A stability index for detonation waves in Majda's model for reacting flow (Q1885888)

In one-dimensional Lagrangian coordinates the Navier-Stokes equations of reacting flow for a one-step reaction may be written in the abstract form \[ u_t +f(u)_x=(B(u)u_x)_x+kq \phi(u) z, z_t=(C(u,z)z_x)_x-k\phi(u)z. \tag{1} \] Under different conditions at \(x=+\infty,-\infty\), there can result a variety of types of waves solving (1): nonreactive gas-dynamical shock and rarefaction solutions (\(z=0\), or \(z=\) constant and \(\phi=0\)), and traveling combustion waves consisting of weak and strong detonations, weak and strong deflagrations, and Chapman-Jouguet detonation and deflagration waves, which are limiting cases dividing weak and strong branches. Similarly as in the case of real, e.g., van der Waals, gas dynamics, the multitude of possible such elementary waves leads to a multitude of possible time-asymptotic states, and these must be classified according to stability. Up to now, essentially all analysis have been carried out for one of three simplified models; (i) ZND model \((B=C=0),\) (ii) CJ or square-wave model \((B=C=0, k\) is taken to be infinite), (iii) Majda model, for which \(u\) is taken to be a scalar and \(B\) and \(C\) are set to \(1\) and \(o\) (or sometimes \(1)\), respectively. At one time, it seems to have been belived, based on analysis of the ZND, that weak detonations and strong deflagrations were unstable, the other types stable at least in moderate ranges. This conjecture on weak detonation is now widely agreed to be false when viscosity and other effects are taken into account. However, rigorous analysis of stability for the full model (1) or comparison with stability for the ZND or CJ appoximations, remain important open problems. The purpose of this paper is to initiate a large-scale study of these problems. In the simple setting of the Majda model, these techniques were designed for the study of systems, \(u\in \mathbb R^n, n>1\), so may be applied also in the case of the full reacting Navier-Stokes equations. The resulting stability condition is satisfied for all nondegenerate, i.e. spatially exponentially decaying, weak and strong detonations of the Majda model in agreement with numerical experiments of [\textit{P. Colella, A. Majda} and \textit{V. Roytburd}, SIAM J. Sci. Stat. Comput. 7, 1059--1080 (1986; Zbl 0633.76060)] and analytical results of [\textit{T.-P. Liu} and \textit{S.-H. Yu}, Commun. Math. Phys. 204, 551--586 (1999; Zbl 0976.76036); \textit{A. Szepessy}, Commun. Math. Phys. 202, 547--569 (1999; Zbl 0947.35019)] for a related model of Majda and Rosales. Also the authors discuss the role in the ZND limit of degenerate, subalgebraically decaying weak detonation and (for a modified,``bump-type'' ignition function) deflagration profiles, as discussed in [\textit{I. Gasser} and \textit{P. Szmolyan}, SIAM J. Math. Anal. 24, 968--986 (1993; Zbl 0783.76099), SIAM J. Appl. Math. 55, 175--191 (1995; Zbl 0814.34028)] for the full equations.