The thermo chemical state of a single-phase reacting fluid having Ns species is uniquely determined by a point in a Ns+1 dimensional space (T, p, and Ns-1 species mass fractions). However, it is widely recognized that low-dimensional manifolds determined by the slow chemical reactions exist in this space. This provides the opportunity to re-parameterize the thermo chemical state of a system with a smaller number of variables, the reaction variables. For a reduced model to be effective, these variables must provide a good lower dimensional approximation of the original thermo chemical state of the system.

In the present paper, a new method for the estimation of the dimensionality of the thermo chemical state of a reacting system and for the selection of “optimal” reaction variables is proposed. Two, high fidelity, experimental data sets, corresponding to a CO/H2/N2 turbulent jet flame and a piloted CH4/Air flame, have been investigated. Principal Component Analysis (PCA) has been employed through an eigenvalue analysis of the data sets defined by instantaneous spatial snapshots of the species' concentrations and temperature.

Results show that, for both flames, PCA is able to achieve a significant size reduction, although a larger number of “optimal” variables are required for the CH4/Air flame due to its greater complexity. Moreover, the influence of proper data preprocessing strategies on the number of components needed to properly reconstruct the original data is assessed.