Efficient slow manifold identification for tabulation based adaptive chemistry

We present a tabulation strategy for the numerical integration of chemical reacting flow processes on the basis of a reduced (non-stiff) model of the original set of ODEs. Both the tabulation and the identification of the reduced model are adaptive and are based on the Computational Singular Perturbation (CSP) method. The tabulation strategy is implemented in order to store and reuse the CSP quantities required for the construction of the reduced, adaptive model. We discuss a particular feature of this algorithm, the "homogeneous correction", that allows for the accurate and efficient identification of the (slow, invariant) manifold(s) within which the system dynamics evolve driven by slow time scales. The improved efficiency in constructing the slow model and computing the system dynamics along the slow manifold is demonstrated.