Proper orthogonal decomposition analysis of autoignition simulation data of nonhomogeneous hydrogen–air mixtures

The proper orthogonal decomposition (POD) method is implemented on unsteady 2D direct numerical simulation of autoignition in nonhomogeneous hydrogen–air mixtures. The analysis is implemented to evaluate requirements for the reproduction of transient, multidimensional and multiscalar processes in combustion. Data reduction is implemented on a set of 30 snapshots of 2D fields of a passive scalar, the mixture fraction, and a reactive scalar, the mass fraction of the intermediate species, HO2. The snapshots cover the evolution of the hydrogen–air mixture from induction to the early stages of high-temperature combustion. The standard method by which the POD technique is measured, the cumulative energy criterion, based on the sum of the largest eigenvalues, suggests that the bulk of this energy may be represented by the first three to four modes for the reactive scalars. However, this criterion may not be sufficient to characterize the performance of the POD reduction approach. Therefore the number of required eigenmodes for each data set is tested. A number of preprocessing strategies of the scalar fields are explored to reduce the number of required eigenmodes. The strategies are designed to reduce the temporal and spatial spans of scalar values. The results show that different preprocessing strategies may yield different outcomes for the passive scalars, represented by the mixture fraction, and reactive scalars, represented by the intermediate species, HO2 mass fraction. More importantly, there are different requirements to reproduce passive and reactive scalars during the autoignition process. The mixture fraction, which is affected by the mixing process only, requires the least number of eigenmodes, and yields a sufficient representation of the original data with only two to three eigenmodes. The reactive scalar reduction improves significantly with preprocessing, which reduces the required number of eigenmodes to approximately six.