Printer Friendly

Preliminary Investigation of Exhaust Pressure Waves in a Single Cylinder Diesel Engine and the Impacts on Aftertreatment Sprays.


Pressure wave actions are commonly encountered in reciprocating internal combustion engines due to the periodic gas exchange process. The pressure wave has been an interesting research topic because of its important impacts on engine performance. Firstly, the engine volumetric efficiency and gas exchange process are affected by pressure waves. Engine efficiency can be improved by properly designing the intake and exhaust system [1,2]. Moreover, the pressure wave induced non-uniform gas flow changes the gas composition of the in-cylinder charge, which can further influence the in-cylinder combustion process and engine-out emissions. In addition, the operation of turbine in turbocharged engines can be affected by the pressure wave action as well [3]. Due to these impacts, the pressure wave is an important consideration in engine control model development [4, 5, 6]. Previous studies on engine pressure waves include both experimental and simulation research work [7, 8, 9, 10]. The experimental method usually involves measurement of the pressure in the intake or exhaust manifolds under various engine operating conditions [8,9]. For pressure wave simulation, many researchers have used the one dimensional (1-D) approximation [7,10]. Since the pressure wave in the engine system primarily changes along the axial direction of the pipe, while the differences in the cross-sectional area have a minor effect, reasonable results have been obtained using the 1-D simulation methods [7,8].

In addition to the above mentioned impacts of pressure waves, there are some other situations where the pressure wave plays an important role. In modern internal combustion engines, sophisticated exhaust after-treatment systems are often used to reduce engine out emissions. These after-treatment systems include the three-way catalyst (TWC) in gasoline engines, selective catalytic reduction (SCR) and diesel particulate filter (DPF) in diesel engines. The distribution of the exhaust gas flow upstream of the aftertreatment converters can affect the active volume and subsequently the conversion efficiency of the converters [11, 12, 13]. Moreover, for the SCR and the DPF in diesel engines, a secondary injector is often used to provide a supplemental liquid such as urea or diesel fuel for emission reduction and catalyst regeneration. Oh et al. [14] investigated the spray distribution of an SCR injector and found that improved uniformity and atomization of the urea spray resulted in a higher NOx reduction efficiency of the SCR. In their experiments, the gas flow was assumed steady. However, the actual exhaust flow can be highly pulsating with a high amplitude of transient velocity and pressure waves. Nishiyama et al. [15] used Endoscopic Stereo Particle Image Velocimetry (PIV) to measure the three dimensional exhaust gas flow around an oxygen sensor mounted on the exhaust manifold of a four cylinder gasoline engine. Their results showed that an exhaust flow with transient velocity greater than 60 m/s was produced after the exhaust valve opened. Under such flow conditions, the after-treatment spray could be affected which may further influence the function of a catalytic converter.

There were several reports on the interactions between exhaust flow and after-treatment sprays, and most of them were conducted on off-engine test benches using optical methods [14,16,17]. The gas temperature and mass flow rate were commonly investigated in these studies assuming steady gas flow conditions [14,16,17]. The main challenges for conducting after-treatment spray research directly on an engine system are the high temperature environment and the limited optical access, which can usually be complemented by the steady flow test benches. Yet, one of the deficiencies for a steady flow test bench is that it is difficult to replicate the pressure waves and pulsating flow conditions of an engine system. There are some other researchers using pulsating flow rigs [13,18] or shock tube devices [19, 20] to simulate the pulsating pressure and gas flow of an engine system. Liu [13] used a pulsating flow rig to study the effect of pulsating gas flow on the performance of a catalytic converter. Chalet et al. [15] developed a 1-D simulation model for calculating the flow at pipe junctions and verified the model with shock tube test results. The verified model was used later on to simulate the exhaust flow in an engine system and the results showed good agreement with the engine test results. These research results provided the background for investigating the after-treatment spray under unstable gas flow test conditions. From the authors' search, there is no comprehensive work on integrating pressure wave measurement on an engine, with off-engine spray research and 1-D computational simulation. This paper is a preliminary step in that direction with the primary objective being to understand the characteristics of the pressure wave under different engine operating conditions. A secondary objective is to study the effect of the pressure wave on the after-treatment injection process.

In this work, the exhaust pressure wave was measured and identified on a single cylinder compression ignition research engine under different engine operating conditions. The measurement results were also used to validate the simulation model based on the 1-D gas dynamic method using AVL BOOST[TM]. The validated simulation model was then used to estimate the exhaust gas flow velocity. Thereafter, an off-engine shock tube test bench was used to create a simulated exhaust gas flow condition to investigate the impacts of the pressure wave and pulsating gas flow on the droplet breakup and spray distribution of an after-treatment injection using optical methods.


Experimental Setup

Engine Test Setup

The engine tests were conducted on a single cylinder research engine which was modified from a four cylinder production diesel engine. The specifications of this engine are listed in Table 1. The intake and exhaust system of the first cylinder were isolated from the other three cylinders which operated at a fixed low load condition to motor and balance the research cylinder during engine tests. The intake and exhaust conditions, as well as the injection pressure, injection timing and duration, of the research cylinder were independently controlled by in-house designed LabVIEW programs. The configuration of the engine test system is shown in Figure 1.

A critical parameter in this study was the exhaust pressure. The exhaust pressure measuring port was located on a straight exhaust pipe 10 cm downstream of the exhaust valve. The diameter of the pipe was 25.5 mm. Further downstream, a diesel oxidation catalyst (DOC) and an exhaust surge tank were installed in the exhaust system. The DOC was used to reduce incomplete combustion products while the surge tank was used to dampen the pressure wave action and provide a steady condition for the in-cylinder combustion research. These components, however, would not affect the pressure wave actions as the measuring port was right after the exhaust valve. The downstream conditions had relatively minor effects on the pressure wave and are outside the scope of this paper. A more comprehensive analysis is provided in [21].

A Kistler 4075A10 piezoresistive pressure transducer was used for the exhaust pressure measurement. The natural frequency of the transducer was 120 kHz. An AVL GU13P piezoelectric pressure transducer was used to measure the in-cylinder pressure. Both the in-cylinder and exhaust pressures were acquired at a sampling rate of 0.1 [degrees]CA and recorded on a Windows based computer. 200 consecutive cycles were recorded at each data point and the averaged value was reported. The engine speed was kept constant at 1500 rpm throughout the tests which resulted in an equivalent sampling frequency of around 90 kHz. A number of thermocouples were installed at different locations in the exhaust system to monitor the exhaust temperature which also provided information for boundary conditions in the simulations. The primary exhaust temperature measurement was performed using a K-type thermocouple which was installed about 5 cm downstream of the exhaust valve on the exhaust pipe. The exhaust temperatures mentioned in the subsequent sections were measured using this thermocouple.

Shock Tube Test Setup

A shock tube device was used to create pulsating flow conditions which had a similar pressure and velocity to the engine exhaust manifold. The driver to driven section pressure ratio was controlled by using different types of diaphragms. The shock tube was closed at one end and open to the atmosphere at the other end. A liquid droplet (Figure 2) or a low pressure after-treatment injector (Figure 3) was mounted at the open end of the shock tube and the interactions between droplets, pressure wave and high velocity gas flow were investigated by optical methods. The total length of the shock tube was 1.36 m with a 1 m long driven section and 0.36 m long driver section. The inner diameter of the tube was 23.8 mm. Compressed air was used as charging gas in the tests. The compressed air was not heated, thus the impact of gas temperature was not investigated in this set of tests.

An SSI Technologies pressure transducer (model number: P51-200-S-A-I36-5V-000-000) was mounted on the driver section 15 cm away from the diaphragm. This transducer was used to measure the charging pressure and record the diaphragm burst pressure. Two Kistler 4075A10 pressure transducers were mounted on the driven section of the shock tube. The first one was mounted 10 cm downstream of the diaphragm location and the second 10 cm before the open end. When the diaphragm burst, a shock wave was generated and transmitted downstream. The first Kistler pressure transducer detected the pressure rise and sent out a trigger signal to initiate data recording. The signals from all the pressure transducers were transmitted and recorded on a real-time controller (RT). A LabVIEW program was used to control the data recording process with a sampling frequency of 100 kHz. The pressure data from 20 ms before until 30 ms after the trigger were recorded

A Vision Research Phantom v7.3 high-speed camera was used to record the breakup process of liquid droplets under pulsating gas flow. Shadow graph images were taken to record the detailed interactions between the gas flow, the pressure wave and a single droplet, while direct images were recorded to show the behavior of a liquid spray under pulsating gas flow conditions. The camera speed was set to 40000 fps with a resolution of 512 x 128 pixels and an exposure time of 8 [micro]s for the shadow graph images. For the direct image of the injection tests, the camera speed was set to 20000 fps with a resolution of 256 x 512 pixels and an exposure time of 47 [micro]s. All the droplet and spray images were acquired at room temperature. The camera was triggered by the same pressure trigger that was used for data recording. The reference time zero was thus from the time when the shock wave arrived at 10 cm downstream of the diaphragm in the driven section.

Simulation Method

Although the transient exhaust pressure could be measured with the fast response pressure transducer, the transient exhaust gas flow velocity was difficult to measure directly. This information was attained through the simulation method based on the measured pressure and temperature information.

The AVL BOOST[TM] software is a 1-D simulation tool often used for engine simulations. This software features models for different components in the engine system, such as cylinder, pipe, plenum and catalytic converter. Three dimensional effects, such as frictional loss at pipe elbows, were simulated by flow coefficients. The simulation boundary conditions, such as pressure and temperature, were set to the same values as in the engine tests. The in-cylinder pressure acquired from engine tests was used as an input boundary condition for the pressure wave simulation.

The pipe flow simulation in AVL BOOST[TM] is based on the Euler equations. An essentially non-oscillatory (ENO) shock capturing scheme is used to solve the non-linear governing equations. The scheme is based on a finite volume approach where the value at the end of the time step is calculated based on the value at the beginning of the time step and the flux through the cells' borders. The mass, momentum, and energy flux are calculated through the conservation equations [22]. The ENO schemes use adaptive stencils to automatically achieve high order accuracy [22].


Engine Exhaust Pressure Wave and Gas Flow Velocity

Engine Test Results

The exhaust pressure was directly measured during engine tests. With the single cylinder setup, the interactions between different cylinders were avoided. The phase of the exhaust pressure wave with respect to the piston and the valve movement was clearly visualized over a complete cycle of 0[degrees] to 720[degrees]crank angle (CA), as shown in Figure 4. The first phase was during the intake stroke where the exhaust valves closed (EVC) at 10[degrees]CA, shortly after the intake TDC (at 0[degrees]CA). Then the intake valves closed (IVC) at 217[degrees]CA. Thereafter, both the intake and the exhaust valves remained closed, and the piston moved towards the combustion TDC at 360[degrees]CA. The exhaust pressure wave was weak during this period when the exhaust valves were closed since the research cylinder manifold was isolated from the other cylinders, and the long exhaust pipe dampened out any pressure waves from the previous cycles. During the compression and combustion process, the in-cylinder pressure and temperature increased. Then at 491[degrees]CA, the exhaust valves opened (EVO), and the piston continued moving towards the expansion BDC. During this process, two distinct compression waves with comparatively high amplitudes were observed in the exhaust pipe. The first compression wave right after the exhaust valve opening (EVO) was created by the pressure difference between the hot gas inside the cylinder and in the exhaust pipe. This process is often referred to as the "blow down" process. As the exhaust valves opened wider and the piston moved to bottom dead center (BDC @ 540[degrees]CA), the in-cylinder pressure gradually decreased. When the piston continued moving towards intake top dead center (TDC @ 720[degrees]CA), it compressed the in-cylinder charge as well as the gas in the exhaust pipe, and produced a second compression wave. This process is often referred to as the "displacement" process which pushes the residual gas out of the cylinder. Both the in-cylinder pressure and the exhaust pressure increased during this phase.

From these observations, it was seen that the exhaust pressure wave was closely related to the in-cylinder conditions when the exhaust valves were open. The corresponding pressures in this process were marked in Figure 4, where pc-Evo indicated the in-cylinder pressure at EVO, pe represented the exhaust backpressure at EVO, and pe_max indicated the maximum exhaust pressure which was caused by the compression pressure wave during the blow down process. The value of pe was thus similar to the value of the mean backpressure due to the minor pressure fluctuation.

Two sets of tests were conducted to investigate the effect of the in-cylinder conditions and the exhaust conditions on the exhaust pressure wave respectively. The test conditions are listed in Table 2. No exhaust gas recirculation (EGR) or intake air heating was used in these tests.

In the first set of tests, the exhaust backpressure was kept constant while the in-cylinder pressure at EVO was gradually increased by increasing the engine load. The intake pressure was set to approximately 2.0 bar absolute, and the load was increased by increasing the quantity of the injected fuel. As more fuel energy was released to heat up the similar quantity of in-cylinder charge, the in-cylinder pressure and temperature increased. Figure 5 shows the in-cylinder pressure under different load levels.

Corresponding to the increase of in-cylinder pressure and temperature, the peak pressure in the exhaust pipe was increased and the duration of the compression pressure wave was longer, as shown in Figure 6. The increased peak exhaust pressure suggested that a stronger compression wave was formed due to the higher in-cylinder pressure at EVO. The phasing of this compression wave was similar, with the peak pressure being achieved at about 530[degrees]CA in all three cases. This phasing was related to the timing when the exhaust valves opened which was the same in all three cases. The pressure wave during the displacement process was not significantly affected by the change of the in-cylinder pressure. The duration and amplitude were similar in all the three cases. This was probably because that the pressure wave during the displacement process was mainly related to the piston movement. In this case, the mass of the in-cylinder charge was similar under the constant boost pressure and the engine rotation speed was also the same. So, the amplitude of the second pressure wave was not changed. The exhaust gas temperature values are also indicated in the figure. As explained previously, a higher exhaust gas temperature was expected due to the increased in-cylinder temperature at a higher load level.

The second set of tests were conducted by maintaining the same in-cylinder pressure at EVO with an almost constant load level, but the backpressure was increased. As shown in Figure 7, the peak pressure in the exhaust pipe was consistently higher with a higher back pressure, but the pressure increase during the blow down process with respect to the backpressure level decreased. The pressure wave induced peak exhaust pressure increased by 71% when the backpressure was 1.3 bar, however it increased by only 29% when the backpressure was 2.2 bar. This suggested that a higher backpressure suppressed the pressure increase after EVO. The phasing of this compression wave was not changed with the backpressure, which was around 530[degrees]CA in all the cases. The pressure wave during the displacement phase was weaker when the backpressure was lower. This can again be explained through the process of the displacement. When the backpressure was higher, there was a higher resistance for the piston upward movement and thus a stronger compression wave was produced. As the injected fuel and the combustion phasing were largely the same, the exhaust temperatures in the four cases were similar, at around 250[degrees]C.

Based on these results, more engine test data were analyzed to further explore the relationship between the in-cylinder pressure, the exhaust backpressure and the exhaust pressure wave. As shown in Figure 8, a linear relationship between the in-cylinder pressure at EVO, the exhaust backpressure at EVO and the peak exhaust pressure was derived at different engine operating conditions. The various engine operating conditions included different loads, backpressure levels, combustion phasing, different injection strategies such as single and multiple fuel injections, and different fuels such as diesel and neat n-butanol fuel. The results under all the test conditions followed the same trend which indicated that the most important factors affecting the exhaust pressure wave were the in-cylinder pressure and exhaust backpressure at EVO. This relationship suggested that under the same backpressure, a higher in-cylinder pressure at EVO created a stronger compression wave during the blow down phase. In contrast, with the same in-cylinder pressure at EVO, a higher backpressure suppressed the pressure rise during the blow down process. It should be emphasized that all the test data shown in Figure 8 was acquired with the single cylinder test setup and the linear trend may not be applicable in multi-cylinder cases.

Simulation Results

The simulation tool AVL BOOST[TM] was used to study the gas flow conditions in the exhaust pipe under different engine operating conditions. The flow coefficients used in the simulation model were tuned to match the experimental results. These parameters were then held constant for all the simulations, while the simulation conditions such as the mean backpressure and the in-cylinder pressure were changed from case to case. They were measured directly during engine tests and used as input boundary conditions.

The simulation results of exhaust pressure wave under four different engine loads were compared with the experimental results. The simulation conditions were set to be the same conditions of test 1 listed in Table 2. The results showed good agreement with regards to both the amplitude and the phasing of the pressure wave as shown in Figure 9. The maximum absolute difference of the pressure within the complete engine cycle was less than 0.1 bar and the relative difference was less than 5% in all the cases. More validation cases can be found in previous work [21]. Overall, the simulation tool was deemed to be capable of simulating the exhaust pressure wave with reasonable accuracy.

One purpose of simulating the pressure wave was to estimate the exhaust flow velocity, which could not be measured by the authors directly. Figure 10 shows the simulated exhaust gas flow velocity under the four different engine loads where the backpressure was the same. The gas flow velocity was pulsating during the engine cycle. The flow velocity increased rapidly in a short period right after the exhaust valve opening at 491[degrees]CA. The peak value of the gas flow velocity was as high as 200 m/s under the load level of 13.6 bar IMEP. This transient velocity increased with engine load, which was consistent with the trend observed for the exhaust pressure wave shown in Figure 6. The velocity during the displacement process was similar in all four cases around 80 m/s. When the exhaust valves were closed, the velocity in the exhaust pipe was very low (around zero) due to negligible gas flow.

The gas flow velocity under different backpressure levels was also calculated using the simulation tool (shown in Figure 11). In this comparison, the in-cylinder pressure at EVO and engine load were kept constant, while the backpressure was elevated from 1.3 bar to 2 bar. The simulation conditions were set to be the same conditions of test 2 listed in Table 2. Similar to the trend of the exhaust pressure waves shown in Figure 7 where the pressure rise decreased with increasing backpressure, the peak gas flow velocity in the exhaust pipe was also lower with a higher backpressure. Generally, a compression wave which resulted in a larger pressure increase in the exhaust pipe after EVO also produced a faster gas flow. However, the absolute value of the exhaust pressure was not an indicator of the gas flow velocity. The velocity during the displacement was similar under the three different backpressure levels, which was different from the trend of the exhaust pressure wave shown in Figure 7 where the wave amplitude was lower under a lower backpressure. The reason for the similar gas flow velocity could be attributed to the same boost pressure and thus similar mass of in-cylinder charge in all the cases. The residual gas was pushed out during the displacement process. With a similar quantity of residual gas and the same engine rotation speed, the gas flow velocity was also similar.

Droplet Breakup under Fast Gas Flow Conditions

Based on the engine test and simulation results, it was observed that the exhaust pressure and gas flow fluctuated with the periodical opening and closing of the exhaust valves during the gas exchange process. The transient gas flow in the exhaust pipe was as high as 200 m/s under some conditions. The effects of this fast gas flow on the droplet breakup and the spray distribution of the after-treatment injection were investigated on the stand-alone shock tube test bench. The gas flow speed was estimated from the one dimensional shock tube theory. AVL BOOST[TM] was also used to simulate the shock wave and gas flow velocity in a shock tube with the driver section closed and driven section open to the atmosphere to provide sample wave profiles (Figure 12). It should be noted that the exhaust pressure wave and the shock wave were not the same type of pressure wave. The shock wave had a much sharper rising edge compared to the exhaust pressure waves as shown in Figure 12. Yet they had some similar features. Both of them were compression waves and they were both accompanied by a pulsating gas stream after the pressure wave. The major purpose of the shock tube tests was to create conditions that can to some extent represent the conditions in the exhaust pipe--similar order of exhaust pressure and gas flow velocity. However, the environment in the exhaust pipe was quite complicated. The pressure and temperature at different conditions were expected to have large variations. In this research, it was not feasible to simulate all the possible conditions with the shock tube test setup. The focus was to create some representative comparable conditions. The effect of temperature was outside the scope of this research and all the shock tube tests were conducted at room temperature.

Firstly, the breakup process of a single droplet under pressure wave conditions was investigated with the shadow graph method. The test conditions are listed in Table 3. The interactions between the pressure wave, the gas flow and the droplet are shown in Figure 13 and Figure 14 respectively.

Two kinds of liquid - water and diesel fuel, were used in the tests. They were tested under similar pressure wave conditions with a pressure ratio of 3.8 and gas flow velocity of roughly 150 m/s, which was in the range of the transient exhaust gas velocity at medium load conditions (Figure 10). Timewise, the flow duration with the high gas velocity in the exhaust pipe during the engine tests was approximately 30[degrees]CA, which was equivalent to 3 ms when the engine rotation speed was 1500 rpm. This duration was also similar to the conditions in the shock tube tests. The droplets were suspended at a needle tip about 3 cm from the open end of the shock tube and all the droplets had a similar diameter of about 1.5 mm.

As shown in Figure 13, the shock wave and the vortex ring structure caused by the sudden expansion of the gas flow were observed outside the open end of the shock tube at around 2 ms after the trigger signal. They both travelled toward the water droplet location. At 2.1 ms the shock wave passed the droplet location but no significant changes were observed on the droplet. However, at 2.2 ms the droplet began to break up when the high velocity gas flow arrived. Subsequently, the droplet was broken up into a mist and drifted away with the gas flow. The droplet breakup process lasted for about 1 ms.

The breakup process of the diesel droplet (Figure 14) was similar to that of water, except that the droplet stayed on the needle tip for a longer time, which might be attributed to the higher viscosity of diesel.

Spray Distribution under Fast Gas Flow Conditions

Knowing the behavior of a single droplet under fluctuating pressure and gas flow conditions, the spray test was conducted to investigate the impacts of pulsating gas flow on a liquid spray using the test setup shown in Figure 3. The test conditions are listed in Table 4.

Before running the spray test, the injector was tested under static conditions to see the original spray pattern. The low pressure liquid spray development under quiescent conditions is shown in Figure 15. Time zero was the time of the injection command. As shown in Figure 15, two water streams were injected from the low pressure injector. There were some large droplets and water streams especially near the nozzle tip. The injection pressure used in the tests was 4 bar. The low injection pressure did not create an optimum atomization of the liquid, which was similar to the case in the application of after-treatment injections where the injection pressure was usually less than 10 bar.

The injector was then placed at the open end of the shock tube. The spray behavior in the presence of the pressure wave is shown in Figure 16. The pressure ratio of the shock tube test was 4.1, which created a gas flow with a speed of approximately 170 m/s. In order to have a fully developed spray at the time when the shock wave arrived, the injector was opened earlier than the time of the trigger to compensate the time difference due to the injector opening delay. As shown in Figure 16, there were two fully developed spray streams 2 ms after the trigger. The interaction between the spray and the gas flow was first observed at 2.1 ms. The increased white area was the water mist. This area continued to expand in the next frame, when the second water stream was broken up. The breakup process continued during the following 2 ms. The atomization of the water spray was considerably improved by the gas flow. However, due to the high speed cross flow, the penetration of the spray was negatively affected. The streams could not penetrate through the whole cross-sectional area and water distribution was uneven with more water mist in the top and less in the bottom. The gas flow velocity gradually reduced as shown by the smaller white area of water mist after 5 ms. Instead of breaking the water droplets into water mist, the gas flow drifted the droplets away and broke them into comparably smaller droplets. 12 ms after the trigger, the gas velocity reduced and the spray went back to the original status.

Since the driven section was open ended and its pressure could not be changed, the pressure ratio was controlled by changing the driver pressure and using different diaphragms. Three pressure ratios, 1.8, 2.4 and 4.1, were used respectively. Under these pressure ratios, the gas flow velocity was about 70 m/s, 110 m/s and 170 m/s, respectively, which was similar to the peak gas flow velocity in the exhaust pipe as discussed previously.

The tests under the aforementioned three conditions showed similar results; the spray was broken up by the high velocity gas flow. In order to compare the difference in the spray breakup between each condition, gray scale images were converted into binary images to calculate the area of the water mist. The water mist was estimated by representing liquid with dark pixels as shown in Figure 17. The spread of the liquid was significantly larger and the process lasted for a longer duration when the pressure ratio was higher (Figure 18).


In both exhaust pipe and shock tube, the gas flow had a very high transient velocity which was much higher than the average velocity. The fast transient gas flow had a pronounced impact on the breakup process of liquid droplet. Due to this impact, the assumption of using the averaged velocity and steady flow condition to study the exhaust gas and after-treatment spray interaction may not be appropriate.

The transient high velocity gas flow can be used to improve the spray atomization and subsequently enhance the liquid evaporation. However, as the high velocity gas flow passed through very fast, as short as 3 ms under current testing conditions, its influence on the spray was limited due to a short duration of interaction. The current engine test setup was a single cylinder engine and the exhaust pressure wave was only affected by the valve and piston movement. In multi-cylinder engines the exhaust pressure wave would be more complicated with the interaction between different cylinders and the impacts on the spray might be more pronounced.


The exhaust pressure wave in a single cylinder engine system was investigated by both experimental and simulation methods. Results showed that:

1. The exhaust pressure wave was related to both the in-cylinder and exhaust backpressure conditions. A higher in-cylinder pressure created a stronger compression wave after the exhaust valve opening, while a higher backpressure suppressed this compression wave.

2. The exhaust gas flow velocity followed a trend similar to the exhaust pressure wave. The velocity was high during the exhaust valve open period and nearly zero when the valves were closed. The amplitude of the transient gas flow velocity during the blow down phase was related to the strength of the compression wave released from cylinder after exhaust valves opened. A stronger compression wave produced a faster exhaust gas flow.

3. A linear relationship was derived which showed that the ratio of the exhaust peak pressure and the exhaust backpressure was linearly related to the ratio of the in-cylinder pressure at EVO and the exhaust backpressure at EVO.

4. A shock tube device was used to create a similar condition as in the exhaust pipe except that the temperature effect was not considered. The breakup process of water and diesel droplets, as well as the water spray distribution under fast gas flow were recorded by a high speed camera. A higher transient velocity led to larger spatial distribution of the spray droplets.


[1]. Kesgin, U., "Study on the Design of Inlet and Exhaust System of a Stationary Internal Combustion Engine," Energy Convers. Manag., 46(13-14), pp. 2258-2287, 2005.

[2]. Galindo, J., Lujan, J. M., Serrano, J. R., Dolz, V., and Guilain, S., "Design of an Exhaust Manifold to Improve Transient Performance of a High-Speed Turbocharged Diesel Engine," Exp. Therm. Fluid Sci., 28(8), pp. 863-875, 2004.

[3]. Palfreyman, D., and Martinez-Botas, R. F., "The Pulsating Flow Field in a Mixed Flow Turbocharger Turbine: An Experimental and Computational Study," J. Turbomach., 127(1), p. 144, 2005.

[4]. Andwari, A. M., Aziz, A. A., Said, M. F. M., and Latiff, Z. A., "Experimental Investigation of the Influence of Internal and External EGR on the Combustion Characteristics of a Controlled Auto-Ignition Two-Stroke Cycle Engine," Appl. Energy, 134, pp. 1-10, 2014.

[5]. Stockar, S., Canova, M., Guezennec, Y., Della Torre, A., , "Modeling Wave Action Effects in Internal Combustion Engine Air Path Systems: Comparison of Numerical and System Dynamics Approaches," Int. J. Engine Res., 14(4), pp. 391-408, 2013.

[6]. Flardh, O., and Martensson, J., "Exhaust Pressure Modeling and Control on an SI Engine with VGT," Control Eng. Pract., 25, pp. 26-35, 2014.

[7]. D'Errico, G., Ferrari, G., and Onorati, A., "Numerical Modeling of Unsteady Reacting Flows in the Exhaust System of a SI Engine Including the Catalytic Converter," FISITA World Automotive Congress, Seoul, 2000.

[8]. Harrison, M. F., and Stanev, P. T., "Measuring Wave Dynamics in IC Engine Intake Systems," J. Sound Vib., 269(1), pp. 389-408, 2004.

[9]. Selamet, A., Kothamasu, V., Novak, J. M., and Kach, R. A., "Experimental Investigation of In-Duct Insertion Loss of Catalysts in Internal Combustion Engines," Appl. Acoust., 60(4), pp. 451-487, 2000.

[10]. Gasser, I., and Rybicki, M., "Modelling and Simulation of Gas Dynamics in an Exhaust Pipe," c Appl. Math. Model., 37(5), pp. 2747-2764, 2013.

[11]. Martin, A., Will, N., Bordet, A., Cornet, P. , "Effect of Flow Distribution on Emissions Performance of Catalytic Converters," SAE Technical Paper 980936, 1998, doi:10.4271/980936.

[12]. Weltens, H., Bressler, H., Terres, F., Neumaier, H. , "Optimisation of Catalytic Converter Gas Flow Distribution by CFD Prediction," SAE Technical Paper 930780, 1993, doi:10.4271/930780.

[13]. Liu, Z., "Pulsating Flow Maldistribution in Automotive Exhaust Catalysts--Numerical Modelling and Experimental Correlation," PhD thesis, Coventry University. 2003.

[14]. Oh, J., and Lee, K., "Spray Characteristics of a Urea Solution Injector and Optimal Mixer Location to Improve Droplet Uniformity and NOx Conversion Efficiency for Selective Catalytic Reduction," Fuel, 119, pp. 90-97, 2014.

[15]. Nishiyama, A., Jeong, H., Ikeda, Y., and Sawada, R., "Application of Endoscopic Stereo PIV to 3D Exhaust Gas Flow Measurements in a Practical SI Engine." Journal of Physics: Conference Series 85, 2007.

[16]. Dev, S., "An Investigation of a Diesel Liquid Injector in a Simulated Exhaust Flow.", Master's thesis, University of Windsor, 2013.

[17]. Grout, S., Blaisot, J.-B., Pajot, K., and Osbat, G., "Experimental Investigation on the Injection of an Urea-Water Solution in Hot Air Stream for the SCR Application: Evaporation and Spray/Wall Interaction," Fuel, 106, pp. 166- 177, 2013.

[18]. Benajes, J., Bermudez, V., Climent, H., and Rivas-Perea, M. E., "Instantaneous Pressure Measurement in Pulsating High Temperature Internal Flow In Ducts," Appl. Therm. Eng., 61(2), pp. 48-54, 2013.

[19]. Chalet, D., and Chesse, P., "Fluid Dynamic Modeling of Junctions in Internal Combustion Engine Inlet and Exhaust Systems," J. Therm. Sci., 19(5), pp. 410-418, 2010.

[20]. Chalet, D., Chesse, P., and Hetet, J. F., "Boundary Conditions Modelling of One-Dimensional Gas Flows in an Internal Combustion Engine," Int. J. Engine Res., 9(4), pp. 267-282, 2008.

[21]. Yang, Z., "Exhaust Pressure Waves in Diesel Engines and the Impacts on Aftertreatment Sprays.", Master's thesis, University of Windsor, 2015.

[22]. AVL, "BOOST Theory," 2011.


Ming Zheng

Department of MAME University of Windsor 401 Sunset Avenue, Windsor ON., Canada, N9B 3P4


This research is supported by NSERC CRD, APC, Discovery, CRC, CREATE programs; the CFI-ORF New Initiative Program, ORF Research Excellence programs; the NCE AUTO21 and BioFuelNet programs; the Ford Motor Company, and the University of Windsor.


CA - Crank angle

DOC - Diesel oxidation catalyst

DPF - Diesel particulate filter

EVC - Exhaust valve closing

EVO - Exhaust valve opening

fps - frame per second

IMEP - Indicated mean effective pressure

IVC - Intake valve closing

IVO - Intake valve opening

SCR - Selective catalytic reduction

TWC - Three-way catalytic converter

Zhenyi Yang, Shouvik Dev, Marko Jeftic, Christopher Aversa, Akshay Ravi, David Ting, and Ming Zheng

University of Windsor

Table 1. Test engine specification

Exhaust Valve      28 mm       Intake Valve Diameter    33.4 mm
Total Engine    1.998 Liters   Maximum Exhaust           8.64 mm
Displacement                   Valve Lift
Bore               86 mm       Maximum Intake Valve     12.5 mm
Stoke              86 mm       Intake Valve Opening    687 [degrees]CA
Connecting         144 mm      Exhaust Valve Opening   491 [degrees]CA
Rod                            (EVO)
Compression         18.2:1     Intake Valve Closing    217 [degrees]CA
Ratio                          (IVC)
Valves per           4         Exhaust Valve Closing    10 [degrees]CA
Cylinder                       (EVC)

Table 2. Engine test conditions

          IMEP     Boost pressure   (bar absolute)   Speed
          (bar)    (bar absolute)   Backpressure     (rpm)

Test 1    4.6             2              2            1500
Test 2   ~7               2            1.3            1500

Table 3. Shock tube tests with a single droplet

                                 Test 1     Test 2

Droplet diameter(mm)             ~1.5       ~1.5
Liquid                           water      diesel
Pressure ratio                    3.8        3.8
Gas                              air        air
Gas velocity(m/s)(estimated)   ~150 m/s   ~150 m/s
Gas temperature ([degrees]C)     26         26

Table 4. Shock tube tests with liquid sprays

Injection liquid                       water
Ambient gas                            air
Shock tube initial pressure ratio      4.1/2.4/1.8
Gas velocity (estimated, m/s)       ~170/110/70
Gas temperature ([degrees]C)          26
Injection pressure (bar absolute)      4
Injection duration (ms)               50
COPYRIGHT 2017 SAE International
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2017 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Yang, Zhenyi; Dev, Shouvik; Jeftic, Marko; Aversa, Christopher; Ravi, Akshay; Ting, David; Zheng, Mi
Publication:SAE International Journal of Engines
Date:Apr 1, 2017
Previous Article:Three-Dimensional Three-Component Air Flow Visualization in a Steady-State Engine Flow Bench Using a Plenoptic Camera.
Next Article:Analysis of NOx Emissions during Crank-Start and Cold Fast-Idle in a GDI Engine.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters