Laminar burning velocity measurements of iso-octane + air mixtures at higher unburnt mixture temperatures

Highlights

• Burning velocity of iso-octane + air mixtures measured at elevated temperatures.

• Burning velocity and temperature exponent reported for equivalence range (0.6–1.4).

• Reaction pathway study compares burning velocities at 470 and 640 K temperatures.

• Role of CH₃ radical in auto-ignition investigated for slightly rich mixtures.

Abstract

Surrogate fuels offer an effective and efficient alternative in predicting the combustion characteristics of various practical fuels such as gasoline, kerosene and diesel. Once standardized, the chosen constituents of the surrogate fuels help reduce the complexity involved in the development of detailed kinetic models for engine simulations at global scales. In this study, with air being the oxidizer, iso-octane as a surrogate constituent of the gasoline fuel is used for measuring the laminar burning velocities at high mixture temperatures. Present investigations report LBV measurements for a temperature range of 300–640 K using an externally heated diverging channel (EHDC) method. The comparison is then carried out with the available experimental data in the literature and kinetic model predictions at 1 atm pressure for an equivalence ratio range of ϕ = 0.7–1.4. Temperature exponent, (α) is derived using the power-law correlation and a good consistency is observed between the current measurements and the kinetic model predictions of LLNL-V3 (2008) and Hasse (2000) for a temperature ratio of 1.7 and 1.8 respectively. Reaction pathway analysis exhibits that the presence of species TC₃H₆CHO at 470 K leads to a lower LBV than at 640 K. Methyl radical, CH₃ plays a significant role in the auto-ignition phenomena as highlighted from the sensitivity analysis at ϕ = 1.1. The present study also becomes vital in terms of improving the detailed kinetic models targeting the prediction of knocking propensity.

Introduction

Fundamental knowledge of combustion characteristics of a commercially available fuel holds the key in improving the performance with minimal chemical emissions in the design of a combustion device. Operating on gasoline, the design, and optimization procedure for Spark Ignition (S.I.) engines call for more simplification on its combustion phenomena. Inherently being composed of several hydrocarbons makes the experimental and computational investigations extremely difficult to highlight its oxidation chemistry more vividly. When oxidized, such large molecules undergo subsequent breakdown into much simpler hydrocarbons (SHC), and determining its combustion chemistry to develop a detailed kinetic model is simpler to investigate and comprehend. Mixtures of such SHC is termed as surrogate fuels.

The chemical constituents of gasoline exhibits compositional variation, on account of several factors such as parent crude oil, seasonal variation, strict regulations, distillation process, etc. Developed surrogate fuels guarantee strict regulations for the chemical composition to retain its accuracy while validating a chemical kinetic model for engine simulations and development. Surrogate fuels can either be a single or a multi-component fuel based on the targeted chemicophysical properties [1] as shown in studies [2], [3], [4]. Surrogate fuels categorize its compound class composition in terms of iso-paraffins, normal paraffins, single-ring aromatics, cyclo-paraffins, olefinic species, and multi-ring aromatics. Iso-octane represents the iso-paraffins of the surrogates [5]. Iso-octane along with n-heptane is considered to be one of the primary reference fuels (PRFs) as they have enough potential to simulate the gasoline reactivity in addition to the octane rating [1], and also for S.I engine investigations [6], [7]. Iso-octane is also capable of assessing fuel volatility [8]. In the aviation sector, it was used in the development of JP-8 surrogate model [9]. For a surrogate fuel to emulate the combustion properties of any practical engine fuel, laminar burning velocity (LBV) plays the role of a matching parameter, which is globally accepted fundamental combustion property to validate a detailed reaction mechanisms [10], [11], [12]. The validation is then followed by simulating the combustion phenomena to optimize the practical combustion systems, which is ostensibly the motive behind the surrogate model development. This surrogate fuel-based model also finds its applications in the homogenous charge compression ignition (HCCI) engines among others [13]. For HCCI and other practical engines, high temperature LBV measurements of a surrogate fuel becomes vital on considering the air temperature after compression process. Since the available data in literature lacks measurements at elevated temperatures (as will be discussed later), this calls for experimental investigations in developing and validating kinetic model at high temperatures.

Noteworthy efforts towards the measurement of the laminar burning velocity of iso-octane + air mixture started towards the end of the 20th century. In 1998, Bradley et al. [14] used a linear-extrapolation technique to obtain the stretch-free laminar burning velocity of the primary reference fuels (PRF). In 2004, Huang et al. [15] performed measurements using a counterflow (CF) burner configuration and digital particle image velocimetry (DPIV). Linear extrapolation was used to report the LBV of PRF. For iso-octane, the measured values were consistently higher than the earlier reported data by Davis et al. [16], using the same technique. A study by Stanglmaier et al. [17] highlighted the fact that the LBV of gasoline and iso-octane are highly susceptible to inconsistencies when they are measured at higher temperatures and pressures. From the investigations of Farrell et al. [18] and Cruz et al. [19] the presence of aromatics in gasoline resulted in different LBV values as compared to the values of PRF mixtures and led to significant errors in flame propagation simulations. A comparison of the measured values by Kumar et al. [20] (with stagnation flame method) with those obtained using the heat-flux (HF) method [21], [22], [23] showed that the CF burner values were relatively higher than HF measurements. This was due to the linear-extrapolation technique used for stretch-correction by Kumar et al. [20]. In 2009, Jerzembeck et al. [24] reported LBV of iso-octane, commercial gasoline + air, and a PRF mixture with a research octane number of 87 using spherically expanding flame (SEF) method for a range of equivalence ratios, ϕ = 0.7–1.2. The measured data with unburnt mixture temperature of 373 K pressurized between 10 and 25 bar was used to validate the reduced kinetic mechanism developed. Though the predictions were consistent with the experimental data for gasoline + air mixtures at lean conditions, a disparity was observed for stoichiometric and rich mixtures at all pressure conditions, which stipulates for further examination of both the measurements and kinetic models.

Kelley et al. [25] measured laminar flame speeds of iso-octane + air mixtures using SEF and CF methods. The reported measurements were consistent at 1 atm, however, considerably lower values were reported at higher pressures, when compared with other measurements. Linear and non-linear extrapolation techniques were used in CF and SEF respectively to report unstretched laminar flame speeds. Galmiche et al. [26] emphasized that the improvement of LBV of iso-octane + air mixtures primarily due to the application of different experimental techniques. Measurements of iso-octane and n-heptane reported by Dirrenberger et al. [23] were consistently lower than the reported values of van Lipzig et al. [21], all measurements using the HF method. This prompted Sileghem et al. [22] to perform experiments using the heat-flux method with minor improvements in the experimental setup to determine the laminar burning velocities of gasoline fuel and its surrogates. The measurements were performed at atmospheric pressure with initial temperatures of 298 K and 358 K for equivalence ratios between 0.7 and 1.3. The measured values showed a consistent agreement with the data reported using the heat-flux method by Dirrenberger et al. [23] in addition to the data of Zhao et al. [27] using stagnation flame method and particle image velocimetry. When compared with the predictions of other kinetic models, no model was found to be good enough to show consistency with the measured values of iso-octane, suggesting further improvement in the kinetic models. Measurements reported by Dirrenberger et al. [23] for commercial gasoline and a ternary mixture of n-heptane + iso-octane + toluene were in good match with the reaction mechanism predictions.

Mannaa et al. [28] reported the laminar burning velocity measurements for various gasoline surrogates and observed that the gasoline surrogates were consistent only for lean and stoichiometric mixtures and highly inconsistent for rich mixtures with the predictions of various kinetic models. In addition to this, due to the formation of cellular flames for highly rich mixtures, the spread of the reported values becomes wider. Recently, Meng et al. [29] measured the LBV of iso-octane, toluene, 1-hexene, ethanol, and the quaternary blends with ethanol using a constant-volume bomb method and Schlieren imaging for temperatures and pressures up to 450 K and 4 bar respectively. A linear correlation was employed to obtain the stretch-free laminar flame speeds. In a recent publication, Han et al. [30] measured the LBV of n-heptane and iso-octane + air mixtures for a temperature range of 298 – 358 K. At atmospheric pressure with ϕ = 0.7–1.6, different approaches for flame cellularity reduction were employed to obtain the LBV measurements of very rich flames. A good match was observed with the kinetic model predictions. Apart from these measurements, for a temperature range of 298 – 700 K, Metghalchi and Keck [31] reported the burning velocities of iso-octane + air using constant volume bomb method. The measurements were performed for a range of equivalence ratios, ϕ = 0.8–1.5 over a pressure range, 0.4–50 atm. Errors due to buoyancy and flame wrinkling were neglected. Based on all the measurements, it is clear that the selection of the extrapolation technique remains uncertain for measurements using the same technique, which needs to be further explored. Some of the important experimental studies are summarised in Table 1. Other concern revolves around the development of a kinetic model focussing on suitable enhancements in the models, and ensure their consistency in better predictions for rich mixtures. Increased inconsistency between the predictions and measured values due to a disparity in the temperature exponent (α) becomes more prominent at higher mixture temperatures (close to engine relevant conditions) [32]. Two iso-octane + air kinetic mechanisms are used in the present analysis: Hasse [33], 29 species and 48 reactions; and LLNL V3 [34], 1550 species and 8000 reactions.

From literature, the associated discrepancies in the measurements motivated the authors to obtain accurate measurements and simultaneously compare the performance of various reaction mechanisms at all mixture conditions [32]. Especially at higher temperatures (>400 K), where discrepancies are relatively higher [32]. The measurements available in the literature (after the work of Bradley [14] (1998)) are presented and compared with the current measurements using externally heated diverging channel method for iso-octane + air mixtures. This work lists out the various data of laminar burning velocities of a mono-component gasoline surrogate (iso-octane) for temperatures up to 640 K using an externally heated diverging channel (EHDC) method. LBV of several liquid fuels at higher mixture temperatures has been reported in the literature [39], [40], [41], [42], [43], [44], [45] using this method earlier.