Ahigh-resolution Eulersolverhas been used to investigatethee owe eld induced by a blast wavein a double-bent duct. A blast-wave/vortex interaction and shock/shock interactions are also studied. A pair of vortices with an opposite rotating direction is found. The trajectories of the induced main vortices are explored for different initial pressure ratios. It was found that thevortex could deform theshock front of the blast wave becauseof compression and expansion involved in their interaction. Moreover, after blast-wave diffraction around the e rst convex corner a locally maximum wall pressure could occur at the second concave corner, and a globally maximum wall pressure could occur at the e rst concave corner, resulting in a large loading at the compression corners. NGINEERS and scientists are quite interested in the problem of blast-wave propagation in ducts because the phenomena of blast-wave propagation often take place,for example,in the exhaust pipes of internal combustion engines or, caused by explosions, in- side a building. A fundamental study of blast-wave propagation in a duct is helpful in the understanding of the blast-wave dynamics. In this study we investigated the unsteady e ow of planar blast-wave propagation in a double-bent duct with two 90-deg turns. The prob- lem involved a complicated unsteady e owe eld of blast-wave/vortex interactions, shock/shock interactions, and blast-wave ree ections. Inthepasttherewereveryfewpapersthatreportedthedetailofthe e owe eldinducedbyablastwaveinadouble-bentduct.Somerelated papers are referred to here. Howard and Mathews 1 experimentally investigated the vortex generation caused by shock-wave diffrac- tion around a corner. Schardin 2 studied a ree ected shock/vortex interaction by using the shadowgraph method. His experimental results provided basic understanding of shock/vortex interaction. Mandella et al. 3 conducted a series of experiments for studying the vortex induced by a shock wave near a corner. They reported that the vortex generation was not caused by the e uid viscosity. Sivier et al. 4 numerically simulated the problem studied by Schardin and obtained good agreement between the numerical and experimental results. Igra et al. 5 used both experimental and numerical methods to study the intensity of a diffracted wave in a branched duct and concluded that the e ow behind the diffracted wave could not be assumed to be quasi-one-dimensional. Ellzey and Henneke 6 inves- tigated the sound generation caused by the shock/vortex interaction and its mechanism. They found that the shock-wave deformation and a compression effect involved in the interaction are the two main factors for sound generation. Grasso and Pirozzoli 7 numeri- callystudiedashock/vortexinteractionandclassie edtheinteraction intothree types.They also characterized theshockandvortexdefor- mationsandthemechanismofsoundgeneration intheinteractionof a cylindrical shock wave with a vortex. The numerical scheme they used is a weighted essentially nonoscillatory (WENO) scheme of