摘要：We have studied cyclic heat release variability in a spark-ignition engine under exhaust gas recirculation (EGR), using nitrogen to simulate EGR. Five EGR levels are examined. We used wavelet analysis to identify the dominant modes of fluctuation and how these modes vary in time. It is found that at a low EGR level, the heat release variations exhibit high-frequency intermittent oscillations. As the EGR level increases, the high-frequency oscillations tend to become more persistent, occurring continuously over many cycles. When the EGR level is sufficiently high, intermittent oscillations are observed at both high and low frequencies. In addition, persistent low-frequency fluctuations are present at the high EGR level. We have fitted theoretical probability models to the empirical heat release distributions. Depending on the EGR level, a three-parameter probability density function such as the generalized logistic distribution, a four-parameter distribution such as Johnson SB, or the five-parameter Wakeby distribution is found to provide a good fit. The goodness of fit of the theoretical distributions is assessed by the Kolmogorov-Smirnov (KS) test statistics. A good understanding of cyclic variability is essential to develop effective control strategies for efficient combustion.
