Analysis and simplification of chemical kinetics mechanisms with CSP-based techniques

The computational singular perturbation (CSP) method is exploited to build a comprehensive framework for analysis and simplification of chemical kinetic models. The necessity for both smart post-process tools, able to perform rational diagnostics on large numerical simulations of reactive flows, and affordable reduced kinetic mechanisms, to make the simulations feasible, is the driving force behind this work. The ultimate goal is to improve the understanding of the fundamentals of chemically reacting flows. The CSP method is a suitable candidate for extracting physical insights from reactive flows dynamics that can be employed for both the generation of simplified kinetic schemes and the calculation of smart and compact diagnostic observables. Among them, the tangential stretching rate (TSR) is an estimate of the system’s driving chemical timescale that can be profitably employed for characterising the reactive flow dynamics in terms of combustion regimes and role of transport with respect to kinetics. The potentials of TSR are extensively highlighted, starting from prototypical combustion models, such as batch reactor and unsteady laminar flamelet, and getting to real-life usage on 3-dimensional direct numerical simulation datasets. The CSP mathematical foundations are then employed for mechanism simplification purposes, where small and accurate kinetic mechanisms are sought after. An existing CSP-based simplification algorithm is improved, aiming at the minimisation of the required user knowledge, which becomes a critical feature of the algorithm when dealing with new fuels. Practical applications of the revised algorithm are shown and discussed. Finally, the focus is shifted from the quest for tight accuracy in the simplified mechanisms towards a much broader question regarding confidence in detailed kinetic schemes. Uncertainty in the kinetic model parameters, such as Arrhenius coefficients, can jeopardize the efforts spent in the reduction challenge. A new, uncertainty-aware, robust CSP simplification strategy is proposed, discussed and employed, and its robustness demonstrated in a test case involving an uncertain -in its Arrhenius pre-exponential coefficients- kinetic scheme.