In-cylinder unburned hydrocarbon visualization during low-temperature compression-ignition engine combustion using formaldehyde PLIF

https://doi.org/10.1016/j.proci.2006.07.044Get rights and content

Abstract

Formaldehyde (H2CO) is visualized by planar laser-induced fluorescence (PLIF) in a heavy-duty direct-injection diesel engine to better understand the sources of unburned hydrocarbon (UHC) emissions. H2CO is used as a tracer for UHC based on chemical kinetic simulations, which show that after the first-stage of ignition, the evolution of H2CO is very similar to that of UHC in the bulk gases. Modern low-temperature, low-nitrogen oxides (NOx) combustion conditions are achieved by diluting the intake stream with nitrogen to 12.7% oxygen by volume, simulating exhaust gas recirculation. A range of ignition delays (IDs) is produced by adjusting the intake temperature and/or the fuel injection timing. Spectral measurements of the laser-induced emission after 355-nm excitation show that broadband emission from sources other than H2CO can be significant, so the PLIF images must be interpreted with care. A frequency-domain covariance analysis is offered as a way to quantify the contribution of H2CO fluorescence in the images. Longer ID conditions are found to have both higher-UHC emissions and stronger H2CO fluorescence persisting late in the cycle, especially near the injector. Based on kinetic model predictions of longer H2CO lifetimes in leaner regions, it is concluded that for long ID conditions, mixtures near the injector after the end of injection are too lean to achieve complete combustion, thus contributing to UHC emissions.

Introduction

Some combustion strategies have emerged to face the challenge of reducing both particle matter (PM) and nitrogen oxides (NOx) from diesel engines, based on new low-temperature combustion (LTC), such as homogeneous charge compression-ignition (HCCI) [1], late injection with exhaust gas recirculation (EGR) and high swirl (e.g., MK combustion) [2], and hybrid double-injection schemes (e.g., UNIBUS) [3]. In general, these strategies use very early or very late fuel injection and/or high EGR rates to extend the ignition delay (ID), which is the time between the start of injection (SOI) and the start of combustion. This can enhance pre-combustion mixing so that both PM and NOx are reduced, but other emissions, including unburned hydrocarbons (UHC), become a concern. One source of UHC is incomplete bulk-gas reactions, which can be quite significant, especially in HCCI engines at low load, for which mixtures are very lean [4], [5]. With long mixing times in LTC diesel engines, it is likely that similar very lean, over-mixed regions also exist, and may contribute significantly to UHC emissions. Currently, however, the importance and evolution of this in-cylinder source of UHC emissions in LTC diesel engines is not well understood.

No suitable planar laser/imaging diagnostic yet exists for direct quantification of UHC in complex mixtures of unburned fuel and combustion intermediates in LTC diesel engines. Diesel-like fuels display a second-stage ignition behavior. Combustion intermediates, including formaldehyde (H2CO), are formed during the first-stage, and are consumed later in the second-stage. H2CO is easily accessible by planar laser-induced fluorescence (PLIF). H2CO might be a suitable, naturally occurring tracer for in-cylinder UHC in lean, over-mixed regions that achieve the first (low-temperature) stage of ignition, but have not yet transitioned to the high-temperature chemistry. Indeed, images of H2CO PLIF in HCCI-type engines have been shown to correlate well with fuel tracer PLIF images during the period between low- and high-temperature ignition [6]. While these observations are encouraging, it is unclear whether the single-component fuel tracer is representative of the multi-component hydrocarbon mixture in partially ignited lean regions.

To better understand the suitability of H2CO as a tracer for the “soup” of UHC after the low-temperature ignition, we simulated the ignition chemistry using a detailed chemical kinetic model for n-heptane [7], [8]. This model has been validated over a wide range of conditions, and has shown success in predicting ignition in HCCI engine simulations [5]. Using the Senkin module of CHEMKIN III [9], we simulated the constant-pressure adiabatic ignition chemistry of a lean mixture with an equivalence ratio (Φ) of 0.5, initially at 770 K and 46.5 bar, which are typical in-cylinder conditions prior to ignition for this study. To simulate EGR, the air was diluted with nitrogen (N2) to 12.7% oxygen, by volume. The simulation results in Fig. 1 show that after an induction period of roughly 10 engine crank-angle degrees (CAD) at 1200 rotations per minute, the temperature rapidly increases by ∼100 K during the first-stage low-temperature ignition when much of the parent fuel molecule is consumed and H2CO is formed. Then, the temperature increases more slowly over the next 30 CAD until the temperature reaches about 1000 K, when the second-stage hot ignition reactions commence and the temperature again rises more rapidly. Also plotted in Fig. 1 is the mass fraction of carbon (C) that is not in either CO or CO2, i.e., the carbon that is in UHC or oxygenated UHC. For demonstration, the H2CO curve has been rescaled so that it overlaps the C curve after the first-stage ignition. Figure 1 shows that although parent fuel molecule is mostly consumed during the first-stage ignition, much of the carbon is still bound in other UHC species. Therefore, the parent fuel molecule is actually a poor marker of UHC after the first-stage of ignition. The same may be true for some other potential tracer species (e.g., ketones) that are commonly doped in the fuel. By contrast, for these conditions, the H2CO and the C curve have a very similar evolution. Therefore, we deem H2CO a suitable and diagnostically convenient tracer for in-cylinder UHC arising from partially burned mixtures.

In this study, in-cylinder UHCs are studied in LTC diesel modes over a wide range of ID, achieved by adjusting the start of injection (SOI) and the intake temperature. In-cylinder H2CO is probed by PLIF as an indicator of in-cylinder UHC, and these observations are compared to trends in UHC measured in the engine exhaust. It also will be shown through spectral analysis that sources other than H2CO may contribute to UV fluorescence, so the fluorescence images must be interpreted with care.

Section snippets

Experimental set-up

The bore, stroke, and displacement of the optically accessible 4-stroke, single-cylinder direct-injection diesel engine (Fig. 2) are, respectively, 139 mm, 152 mm, and 2.34 L. It has a quiescent (low swirl) combustion chamber, and the geometric compression ratio is 10.75. The bowl diameter is 97.8 mm, and a section of the bowl-wall is cut out to enable laser-sheet access when the piston is near top dead center. The axes of the jets from the centrally located, 8-hole common-rail mini-sac injector

Experimental conditions

By adjusting SOI timing and/or the intake temperature, the ID was made to vary from 4.4 to 15.2 CAD as reported in Table 1. These conditions were intentionally designed to yield a wide range of exhaust UHC, as affected by ID. The intake air was diluted with nitrogen to approximately 12.7 vol.% oxygen to simulate EGR, and its temperature (Tin) and pressure (Pin) were adjusted to maintain constant intake density. Fluorescence interference from conventional diesel fuel was found to be

Formaldehyde detection and signal processing

Figure 3a is an ensemble average of 20 instantaneous UV fluorescence images from the ID = 5.4 CAD condition, acquired during the peak fluorescence intensity, at a 3° after top dead center (ATDC). The field of view in this image only shows the region around the fuel jet that is illuminated by the laser sheet. The fuel injector is located on the bottom side of the image, and residual elastic scattering from the liquid fuel spray of the eight jets is clearly distinguishable in the image. The curved

Conclusions

Chemical kinetic model results for dilute, low-temperature combustion conditions show that after the first-stage ignition, the evolution of H2CO is very similar to that for UHCs. H2CO is readily accessible by laser-induced fluorescence diagnostics, so it is a good tracer candidate for the complex post-ignition mixture of UHCs. Furthermore, the parent fuel molecule is actually a poor indicator of these UHCs.

For some engine operating conditions, significant broadband laser-induced emission was

Acknowledgments

This study was performed at the Combustion Research Facility, Sandia National Laboratories, Livermore, CA. Support for this research was provided by the US Department of Energy, Office of FreedomCAR and Vehicle Technologies. Sandia is a multi-program laboratory operated by Sandia Corporation, a Lockheed Martin Company for the United States Department of Energy’s National Nuclear Security Administration under contract DE-AC04-94AL85000. The authors thank Dave Cicone for assistance in operating

References (14)

  • H.J. Curran et al.

    Combust. Flame

    (1998)
  • C.K. Westbrook

    Proc. Combust. Inst.

    (2000)
  • C. Brackmann et al.

    Spectrochim. Acta A

    (2003)
  • A. Burkert et al.

    Proc. Combust. Inst.

    (2000)
  • F. Zhao et al.

    Homogeneous Charge Compression Igntion (HCCI) Engines

    (2003)
  • S. Kimura et al.

    SAE Trans.

    (1999)
  • R. Hasegawa et al.

    SAE Trans.

    (2003)
There are more references available in the full text version of this article.

Cited by (0)

View full text