of the mixtures with H2, C3H8 or C2H4 as fuels considered in this work. The comparison of their measured and calculated widths shows an agreement better than or within the accepted experimental uncertainties, depending on the quality of the chemical kinetic scheme used for the ZND calculations. However, the comparison for CH4:O2 mixtures shows high overestimates, indirectly confirming that the detonation reaction zones in these mixtures certainly include other ignition mechanisms contributing to the combustion process, such as turbulent diffusion. In these situations, the cell widths on longitudinal soot recordings have a very large dispersion, so a mean width may thus not be a relevant detonation characteristic length. The model is easily implementable as a post-process of ZND profiles and provides rapid estimates of the cell width, length and reaction time.