The existence of a secondary discontinuity at the rear of a detonation fron
t shown in experiments by Peraldi and Veyssiere (1986) in stoichiometric hy
drogen-oxygen mixtures with suspended 20-mu m starch particles has not been
explained satisfactorily. Recently Veyssiere et al. (1997) analyzed these
results using a one-dimensional (1-D) numerical model, and concluded that t
he heat release rate provided by the burning of starch particles in gaseous
detonation products is too weak to support a double-front detonation (DFD)
; in contrast to the case of hybrid mixtures of hydrogen-air with suspended
aluminium particles in which a double-front detonation structure was obser
ved by Veyssiere (1986). A two-dimensional (2-D) numerical model was used i
n the present work to investigate abovementioned experimental results for h
ybrid mixtures with starch particles. The formation and propagation of the
detonation has been examined in the geometry similar to the experimental tu
be of Peraldi and Veyssiere (1986); which has an area change after 2 m of p
ropagation from the ignition point from a 69 mm dia. section to a 53 mm x 5
3 mm square cross section corresponding to a 33% area contraction. It is sh
own that the detonation propagation regime in these experiments has a diffe
rent nature from the double-front detonation observed in hybrid mixtures wi
th aluminium particles. The detonation propagates as a pseudo-gas detonatio
n (PGD) because starch particles release their heat downstream of the CJ pl
ane giving rise to a non-stationary compression wave. The discontinuity wav
e at the rear of the detonation front is due to the interaction of the lead
ing detonation front with the tube contraction, and is detected at the fart
hest pressure gauge location because the tube length is insufficient for th
e perturbation generated by the tube contraction to decay. Thus, numerical
simulations explain experimental observations made by Peraldi and Veyssiere
(1986).