A reactive system’s slow dynamic behavior is approximated well by evolution on manifolds of dimension lower than that of the full composition space. This work addresses the construction of one-dimensional slow invariant manifolds for dynamical systems arising from modeling unsteady spatially homogeneous closed reactive systems. Additionally, the relation between the systems’ slow dynamics, described by the constructed manifolds, and thermodynamics is clarified. It is shown that other than identifying the equilibrium state, traditional equilibrium thermodynamic potentials provide no guidance in constructing the systems’ actual slow invariant manifolds. The construction technique is based on analyzing the composition space of the reactive system. The system’s finite and infinite equilibria are calculated using a homotopy continuation method. The slow invariant manifolds are constructed by calculating attractive heteroclinic orbits which connect appropriate equilibria to the unique stable physical equilibrium point. Application of the method to several realistic reactive systems, including a detailed hydrogen-air kinetics model, reveals that constructing the actual slow invariant manifolds can be computationally efficient and algorithmically easy.
An Eulerian–Lagrangian modelling approach was employed in order to investigate the flow field, heat transfer and particle distribution in nanofluid flow in a parallel-plate microchannel, with a focus on relatively low Reynolds numbers ( $Re\leqslant 100$ ). Momentum and thermal interactions between fluid and particle phases were accounted for using a transient two-way coupling algorithm implemented within an in-house code that tracked the simultaneous evolution of the carrier and particulate phases while considering timescale differences between the two phases. The inaccuracy of assuming a homogeneous particle distribution in modelling nanofluid flow in microchannels was established. In particular, shear rate and thermophoresis were found to play a key role in the lateral migration of nanoparticles and in the formation of particle depletion and accumulation regions in the vicinity of the channel walls. At low Reynolds numbers, nanoparticle distribution near the walls was observed to gradually flatten in the streamwise direction. On the other hand, for relatively higher Reynolds numbers, higher particle non-uniformities were observed in the vicinity of the channel walls. Furthermore, it was established that convective heat transfer between channel walls and the bulk fluid can either improve or deteriorate with the addition of nanoparticles, depending on whether the flow exceeded a critical Reynolds number of enhancement. It was also established that Brownian motion and thermophoresis had a major role in nanoparticle deposition on the channel walls. In particular, Brownian motion was the main deposition mechanism for nano-sized particles, whereas due to thermophoresis, nanoparticles were repelled away from channel walls. The result of the competition between the two is that deposition gradually increased along the streamwise direction.
A new skeletal mechanism, consisting of 16 species and 72 reactions, has been developed for lean methane–air premixed combustion from the GRI-Mech 3.0. The skeletal mechanism is validated for elevated unburnt temperatures (800 K) and pressures up to 4 atm, thereby addressing realistic gas turbine conditions. The skeletal mechanism is obtained by applying the directed relation graph method and performing sensitivity analysis on the detailed mechanism. The mechanism has been validated for flame speed and flame structure in a wide range of conditions and configurations. A good agreement between the skeletal mechanism and GRI-3.0 was obtained. The configurations considered include one-dimension laminar premixed flames, laminar non-premixed counterflow burners, and two- and three-dimensional unsteady configurations with variations of temperature, pressure, and composition. The skeletal mechanism allows for the inclusion of accurate finite rate chemistry in large-scale direct numerical simulations of lean turbulent premixed flames. In a large-scale direct numerical simulation, the use of the skeletal mechanism reduces the memory requirements by more than a factor of 3 and accelerates the simulation by a factor of 7 compared with the detailed mechanism. The skeletal mechanism is suitable for unsteady three-dimensional simulations of methane turbulent premixed, non-premixed, and globally lean partially premixed flames and is available as supplementary material.
1 April 2004 in International Journal of Heat and Mass Transfer