Impact of Multi-Valve Exhaust Gas Recirculation (EGR) System on Nitrogen Oxides Emissions in a Multi-Cylinder Engine

Abstract

Exhaust gas recirculation (EGR) systems, in addition to catalytic reactors, are now widely used in reciprocating internal combustion engines to reduce oxides of nitrogen (NOx) in the exhaust gases. They are characterized by the fact that part of the exhaust gas from the exhaust manifold is recycled and directed to the intake manifold through a special valve. This valve, depending on the current engine load and velocity, doses an appropriate amount of exhaust gas which, with each new charge, is fed to the individual engine cylinders. In addition, the positioning of the valve has a significant effect on the formation of nitrogen oxides in the exhaust gas from individual engine cylinders, which is due to the uneven distribution of exhaust gas into the intake manifold channels. Tests were carried out on a power unit equipped with a symmetrical intake manifold with a centrally located EGR valve. The article presents the results of tests on a system in which each cylinder was supplied with a separate EGR valve. This solution made it possible to charge each cylinder with the same mass of recirculated exhaust gas, which was dependent on engine velocity and load. The exhaust nitrogen oxides emissions were measured for the originally manufactured system and compared with the multi-valve system. The results confirmed the need for individual selection of the dose of recirculated exhaust gas for particular cylinders, as the multi-valve system equalized the levels of nitrogen oxides emissions in the exhaust gases coming from individual cylinders of the internal combustion engine.

1. Introduction

One of the most difficult challenges faced by current engineers designing internal combustion engines, both compression-ignition and spark-ignition, is meeting the strict emission standards set by environmental advocates [1]. This is due to the fact that automotive transport has a significant share in the emission of toxic compounds, particularly nitrogen oxides, carbon oxides, and particulate matter, which are predominant in compression-ignition engines. The impact of these emissions negatively affects the natural environment and human health, leading to chronic and long-developing diseases.

Nitrogen oxides are among the most harmful compounds produced during the combustion of fossil fuels at high temperatures. Their negative impact on the human body is several times greater than that of carbon monoxide (CO). This group of compounds is generally referred to as NOx. It includes nitrous oxide (N₂O), nitric oxide (NO), nitrogen dioxide (NO₂), dinitrogen trioxide (N₂O₃), dinitrogen tetroxide (N₂O₄), and dinitrogen pentoxide (N₂O₅). The sources of NOx formation during combustion are the nitrogen contained in the fuel and the molecular nitrogen from the air. The concentration of nitrogen oxides in exhaust gases is highest when burning mixtures in the range of 1.05 < λ < 1.1. Combustion of mixtures outside this range (richer or leaner mixtures) results in reduced NOx emissions. However, it should be noted that the greatest impact on the formation of nitrogen oxides comes from the high temperature and pressure in the combustion chamber, which are a result of engine load [2]. One method to reduce nitrogen oxides in engines is the use of exhaust gas recirculation (EGR) systems, which promote fuel evaporation and oxidation of unburned hydrocarbons [3,4]. The primary goal of EGR systems is to lower the combustion temperature, but this process, in turn, leads to an increase in particulate matter emissions, which is an inherent problem in compression-ignition engines [5]. Reducing nitrogen oxides during combustion is also possible through increased fuel injection pressure, as well as the use of properly designed combustion chambers and intake systems [6,7]. Another essential component of modern compression-ignition engines is the additional selective catalytic reduction (SCR) system. This method involves injecting a urea solution before the catalyst, which then decomposes into ammonia and carbon dioxide. In the catalyst, a reaction occurs between ammonia and nitrogen oxides, producing nitrogen and water vapor. However, the efficiency of NOx reduction is highly dependent on the metal oxides used in the catalyst and the operating temperature of the catalyst [8,9,10].

Current research focuses on reducing nitrogen oxides in exhaust gases using various EGR system configurations. These include low-pressure systems (exhaust gas intake after the turbocharger), high-pressure systems (exhaust gas intake before the turbocharger), hybrid systems, and both low- and high-temperature systems [11,12,13]. In publication [14], it was shown that increasing the level of “hot” EGR results in a decrease in the density of the oxygen-rich charge, leading to reduced NOx emissions but causing an exponential increase in PM emissions. Much better results were achieved when cooled exhaust gases were used. At high loads, there was a significant reduction in nitrogen oxides and particulate matter in the exhaust gases. However, a slight increase in NOx at lower loads was due to the delayed autoignition of diesel fuel injected into the lower-temperature charge. According to [15], NOx emission reduction can be almost proportional to the degree of exhaust gas recirculation, with up to a 50% reduction in NOx achievable at just 20% EGR usage without an increase in unburned hydrocarbons or particulate matter. Studies [16,17,18] have discussed the effects of different engine boost levels using both high- and low-pressure EGR on nitrogen oxide emissions, particulate matter, and fuel consumption. The authors of these studies achieved a NOx reduction in the range of 60–70% using only a high-pressure EGR system. Achieving such a high NOx reduction in hybrid systems is only possible with a properly calibrated system and depends on the engine’s operating parameters, as also confirmed in studies [19,20].

An interesting solution for reducing nitrogen oxides without a significant increase in particulate matter and carbon monoxide, when using an EGR system in a homogeneous charge compression-ignition HCCI engine, was presented in study [21]. The research focused on four methods of exhaust gas delivery, where the percentage distribution of exhaust gases between a direct channel and a swirl channel supplying a single cylinder was varied. According to the authors, the greatest reduction in nitrogen oxides and particulate matter was achieved in the charge where the highest percentage of recirculated exhaust gases was directed to the swirl channel.

This article presents a study on the impact of the use of the exhaust gas recirculation systems that evenly and unevenly dose exhaust gases into the intake manifold ducts of a multi-cylinder engine, and their impact on the level of emissions and concentrations of toxic substances from individual cylinders. This issue was chosen as a result of preliminary tests for an engine equipped with a single EGR valve, which yielded different concentrations of harmful substances from each cylinder, regardless of its operating parameters [22]. Final tests on the EGR systems made it possible to determine if and in what manner harmful substances emissions could be balanced and reduced while standardizing the combustion process in multi-cylinder engines.

2. Materials and Methods

2.1. Research Object

The engine used in the study was a turbocharged, 2 L, sixteen-valve compression-ignition engine from the Volkswagen Group (Wolfsburg, Germany), 2.0 TDi BKD, equipped with a symmetrical intake manifold with a centrally located EGR valve (see Figure 1). The manifold used has eight separate pipes supplying charge to the cylinders (two per cylinder). Each pair of pipes is connected to the head inlet ducts, whose function is to generate charge turbulence in the cylinder. Their construction and arrangement for a single cylinder are shown in Figure 2.

In order to carry out bench tests, it was necessary to individually measure the toxicity of the exhaust gases from individual cylinders of the tested power unit. This was realized by using four connecting ports located in the exhaust manifold, with the exhaust intake located in the engine head at the connection point of each pair of exhaust ducts of a single cylinder. In Figure 3, the connection ports for cylinders one to four were marked with numbers 1 to 4, while the connection port located behind the turbine for collecting the exhaust gas was marked with number 5. A Portable Automated Measuring System HORIBA PG350 (Irvine, CA, USA) gas analyzer and a Smoke Meter AVL 415S opacimeter (Graz, Austria) were used to measure the exhaust gas composition of the engine.

2.2. Nitrogen Oxide Measurement

The tests were carried out in two stages. The first stage was aimed at obtaining reference values for the emissions of nitrogen oxides in the exhaust gases of the originally manufactured intake system equipped with a single exhaust gas recirculation valve. In the second stage, it was necessary to replace the existing single EGR valve solution with a device that evenly distributes the exhaust gases between the cylinders of the combustion engine. The designed system was intended to maintain the pressures and velocities of the exhaust gas flow from the original system, and to ensure its continuous flow into all channels of the intake manifold by means of splitting the main charge symmetrically into four smaller streams. The completed system, with the even exhaust gas distribution, is shown in Figure 4.

The system consisted of a spiral duct connected to the exhaust gas cooler, a symmetrical distribution manifold, connecting pipes, four EGR valves, and a spacer connecting the intake manifold to the head. The first component of the system was the duct connecting the exhaust gas cooler to the distribution manifold. Its purpose was to stabilize the flow of gases coming out of the cooler. The elimination of internal flow disturbances, which possibly occur due to increased duct thickness at the seams, was achieved by using a precision seamless pipe. The inner diameter of the pipe was 24 mm, which equaled the diameter of the original pipe connecting the cooler with a single EGR valve. The radius of the spiral coil was 200 mm and its pitch was 115 mm. A symmetrical distribution manifold was fitted to the end of the duct (Figure 5), which was designed to split the inlet flow into four equal outlet streams.

2.3. Numerical Modeling

For the designed system, numerical calculations were performed and flow parameters were checked for the split gas stream using finite volume computer modeling in Ansys Fluent v.12 Fluid Simulation Software (Ansys Inc., Canonsburgu, PA, USA). Figure 6 shows an example of the distribution of the gas velocity vector field for the initial part of the modeled system. The inlet velocity of 8.7 m/s was calculated based on engine specifications at 800 rpm.

The model was calculated using the RANS method with a two-dimensional K-epsilon turbulence model applied. The Y+ value on the wall was maintained within the range of 30–100. The computations were conducted with an accuracy level of 10⁻⁴. The mesh consisted of 3,670,059 triangular cells, 9,022,709 faces, and 1,749,815 nodes. Boundary conditions for inlet were the following: Type—velocity-inlet, Velocity Specification Method—Magnitude, Normal to Boundary, Reference Frame—Absolute, Velocity Magnitiude—8.7 m/s, Turbulence Specification Method—Intensity and Hydraulic Diameter, Turbulent Intensity—10%, Hydraulic Diameter—24 mm. Boundary condition for outflow: Type—outflow, Flow Rate Weighting—1. Fluid (exhaust gases) specification: density—1.3662 kg/m³, Cp (Specific Heat)—1146.46 J/kg×K, thermal conductivity—0.02182 W/(m×K), viscosity—1.4834 × 10⁻⁵ kg/(m×s), molecular weight—28.84 kg/kmol.

The marked inlet velocity contour fields are practically even over the entire area and result from the defined value of the initial velocity, whose vector was oriented perpendicular to the inlet. In order to demonstrate the stabilisation of the flow in the main duct, cross sections A-A, B-B, C-C, and D-D show successive distributions of the velocity fields. The purpose of the distribution manifold used was to generate four identical exhaust gas streams, which further supplied the individual EGR valves. As the gas density was assumed to be constant throughout the entire volume of the duct, according to Equation (1):

m = ρ × v × A,

(1)

where m [kg/s]—mass flow rate, ρ [kg/m³]—density, v [m/s]—velocity, A [m²]—surface area, then an even mass flow rate in the distribution manifold channels depended on the volumetric flow rate described by Equation (2):

Q = v × A,

(2)

where Q [m³/s]—volumetric flow rate.

Therefore, identical gas streams could be obtained if the even velocity field distribution in the duct fed the exhaust gas to the distribution manifold. Figure 7c shows a view of this duct at its end, just before the distribution manifold in the C-C section. This distribution was achieved by using a rectilinear flow-stabilizing duct. Its length was determined on the basis of PN-93M-53950 01 [23].

Comparing the results with those on cross sections B-B, a significant improvement in the geometric distribution of the velocity field in the flow channel can be seen, in which a laminar flow with the highest velocity in the central part of the duct can be observed.

The final component of the system tested was the exhaust manifold, in which the exhaust gas stream should be evenly distributed. Figure 7d shows the contour distributions of the velocity fields for individual outlets labeled w1, w2, w3 and w4. For the given initial conditions, a similar distribution of velocity contours could be observed in the individual outlet channels, with a concomitant increase in their values due to a reduction in the total area of the outlet relative to the area of the inlet.

Table 1. Flow parameters for individual distribution manifold outlets at v = 8.7m/s.

Outlet No.	Velocity [m/s]	Absolute Pressure [Pa]	Mass Flow [kg/h]	Volumetric Flow [m3/h]
w1	18.18	101,041	4.319	3.52
w2	18.17	101,038	4.319	3.52
w3	18.17	101,041	4.319	3.52
w4	18.18	101,037	4.319	3.52

and

Table 2. Flow parameters for individual distribution manifold outlets at v = 32m/s.

Outlet No.	Velocity [m/s]	Absolute Pressure [Pa]	Mass Flow [kg/h]	Volumetric Flow [m3/h]
w1	66.38	97,873	15.88	12.96
w2	66.37	97,823	15.88	12.96
w3	66.38	97,874	15.88	12.96
w4	66.40	97,820	15.88	12.96

summarize the results of the key flow parameters generated in Ansys Fluent for the minimum and maximum exhaust gas flow rates for which tests were carried out. The minimum value corresponded to an idling velocity of 800 rpm and the maximum value to a velocity at which the EGR valve was still active and was 3000 rpm.

The flow values were the average obtained at the outlet of a single channel. The analysis of flow similarity between the distribution manifold outlets was based on the percentage difference calculated between the maximum and minimum values of the parameters in relation to their average value obtained from the four channels. At the inlet velocity v = 8.7 m/s, the difference in velocity was ∆v = 0.0847%, pressure ∆p = 0.0033%, mass flow ∆ṁ = 0.000058%, and volume flow ∆Q = 0.000059%, while for the inlet velocity of v = 32 m/s, ∆v = 0.0484%, ∆p = 0.0542%, ∆ṁ = 0.000138%, and ∆Q = 0.000136%, respectively.

The numerical calculations carried out and the calculations describing the maximum differences in flow, designed for the initial part of the system with defined geometrical dimensions, made it possible to conclude unambiguously that the system meets the requirements for even distribution of the main exhaust gas stream into four smaller ones.

The exhaust gases from the system described above were fed to four EGR valves, which were connected to a spacer mounted between the original intake manifold and the head. This solution allowed the exhaust gases from the EGR valves to be fed directly to the individual intake manifolds of individual cylinders. Measurement points for rotational speeds and loads were determined based on bench tests, for which the EGR valve was open. The tests were carried out for rotational speeds in the 1000, 1200, 1400, 1600, 1800, 2000, and 2200 rpm range, and 20, 60, and 110 Nm loads [24].

3. Results

In both cases, when measured using the original EGR system and the controlled exhaust gas dosing system, there were different nitrogen oxides emissions and opacity emitted from individual cylinders. The reason for this is the uneven charge of the cylinders, resulting from wave phenomena and flow resistance occurring in the intake system, and the different residual exhaust gases in the cylinders from the previous combustion process. Given the uneven charge of the cylinders (the smallest degree of new charge in the first cylinder), it was decided to carry out another series of measurements, in which exhaust gases were dosed only to cylinders 2, 3, and 4.

The following diagrams present measurements of nitrogen oxide concentrations and smoke opacity of the systems tested, with a single diagram showing the range of changes in concentrations of a single substance from individual engine cylinders and the exhaust pipe; the results in Figure 8, Figure 9 and Figure 10 marked with letter:

(a)

denoted the “zero state” and characterized the changes in nitrogen oxide (NOx) concentrations obtained for the engine with the EGR valve on;

(b)

were carried out for a system using four EGR valves supplying exhaust gas only to the swirling ducts of the intake manifold;

(c)

were performed for the system from point (b) with the disconnected valve of the first cylinder.

For the factory EGR valve at a 20 Nm load (Figure 8a), there was a variation in NOx emissions from individual cylinders across all rotational speeds, but the lowest emission almost always came from cylinder no. 1 (cyl 1), with the exception of the speed of 2000 rpm. In the case of using a system that distributes exhaust gases evenly (Figure 8b), an increase in NOx emissions was observed for all rotational speeds, along with greater unevenness in emissions between cylinders compared to the factory system. For the setup with the EGR system of the first cylinder disconnected (Figure 8c), despite a general increase in NOx emissions from all cylinders, there was a leveling of NOx emissions at speeds of 1000, 1200, and 1400 rpm. In all cases (Figure 8a–c), the NOx emission from the collector pipe (exh) was higher than from any individual cylinder.

For the factory EGR valve at a 60 Nm load (Figure 9a), similar to the 20 Nm load, there was a slight variation in NOx emissions from individual cylinders across all rotational speeds. However, the lowest emission from cylinder no. 1 only occurred at speeds of 1000, 1400, and 1600 rpm. In the case of using the system that distributes exhaust gases evenly (Figure 9b), an increase in NOx emissions was observed for all rotational speeds, along with greater unevenness in emissions between cylinders compared to the factory system. For the setup with the EGR system of the first cylinder disconnected (Figure 9c), an increase in NOx emissions from all cylinders was noted at speeds from 1200 to 1600 rpm compared to the factory solution. Additionally, at speeds from 1000 to 1600 rpm, the emission from the first cylinder was the highest.

For the factory EGR valve at a 110 Nm load (Figure 10a), similar to the previous loads, there was a variation in NOx emissions from individual cylinders. Additionally, significant fluctuations in NOx emissions were observed at different rotational speeds. In the case of using the system that distributes exhaust gases evenly (Figure 10b), an increase in NOx emissions was recorded for all rotational speeds; however, the unevenness of emissions was comparable to the factory system. For the setup with the EGR system of the first cylinder disconnected (Figure 10c), an increase in NOx emissions from all cylinders was noted at speeds of 1400 and 1600 rpm compared to the factory solution. In other cases, the emission values were lower, and the uneven distribution was comparable to the factory configuration.

4. Discussion

With the use of four EGR valves directing exhaust gases into the swirling ducts (Figure 8, Figure 9 and Figure 10), an increase in NOx concentration in individual cylinders was observed compared to the scenario with a single EGR valve. This change occurred because the main exhaust gas stream was divided within the distribution manifold into four smaller streams. The 8 mm diameter ducts caused a reduction in exhaust gas pressure, leading to a smaller proportion of these gases in the charge delivered to the cylinders. Furthermore, since the exhaust gases were introduced exclusively into the swirling ducts, the central region of the combustion chamber’s charge likely lacked recirculated exhaust gases. This resulted in a rise in combustion temperature within the chamber. In the analyzed case, there was a greater disparity in NOx concentration distribution across cylinders compared to when a single EGR valve was used. This phenomenon was observed across all load conditions, though it varied with engine speed. The uneven distribution of exhaust gases, coupled with the irregular intake of clean air into the cylinders due to wave effects in the intake manifold, was a significant factor. Given the sequence of intake valve openings (1–3–4–2 for the tested engine), the closure of the intake valve in cylinder 2 caused a rebound in the airflow. This disrupted the charge flow to cylinder 1, whose valve opened next—a similar interaction occurred between cylinders 3 and 4. The reduced filling rate, combined with forcing equal exhaust gas flows through all EGR valves, led to lower NOx concentrations in cylinders 1 and 4 compared to cylinders 2 and 3 during combustion. In systems with evenly distributed exhaust gas doses, uneven engine performance was noted at loads of 60 and 110 Nm. This occurred at engine speeds where substantial disparities in charge composition between cylinders resulted in uneven NOx concentrations.

When comparing the graphs in Figure 8c, Figure 9c and Figure 10c, it was found that NOx concentrations between cylinders became more balanced at most measured operating points. Exceptions included engine speeds between 1000 and 1400 rpm at a 60 Nm load, and speeds of 1200 and 1400 rpm at a 110 Nm load. At these points, a higher clean air filling rate in the first cylinder—lacking recirculated exhaust gases—caused an increase in NOx concentrations.

5. Conclusions

The analysis of the test results revealed that that the application of systems evenly dosing exhaust gases into the intake manifold channels led to the equalisation of toxic substance concentrations between the cylinders of a multi-cylinder engine. The highest percentage equalization of the tested concentrations of toxic substances was obtained for the systems:

(1)

Evenly dosing the exhaust gas into the four swirling ducts (Figure 8b, Figure 9b and Figure 10b), at loads of: 20 Nm and rotational velocities of 1800 and 2000 rpm, 110 Nm and rotational velocities of 1800 and 2000 rpm;

(2)

Evenly dosing the exhaust gas into the three swirling ducts (Figure 8c, Figure 9c and Figure 10c), at loads of: 20 Nm and rotational velocities of 1200 and 1600 rpm; 60 Nm and rotational velocities of 1600 and 1800 rpm; 110 Nm and rotational velocities 1600, 1800, and 2000 rpm.

On the basis of the test results, it was concluded that it is possible to standardize and reduce the values for the achieved concentrations of nitrogen oxides and consequently toxic substances between the cylinders of an internal combustion engine by independently enriching the individual charges with recirculated exhaust gases. This turns out possible on the condition that the filling rate of each cylinder with a new charge is taken into account. In addition, the uneven distribution of the recirculated exhaust gas results in different combustion processes in the individual cylinders, and thus uneven energy yields from the individual cylinders, which can affect the variable rotational velocity of the crankshaft during the operating strokes.