Kinetic analysis of the role of selective NO_x recirculation in reducing NO_x emissions from a hydrogen engine

Highlights

• NO_x reduction can be obtained even without an interaction between fuel and NO_x.

• NO injection results in decrease in further NO formation at high equivalence ratios.

• Thermodynamic limitations used to explain the trends for various amounts of fuel.

• NO conversion efficiency remains constant with NO injection for $0.65\le \mathrm{\Phi }\le 0.9$.

Abstract

The kinetics of NO formation is used to study the effect of NO injected into a hydrogen engine on its emission characteristics. Considering the engine to be a variable-volume reactor, it is shown that the injection of NO reduces its further formation. In contrast to the earlier studies, it is highlighted that a reduction in the amount of NO can be achieved even if the reactions between the fuel and NO do not take place. Additionally, the predicted reduction in NO is not due to the lowering of temperature, which is the focus of most of the exhaust gas recirculation studies. The NO conversion efficiency is predicted to be low for equivalence ratios in the range of 0.4–0.5. However, for higher equivalence ratios between 0.65 and 0.9, the model predicts NO conversion efficiencies above 90% for various amounts of NO injected. Consistent with the data reported in the literature, the model predicts that the NO reduction is a weak function of the engine speed and that the NO conversion efficiency is almost independent of the amount of NO injected for equivalence ratios greater than 0.6. A significant reduction in the NO emissions can be achieved, provided that the temperature generated is high enough for the NO formation reactions to approach equilibrium. It is deduced that conditions such as high equivalence ratio, high load, and high compression ratio will result in a high reduction of NO_x. These results can be used to design selective recirculation systems for an enhanced removal of NO_x. Even though H₂ is used as a fuel in the present work, the insights obtained are equally applicable to other fuels as well.

Introduction

The global consumption of marketed energy is projected to increase by 48% between 2012 and 2040 [1]. A major fraction of the energy is provided by fossil fuels, which are depleting fast due to their rapid consumption. Since the combustion of fossil fuels produces CO₂ which is considered to be a major greenhouse gas, the researchers are focusing on renewable and cleaner fuels. Hydrogen, due to its various characteristics such as low cost per unit energy produced, high auto-ignition temperature, high burning velocity, and low ignition energy, has caught the researchers’ attention [2]. H₂ is considered to be a clean fuel as its combustion generates only water and the formation of any carbon-based pollutants is minimal. However, because of the high temperatures generated within the engine, the thermal fixation of N₂ can result in the formation of NO_x.

Various modeling studies have been reported on the combustion of H₂ and the associated NO_x formation. Kinetic models consisting of various species and reactions are available in the literature [3], [4], [5], [6], [7], [8]. These kinetic models have been validated against experimental data from shock tubes, flow reactors, and/or laminar premixed flames. Additionally, some of the models have been used to study the combustion of H₂ in internal combustion engines [9], [10]. The formation of NO_x and the various mechanisms contributing to its formation have been studied by various researchers [11], [12], [13]. Some of the NO_x formation and reduction studies have been specifically performed with H₂ as the fuel [7], [14].

Various methods have been used to reduce the engine-out NO_x emissions some of which include injection of water, charge dilution, throttling, and modification of operating parameters such as spark timing and fuel injection timing [15], [16], [17], [18], [19]. Lee et al. [20] showed that the variation of valve timing and lean boosting can be used to achieve low amounts of NO_x. The reburning of NO_x using a fuel has been studied as a potential method for NO_x reduction [21], [22]. In order to elucidate the kinetics of the reburn chemistry, modeling as well as experimental studies have been performed [23], [24], [25]. Glarborg et al. [23] showed that NO catalyzes the removal of hydrogen atom under fuel-rich conditions and obtained a rate constant for the H + NO + N₂ reaction. Exhaust gas recirculation (EGR) is another NO_x reduction method which makes use of the combustion products for dilution and has been widely used for various fuels including gasoline and diesel. The combustion products from the previous engine cycle are recycled back and mixed with the incoming fresh fuel-air mixture in the present cycle. The recycled gases not only decrease the relative in-cylinder oxygen concentration but also increase the specific heat of the intake gaseous mixture [26]. This reduces the in-cylinder temperature which helps in the reduction of thermal NO_x.

Many researchers have used EGR for the reduction of NO_x emissions [26], [27], [28], [29], [30], [31], [32]. However, excessive use of EGR can result in a reduction in the power obtained. Selective NO_x recirculation (SNR) is one of the alternative methods where NO_x is selectively recirculated into the engine at concentrations higher than those with the conventional EGR. A sorbent material is used to adsorb the NO_x generated, which is then desorbed and recycled back to the engine. Sorbent technologies Inc. developed a sorbent bed made of a special type of sulfur-tolerant carbon granule material [33]. The bed was used downstream of a Cummins engine and experiments were performed to study the amount of NO_x adsorbed at various temperatures. Monticelli et al. [34] studied the potential of Y-Zeolites for NO_x adsorption and desorption using a synthetic exhaust gas with the temperature and composition representative of a lean-burn internal combustion engine. They carried out NO_x adsorption at low temperatures followed by its desorption at higher temperatures. It was observed that at temperatures above 100 °C, NO₂ can be selectively desorbed from the synthetic exhaust gas using the ion-exchanged Y-zeolites. Clark et al. [35] studied the effect of recirculated NO_x on its decomposition within the engine cylinder for lean-burn natural gas engines. The NO_x decomposition rate was shown to be affected by the air-fuel ratio, injected NO quantity, EGR amount, and the engine operating points. They developed a homogenous batch reactor model to study the effect of operating conditions on the reduction in NO_x. NO_x decomposition rates were predicted to be low for lean conditions and high for rich conditions. However, a complete engine model was not used and calculations were performed for fixed temperatures wherein the predictions were done for time periods obtained from the duration of normal combustion. Krutzsch et al. [36] studied the combustion as well as the NO decomposition mechanism during SNR application by experimental as well as theoretical techniques for two diesel engines and one gasoline engine. It was reported that the NO conversion efficiencies of up to 90% were achievable for a gasoline engine under rich or stoichiometric conditions whereas less than 20% conversion was obtained under lean conditions. The reaction paths by which the decomposition of NO occurs were also identified and analyzed by performing numerical simulations. Tissera et al. [37] studied the performance of SNR for lean-burn natural gas-based engines. It was reported that the engine load, amount of NO injected and the air-fuel ratio significantly affected the NO conversion rate while the engine speed did not affect the NO conversion rate if the amount of injected NO was kept constant.

Most of the studies with either SNR or injection of NO_x justify the reduction in the net amount of NO_x by the reaction of NO_x with the fuel, e.g., it was reported by Clark et al. [35] that the additional fuel injected creates a secondary re-burn zone, in which the NO_x formed in the main combustion zone gets converted to nitrogen. Moreover, it was assumed in their modeling studies that NO did not react once the conversion of the fuel (CH₄) reached 98%. They also noted that NO decomposition in a gasoline engine cannot be directly extended to a natural gas engine because of the low reactivity of natural gas. Similarly, Krutzsch et al. [36] proposed that the reduction in NO_x is primarily attributed to the chemical reaction between NO and ketenyl radicals, given by $\mathit{HCCO}+NO\rightleftharpoons HCNO+CO$. Glarborg et al. [24] also studied the reduction of NO with C₁ and C₂ hydrocarbons in a flow reactor and their modeling studies showed that NO is predominantly reduced by HCCO + NO reaction when using natural gas or C₂ hydrocarbons as the reburn fuel. However, we have shown that even if NO_x does not interact with the fuel, a net reduction in NO_x can still be achieved. Moreover, to the best of our knowledge, the injection of NO_x into a H₂ engine and its impact on the net NO_x emissions has not been reported elsewhere.

This work focuses on elucidating the reasons for the net reduction in NO_x upon injection of NO_x during the intake stroke of a H₂ engine and finding the conditions which favor the net reduction. A model is developed based on the coupling of basic conservation equations and the NO_x formation kinetics and is used to predict the effect of injected NO on the emission characteristics of the engine. The temporal profiles of in-cylinder pressure, temperature, and H₂O concentration are analyzed. It is shown that for the conditions of our study, these variables remain nearly unchanged for various amounts of injected NO and hence are not responsible for the reduction in the amount of NO formed during the power stroke. It is shown that even if the chemistry of NO_x reactions with the fuel is not considered, the amount of ‘fresh’ NO generated decreases with an increase in the amount of NO injected. It is also shown that a reduction in NO_x generation takes place at high equivalence ratios ($0.65\le \mathrm{\Phi }\le 0.9$) and significantly lower reduction is predicted at lower equivalence ratios. These observations are related to the attainment of near-equilibrium conditions, which occur only at high equivalence ratios.