Development of an N-Heptane/Toluene/PAH Mechanism and Its Application for Soot Prediction

Development of an n-heptane/toluene/PAH Mechanism and its Application for Soot Prediction

Hu Wang1, 2, Qi Jiao2, Mingfa Yao1, Rolf D. Reitz2

1. Engine Research Center, University of Wisconsin-Madison, Madison, Wisconsin, 53705, U.S.A

1

2. State Key Laboratory of Engines, Tianjin University, Tianjin, 300072, P. R. China

Abstract

A chemical reaction mechanismhas beendeveloped for modeling the combustion process and poly-aromatic hydrocarbon (PAH) formation of diesel and n-heptane/toluene fuels. The mechanism consists of 71 species and 360 reactions. This mechanismwasextensively validated with available experimental data. Apractical multi-step soot model was integrated with the PAH kinetics modelto predict soot emissionsofdiesel and n-heptane/tolueneDIengine data.Constant volume combustion vessel simulations were also conducted. The results show that the present mechanism provides promising agreement in term of PAH prediction for various fuels in premixed flames. HCCI combustion and DI spray combustion simulation results confirm that the present mechanism gives reliable predictions of combustion and soot emissions for both diesel and n-heptane/toluene fuels under various conditions. The study also highlights the importance of aromatics on the PAH formation and soot emissions.

1. Introduction

Diesel and gasoline fuels are complex mixtures of paraffins, aromatics, cycloalkanes and olefins[1]. A significant amount of aromatic compounds is present in both gasoline (~35%) and diesel fuels (~30%). Toluene is a major aromatic component in practical fuels and also a desired component for diesel and gasoline surrogates. It is one of the simplest derivatives of benzene, and its chemistry is a foundation for large aromatic formation. Therefore, many kinetic modeling studies of toluene and TRF fuels have beenconducted. Davidson et al.[2], Shen et al.[3]and Herzler et al. [4]presented measurements of ignition delay of toluene/air and toluene/n-heptane/air mixtures at engine relevant conditions and different equivalence ratios. Bounaceur et al.[5]and Andrae et al. [6] proposed detailed toluene and TRF kinetic mechanisms. Based on the toluene mechanisms developed by Bounaceur et al.[5] and Andrae et al.[6], Ra et al.[7] developed a reduced toluene mechanism and incorporated it into a multi-chemistry mechanism.

It is clear that detailed and accurate PAH growth mechanisms are necessary if accurate simulation of soot inception is to be realized.Xi et al.[8]proposed a reduced chemical mechanism for the chemical kinetics of n-heptane oxidation in modeling PAH formation in dieselcombustion.Vishwanathan et al.[9] combined this reduced PAH mechanism with the PRF mechanism developed by Ra et al.[10], and coupled it with a multi-step soot model to predict the soot emissions of diesel fuel. Recently Slavinskaya et al. [11]developed a detailed PAH mechanism for methane, ethane and ethylene flames. The proposed PAH mechanism is highly successful at predicting concentrations of aromatics containing two to four rings.

The presence of aromatics promotes the formation of PAHand soot particles as benzene (A1) is one of the major species in both the toluene oxidation and PAH growth processes.Therefore, an accurate PAH formation model is required for better soot prediction, especially for fuels composed of aromatic compounds.Choi et al.[12]found that PAH initially increased and then decreased with the toluene ratio, exhibiting a synergistic effect.The soot formation increased monotonically with the toluene ratio; however the effect of toluene on sootformation was minimal for relatively small toluene ratios.

The objective of the present study wasto developa reduced n-heptane/toluene/PAH mechanism which can be applied tomultidimensional CFD simulations forcombustion and soot predictions of diesel, n-heptane and n-heptane/toluene fuels.

2. Mechanism formulation

The current investigation is based on the n-heptane-n-butanol-PAH mechanism of Wang et al.[13], wherein a reduced mechanism was developed to predict the combustion and soot emissions of both non-oxygenated and oxygenated fuels. The base n-heptane-PAH mechanisms were used in this study. However, updates were made to improve the PAH predictions in ethylene and n-heptane premixed flames and the ignition delays andbenzene predictions of the reduced toluene mechanism developed by Ra et al. [7].In addition, an n-heptane/toluene/PAH mechanism was formulated based on these mechanisms.

Several reactions were adjusted or added in this n-heptane mechanism to improve itsprediction of C2H2 concentrations and ignition delays under low oxygen concentrations, viz.

The pre-exponential factor of theforward rate of R1 was reduced by a factor of 15 to improve both the C2H2and PAH predictions. The original higher reaction rate constant of this reaction converts most of the C2H4 to CH2O, and this dramaticallyreduces the building block pool for PAH formation, and thusgreatly affects the formation of PAH molecules. The rate constants of R2-R4 were taken from Ra et al.[7]. These three reactions can greatly improve the performance of the mechanism under low oxygen concentration conditions without penalty on the ignition delay prediction under high oxygen concentration conditions.

The reduced toluene mechanism developed by Ra et al.[7] was taken as the base toluene mechanism. Initial toluene premixed flames simulation results showed that the original toluene mechanism in this multi-chemistry mechanism over-predicts benzene (A1). Through rate of production and sensitive analysis, the following three reactions were found to be important to the production of benzene (A1), and were adjusted to improve the prediction of A1:

The pre-exponential factor of theforward rates of R4 and R6 were reduced by 10 times while R5 was increased by 10 times to reduce the production of A1 when toluene was applied. However, it is also necessary to adjust the following two reactions to match the ignition delays:

These two reactions are the most ignition-delay-sensitive reactions in this reduced toluene mechanism.Based on the present reduced toluene and n-heptane/PAH mechanisms, ann-heptane/toluene/PAH mechanismwas formulated by adding the toluene mechanism into the base n-heptane/PAH mechanism. The finaln-heptane/toluene/PAH mechanism consists of 71species and 360 reactions. The cross reactions between then-heptane and toluene mechanisms were also considered and taken from [14].

3. Validation of mechanism

3.1 Ignition delay timings

Ignition delay timings predicted by the present reaction mechanism were compared and validated using experimental data available from the literature. Figure 1 shows comparisons of predictions of ignition delay times between the present reaction mechanism and shock tube data at different equivalence ratios for n-heptane, toluene and toluene/n-heptane fuels.

Figure 1 Comparison of predicted constant volume ignition delay times of n-heptane/air, toluene/air and toluene/n-heptane/air between the present mechanism and shock tube test data. (a) n-heptane/air, phi=1.0; (b) n-heptane/air, phi=0.5; (c) toluene/air, phi=1.0; (d) toluene/air, phi=0.5; (e) toluene/n-heptane/air, phi=1.0; (d) toluene/n-heptane/air, phi=0.5.

It is seen that good agreement of ignition delay between the experimental data and the present reaction mechanism can be achieved over wide temperature ranges and at different equivalence ratios for n-heptane fuel. The predicted toluene ignition delays are generally in good agreement with the experimental data at the two different equivalence ratios. However, it can be seen that even under the same conditions, the experimental data sets are different from each other. A possible reason for this difference between these two sets of experimental data is the pre-ignition energyrelease caused by soot depositsin the shock tube, as discussed by Shen et al. [3]. The comparisons of predicted ignition delays of toluene/n-heptane/air mixtures at constant volume between the present mechanism and experimental measurements were also presented in Figure 1-(e) and (f). It is seen that although the ignition delays of pure n-heptane and toluene are well predicted by the current mechanism, when mixture of n-heptane and toluene is applied, the predictions are only roughly acceptable. However, it can be seen that the present mechanism captures the ignition delay trends for changes in temperature and pressures.

3.2 Premixed flames

This section presents comparisons between the simulation results obtained using the present mechanism and available experimental data in premixed flames. Table 1 provides the details of the premixed flames simulated in this work for the mechanism validation. The simulation was conducted with the CHEMKIN package, and temperature profiles from the experiments were used for the calculations.Figure 2 shows comparisons of the concentration profiles between experiments and simulations. It is seen that the reduced mechanism describes satisfactorily the concentrations of C2H2, C2H4, benzene, and toluene and PAH molecules in various flame test conditions.

Table 1 - Test conditions of premixed flames.

Figure 2 Prediction of benzene, C2H2, C2H4 and PAH in benzene premixed flames. (a),C2H4 flame,flame 1; (b) n-heptane flame, flame 2; (c) CH4/toluene flames, flame 4; (d) benzene flame, flame 5; (e) and (f) benzene flames, flame 3 and 6.

3.3 HCCI combustion

HCCI engine experiments were conducted and the experimental results were used for validation of the current mechanism. Mixtures of n-heptane and toluene, with liquid volume fractions of 80%/20% and 60%/40%, respectively, named TRF20 and TRF40, were used. The engine operating conditions are listed in Table 2.

The simulations were performed using the KIVA code coupled with the CHEMKIN for the chemistry calculation. A 45 degree sector mesh with a cell number of around 10,000was used, as shown in Figure 3-(a). The computationsstarted from intake valve closure (IVC) with anassumed uniform mixture distribution in the cylinder andended at exhaust valve open time.Figure 3-(b) and (c) show comparisons of the pressure profiles and apparent heat release rate (AHRR) between the experimental data and predictions obtained using the current n-heptane/toluene/PAHmechanism at EGR rates of 15% and 45%. It can be seen that the ignition timings are well predicted by the present mechanism. The in-cylinder peak pressures at lower EGR conditions for both TRF fuels are slightly over predicted while for higher EGR conditions, the agreements between simulation and experiments are better.

Table 2 - Engine operating conditions

Figure 3 Comparisons of predicted pressure trace and AHRR with experimental data in TRF fuelsHCCI combustion at different EGR rates. TRF20 fuel, EGR 15/45%, phi=0.38/0.48; TRF40 fuel, EGR 15/45%, phi=0.37/0.46.

3.4 DI spray combustion

DI diesel experiments were also conducted for the validation of the proposed mechanism. The engine operating conditions are shown in Table 3.In order to predict soot emissions, the practical phenomenological sootmodel developed by Vishwanathanet al.[9] was applied and coupled with the present mechanism. In this soot model, PAH is treated as the soot precursor species,and soot inception starts with pyrene (A4). The main processes related to the soot formation and oxidations processes and the reaction rates of these steps and further details about the soot model can be found in [9].

Table 3 - Engine operating conditions

Figure 4 Predicted in-cylinder pressure traces and AHRR profiles, NOx and soot emissions for diesel, TRF15and TRF30fuels

The predictions of in-cylinder pressure and AHRR of diesel, TRF15 and TRF30 fuels at EGR rates of 25.2% and 40.6% using the current mechanism are shown in Figure 4-(a) and (b). It can be seen that current mechanism can well predict the in-cylinder pressure traces, as well as AHRR with the different fuels at the different EGR conditions. Also, the proposed mechanisms capture the difference between different fuels well. Figure 4-(c) and (d) show comparisons of NOx and soot emissions between the experimental data and simulation results. It can be seen that NOx emissions are predicted well by the current mechanism. As for soot prediction, the general trend is also captured quite well by the current mechanism. The difference in soot emissions among the three fuels is well captured by the present mechanism. It can be seen that, for the two TRF fuels, the one with higher toluene content actually produces the lowest soot emissions in all EGR conditions in both experiments and simulations. This can be attributed to the slightly longer ignition delay of TRF30 fuel, which is helpful to improve the fuel-air mixing process. This also indicates a higher aromatic in the fuel does not necessarily produce higher soot.

3.5 Constant Volume Combustion Chamber Simulations

The present mechanism was also applied to simulate Sandia ECN constant volume chamber data [21]. The purpose of the simulations was to further explore the effects of the aromatic content of the fuel on sooting propensity. A 15% ambient oxygen concentration, 30.0 kg/m3 ambient density case was chosen for the simulations, in which n-heptane was progressively replaced with toluene (20% steps of mass fraction). The same CFD code, related sub models and soot model used in the above DI engine simulations were used. Two different sets of tests were conducted. In the first set, the physical properties of the blended fuels were represented with those ofn-heptane to highlight the effects of thearomatic chemical properties on mixing and soot emissions. In the second set, boththe physical properties and chemical reaction mechanisms were applied for each fuel. Through these comparisons, the effects of both chemical and physical properties of toluene could be evaluated.

Fig.5 Predicted distributions of soot volume fraction, equivalence ratio, A4 concentration and temperature for the two test sets.(a) The first set; (b) the second set. From left to right: pure n-heptane, 80% n-heptane + 20% toluene, 60% n-heptane + 40% toluene and 40% n-heptane + 60% toluene fuels. From up to bottom: soot volume fraction, equivalence ratio, A4 mass fraction and temperature. Soot volume fraction is averaged between 2.5 to 5.5 ms ASI; Equivalence ratio, A4 concentration and temperature at 5.0 msASI.

Figure 5-(a) shows the predicted distributions of soot volume fraction, equivalence ratio, A4 mass fraction and temperature for the first set. It is seen that the lift-off length increases as the toluene content increases in the fuel, which is helpful to improve the fuel-air mixing. This is helpful to reduce the soot formation through improved mixing process. However, the predicted A4 mass fraction demonstrates a dramatic increase due to the introduction of toluene. Thus, on the one hand, the longer ignition delay and lift-off length hinders the formations of PAH and soot and also enhances the soot oxidation. But, on the other hand, the aromatic greatly promotes the PAH and sootformation. Therefore, the final soot reflects the competitiveoutcomeof these two processes.

These results clearly show the importance of the mixing process and the effects of the fuel on soot emissions. Figure 5-(b) shows the predicted distributions for the second set, in which the real physical properties of n-heptane and toluene were applied. It is seen that the general trend of this set is quite similar to that of the first set. The soot volume fraction increases as the toluene content in the blended fuels increases, accompanied by a lower equivalence ratio, longer lift-off length, higher A4 concentration and similar flame structure and temperature distributions.

Through the comparison of these two sets of simulations, it can be concluded that it is due to toluene’s physical property effects on the spray development, evaporation and mixing that the improvement in equivalence ratio (mixing process) of the second set is not as good as the first set. This results in slightly higher soot volume fractions. Besides these relatively small differences, the overall distributions are quite similar.

4. Conclusion

A reduced chemical reaction mechanism has been developed for modeling the combustion process and PAH formation of diesel and n-heptane/toluene fuels. The final mechanism consists of 71 species and 360 reactions. The mechanism was extensively validated against available experimental data, and shows promising agreements with experimental results. In addition, a multi-step soot model was integrated with the mechanism to predict the soot emissions. The mechanism predicts the combustion processes of the different fuels quite well. The general combustion characteristics and NOx and soot emissions of the tested fuels are well captured under different EGR conditions. The overall results show that the present mechanism provides promising agreements in terms of PAH prediction for various fuels in premixed flames and highlights the importance of aromatics on PAH formation and soot emissions.