Symbolic analysis of the cycle-to-cycle variability of a gasoline–hydrogen fueled spark engine model.

Abstract

An study of temporal organization of the cycle-to-cycle variability (CCV) in spark ignition engines fueled with gasoline–hydrogen blends is presented. First, long time series are generated by means of a quasi-dimensional model incorporating the key chemical and physical components, leading to variability in the time evolution of energetic functions. The alterations in the combustion process, for instance the composition of reactants, may lead to quantitative changes in the time evolution of the main engine variables. It has been observed that the presence of hydrogen in the fuel mixture leads to an increased laminar flame speed, with a corresponding decrease in CCV dispersion. Here, the effects of different hydrogen concentrations in the fuel are considered. First, it is observed that return maps of heat release sequences exhibit different patterns for different hydrogen concentrations and fuel–air ratios. Second, a symbolic analysis is used to characterize time series. The symbolic method is based on the probability of occurrence of consecutive states (a word) in a symbolic sequence histogram (SSH). Modified Shannon entropy is computed in order to determine the adequate word length. Results reveal the presence of non-random patterns in the sequences and soft transitions between states. Moreover, the general behavior of CCV simulations results and three types of synthetic noises: white, log-normal, and a noisy logistic map, are compared. This analysis reveals that the non-random features observed in heat release sequences are quite different from synthetic noises.