High-speed imaging of OH* and soot temperature and concentration in a stratified-charge direct-injection gasoline engine

https://doi.org/10.1016/j.proci.2004.08.021Get rights and content

Abstract

In-cylinder soot formation and oxidation in a stratified-charge spark-ignited direct-injection (SIDI) engine have been studied with a high-speed (9000 frames/s) camera system with three intensified detectors. Continuous records of soot temperature and relative soot concentration within individual engine cycles are evaluated from images of soot radiation at two wavelengths (650 and 750 nm). Combustion is followed simultaneously by imaging OH* chemiluminescence (306 nm). The spatially and temporally resolved data allow the two most important soot sources to be quantified separately for a typical part-load operating condition. (1) Soot first forms in partially premixed flame propagation through rich zones, but the high soot temperatures (∼2000–2400 K) and rapid mixing with surrounding hot lean regions lead to rapid soot oxidation and burnout. (2) Soot formation in pool fires (diffusion flames fed by thin films of liquid fuel on the piston) begins later and continues until late in the cycle, when further significant soot oxidation is unlikely (soot temperatures have dropped to ∼1500 K, and OH* chemiluminescence is no longer detectable). For this SIDI engine design and operating condition, pool fires are the dominant source of engine-out soot.

Introduction

Particulate (soot) emissions are of concern for many practical combustion devices, including internal-combustion (IC) engines [1] and gas turbines [2]. The soot emitted from a combustion device results from the competition of soot formation and oxidation processes, which depend on the local stoichiometry, temperature, pressure, and mixing. In situ diagnostics for soot formation and oxidation are therefore valuable in understanding the fundamental processes, in modeling combustion devices, and in controlling soot emissions.

Two-color pyrometry has long been used to evaluate line-of-sight-averaged (LOSA) soot temperature and relative concentration from the intensity of thermal radiation from soot at two wavelengths [3], [4]. The technique has been previously employed in IC engines (mostly diesels) either for single-shot quantitative imaging [5], [6] or for continuous time-resolved measurements, integrated either along a narrow path [7] or over a large portion of the combustion chamber [8], [9], [10], [11]. A few previous applications of high-speed imaging to two-color pyrometry have used photographic film or a streak camera [4].

This paper describes two-color pyrometry using high-speed (9000 frames/s) image-intensified digital cameras, together with in situ calibration and automated image processing. The digital detectors are much easier to control and calibrate than photographic film. Image intensification facilitates measurement of low soot concentrations and temperatures. This approach provides a continuous series of two-dimensional images of LOSA soot temperature and relative concentration for many engine cycles, permitting analysis of soot formation and oxidation on a cycle-by-cycle basis. This is crucial because engine combustion in general, and these processes in particular, varies significantly from cycle to cycle.

The application here is to a spark-ignited (SI) direct-injection (DI) engine of contemporary design. SIDI engines offer a significant fuel-economy improvement over conventional premixed-charge SI engines by operating nearly unthrottled at part load [12]. At light-load, the overall fuel–air equivalence ratio is often ∼0.25–0.35, which is so lean that the fuel must be stratified or concentrated around the spark gap at the time of ignition. This presents significant challenges in terms of ignition stability [13], gas-phase emissions (unburned hydrocarbons and NOx), and particulate emissions due to locally rich, sooting combustion (although SIDI engine-out soot emissions are typically much lower than those from a diesel engine of comparable output).

In the “wall-guided” [12] type of SIDI engine studied here, fuel–air mixture preparation depends critically on the interaction of the fuel spray with a contoured combustion bowl in the piston. Two primary soot sources have been identified [14], [15], [16], [17], [18]: (1) rich partially premixed combustion in the bulk gases and (2) pool fires (diffusion flames) fed by a thin film of liquid fuel deposited by the spray on the piston. Recent measurements have demonstrated that soot emissions from the engine studied here strongly correlate with the mass of fuel on the piston [19]. In this study, high-speed two-color pyrometry, together with simultaneous OH* chemiluminescence imaging, shows clear and practically important differences between the pool-fire and partially premixed-combustion modes of soot formation and oxidation.

Section snippets

Engine

Figure 1 shows the apparatus. Fuel is injected obliquely into the cylinder of the four-valve SIDI engine (86-mm bore and stroke, compression ratio 10.3) using a high-pressure (11 MPa) six-hole injector mounted between the intake valves. The six spray plumes are evenly spaced with a 50° angle between opposite plumes. For stratified operation, fuel is injected shortly before ignition; specifically, the end-of-injection (EOI) command was at crank angle 50°BTDC, and the spark was triggered at

Results

Figure 6 follows the evolution of the soot temperature, KL distribution, and combustion in one engine cycle from shortly after ignition until shortly before the disappearance of soot luminosity. Figure 7 presents a more global picture in which the spatially averaged soot temperature and spatially integrated KL factor (excluding invalid or suspect values) have been ensemble averaged over 12 cycles. Because the pool fires can be distinguished both spatially and temporally from the earlier rich

Discussion

The literature contains a wide range of temperatures for soot formation (∼1000–2800 K [1], [25]) and oxidation (∼1400–1900 K [24], [25], [26]). Temperatures from diesel engines may be biased high because many of the measurements used two-color pyrometry [4], [7], [9], [21]. The ability illustrated here to follow the spatial distributions of OH* chemiluminescence, soot temperature, and KL factor continuously in time has proven very helpful in interpreting the experimental results, although caution

Conclusions

High-speed two-color pyrometry and simultaneous OH* imaging have provided a much more detailed picture of soot formation and oxidation in a wall-guided SIDI engine than has been available previously. The two principal soot sources for warmed-up stratified-charge operation have been quantified. (1) Soot is formed first as a partially premixed flame propagates through locally rich zones, but this soot burns out rapidly due to its high temperature (∼2000–2400 K) and rapid mixing with surrounding

Acknowledgments

M.E. Rosalik provided outstanding project engineering support in all phases of these experiments. B.D. Stojkovic’s doctoral work was supported in part by DOE PAIR Grant DE-FG02-98-ER14915.

References (32)

  • H. Zhao et al.

    Prog. Energy Combust. Sci.

    (1998)
  • F. Zhao et al.

    Prog. Energy Combust. Sci.

    (1999)
  • H.N. Najm et al.

    Combust. Flame

    (1998)
  • J.B. Heywood

    Internal Combustion Engine Fundamentals

    (1988)
  • A.H. Lefebvre

    Gas Turbine Combustion

    (1983)
  • H.C. Hottel et al.

    Ind. Eng. Chem. Anal. Ed.

    (1932)
  • J.S. Shakal, J.K. Martin, SAE Paper 940903,...
  • C. Arcoumanis, C. Bae, A. Nagwaney, J.H. Whitelaw, SAE Paper 950850,...
  • R.C. Peterson, K.-J. Wu, SAE Paper 861565,...
  • J. Yan, G.L. Borman, SAE Paper 891901,...
  • X. Li, J.S. Wallace, SAE Paper 950848,...
  • U. Spicher, A. Kölmel, H. Kubach, G. Töpfer, SAE Paper 2000-01-0649,...
  • K. Schänzlin, T. Koch, K. Boulouchos, SAE Paper 2002-01-0834,...
  • T.D. Fansler et al.

    Int. J. Engine Res.

    (2003)
  • E. Stevens, R. Steeper, SAE Paper 2001-01-1203,...
  • T. Alger, Y. Huang, M. Hall, R.D. Matthews, SAE Paper 2001-01-1204,...
  • Cited by (138)

    View all citing articles on Scopus
    View full text