Printer Friendly

Instantaneous Flow Rate Testing with Simultaneous Spray Visualization of an SCR Urea Injector at Elevated Fluid Temperatures.


Nitrogen oxides (N[O.sub.x]) emissions reduction legislation worldwide has led to the introduction of new exhaust aftertreatment systems, particularly for lean-burn technologies such as compression-ignition (diesel) engines. Innovative exhaust aftertreatment technologies have been industrialized that treat N[O.sub.x] under the high-oxygen exhaust gas stream conditions typical of lean-burn engines.

One of these technologies comprises a catalyst that facilitates the reactions of ammonia (N[H.sub.3]) with the exhaust N[O.sub.x] to produce nitrogen ([N.sub.2]) and water ([H.sub.2]O). This technology is referred to as Selective Catalytic Reduction (SCR). Many original equipment manufacturers of passenger cars implement this type of system on their diesel powertrains today. A good example of various innovations that have been implemented for SCR--close-coupled injection, SCR on Diesel Particulate Filters (DPF), and others--was described in reference [1] (1).

Ammonia is difficult to handle in its pure form in the automotive environment, therefore it is customary with these systems to use a liquid aqueous urea solution, typically at a 32% concentration of urea ([CO(N[H.sub.2]).sub.2]). The solution is referred to as AUS-32, and is also known under its commercial name of AdBlue[R] in Europe, and DEF--Diesel Exhaust Fluid--in the USA.

The urea solution is injected into the exhaust and transformed to N[H.sub.3] for the SCR reactions over the SCR catalyst. The preferred transformation of urea is through thermolysis to ammonia directly, and through hydrolysis of iso-cyanic acid (HNCO) to ammonia as an intermediary step [2, 3]:

* Thermolysis: [CO(N[H.sub.2]).sub.2] [right arrow] N[H.sub.3] + HNCO

* Hydrolysis: HNCO + [H.sub.2]O [right arrow] N[H.sub.3] + C[O.sub.2]

The spray preparation of the injected urea solution is thought to have a significant influence on the evolution of these urea decomposition processes in the exhaust stream. Until recent years, there has been only limited information about spray quality as the AUS-32 fluid temperature in the injector increased to typical hot levels in exhaust-mount urea injection applications.

As a first step to this improved understanding, high speed video images were taken at the injector tip at different fluid temperatures. These images revealed significant changes in spray structure and also fluid delivery as temperatures approached the atmospheric boiling point of AUS-32 (104[degrees]C) [4, 5]:

* The main spray structure exhibited a higher vapor fraction as the temperature increased to and surpassed the atmospheric boiling point of the fluid--this included an expansion of the spray cone angle;

* At elevated temperatures, a non-negligible amount of spray and flow activity occurred after injector closing.

In an effort to clarify the behavior of the spray and the evolution of the spray structure when injecting into a hot environment, as opposed to a room temperature environment, backlighting spray measurements were performed using a hot air flow bench, and using a prototype injector that permitted active heating of the fluid inside the injection unit [6]. An example of the spray structure evolution from a liquid spray to a flash boiling condition at an exhaust temperature of 490[degrees]C is shown in figure 1.

This testing confirmed the observations made at lower ambient air temperatures, including the post-closing spray activity at high temperature. This was subsequently attributed to the evacuation of the injector "sac volume" - the volume between the injector sealing band and the orifice disk exit--figure 2.

In addition to spray diagnostics, it is important to understand the evolution of the injector flow behavior at the elevated temperatures under consideration. Traditional instantaneous flow rate (or "shot-to-shot") measurement devices are typically based on the so-called Zeuch method, which was developed into commercial measurement equipment in the 1980s by Ono Sokki for diesel injection systems [7] - see figure 3.

The measurement principle is based on injection of the fluid into a closed vessel and monitoring the pressure rate increase, which is proportional to the instantaneous volumetric flow rate as given by the following relation:

dV/dt = V/K dP/dt


V = vessel volume

P = vessel pressure

K = bulk modulus of the fluid.

This approach worked quite well with high pressures typical of diesel applications, and also with gasoline direct injection applications where the pressure increase in the vessel during the injection event was negligible with respect to the system supply pressure. However with low pressure applications typical of port injection or exhaust aftertreatment injection systems, the decrease in delta pressure became significant and reduced the representativity of the flow measurement.

An additional drawback with the Ono Sokki device (and other similar devices working on the same principle) was that injection into a closed liquid volume was also not representative of the real world where injection is into a gaseous medium. These devices were also unable to provide reliable flow information in the case of flash boiling sprays, where the dual phase nature of the spray rendered a determination of the bulk modulus of the fluid in the injection volume difficult or even impossible. A final disadvantage to these devices is the inability to perform simultaneous flow measurements and spray diagnostics.

The University of Perugia has developed a device that, while still relying on the Zeuch principle, performs the analysis on a fluid volume on the supply side of the injector. This approach recovers the representativity of a spray injecting into a gaseous environment, and also provides the opportunity to perform simultaneous flow and spray diagnostics.

Measurements were made with the University of Perugia measurement device, hereafter referred to as "dINJ", including validation of the dINJ in comparison with the traditional Zeuch principle flow meters. That testing culminated in measurements under flash boiling conditions with injection of AUS-32--these are summarized in figure 4 [8].

The measurements permitted a clearer understanding of the instantaneous flow behavior under flash boiling conditions. In particular, further insights were obtained to explain a generally observed increase in the mass flow per injection pulse at the flash boiling condition, however the precise details concerning this increase still remained hypothetical.

As mentioned, a unique advantage to the dINJ device is the ability to conduct simultaneous diagnostics of injector flow and spray behavior - this paper presents the results of those continued investigations. One objective of these investigations was to illuminate further the flow evolution leading to the flash boiling flow increase, and to study the effects of the oscillatory pressure behavior on the spray.

A description of the test setup and procedures is followed by a presentation of the simultaneous imaging and instantaneous flow measurement results. A discussion follows examining in more detail the injector opening and closing transients. The paper closes with a summary of the conclusions that can be drawn from the present testing campaign.


The dINJ instantaneous flow meter is shown schematically in figure 5.

The principle of operation consists of stabilizing the fluid volume in the upstream chamber (green) and then activating the test unit for the desired injection pulsewidth. The fluid temperature in the vessel is monitored, and the fluid pressure evolution is acquired with a pressure transducer. Once the test injection is complete, a separate filling valve is actuated to refill and restabilize the pressure in the fluid chamber before the next test cycle.

A typical pressure signal and derived volumetric flow rate is shown in figure 6 for one complete cycle.

The raw injection rate signal is subject to some oscillation/noise after the opening and closing transients--this is due to a number of factors related to the vessel geometry, relative location of the test injector and filling valve, and noise on the pressure derivative signal. In order to mitigate the effects of this noise, an 8th order Bessel low pass filter is applied with a cutoff frequency typically in the range of 1 - 2 kHz.

In addition, a Siemens SITRANS MASSFLO Coriolis meter is installed upstream of the filling valve to provide a mean injected volume signal for comparison with and calibration of the instantaneous measurements - this measurement allows for a calculation of the V/K term in the Zeuch methodology.

A fully built Actively Heated Reductant Dosing Unit (AH RDU) prototype equipped with an "SVS" injector--an injector design with a secondary heating coil for inductive heating of the injected fluid - was used for this testing. The RDU was equipped with a thermocouple in the fluid path at the injector tip ([T.sub.tipfluid]), as well as a thermocouple welded to the skin of the valve body, also at the injector tip ([T.sub.tip])--see figure 7.

The RDU was operated with injection into open air in a laboratory environment (i.e. room temperature with minimal air flow)--figure 8.

Imaging was performed with a high speed video camera at an acquisition rate of 10000 frames per second and an exposure time of 3 [micro]s per frame. The video resolution was 432x820 pixel; backlighting was provided by a continuous LED source. The reported videos of 13 ms duration captured a single injection event at a pulsewidth of 5.0 ms--the video image was triggered by the injector actuation pulse. Each image in the video was time stamped from the trigger point.

The heater controls used a heater control unit (iHCU) in conjunction with a controller operating in a LabVIEW environment.

For all the testing, the basic procedure for the flow measurement was similar, and consisted of the following:

* A warmup phase to stabilize the system--typically 1000 injections;

* An acquisition phase for the instantaneous rate measurements - typically 200 or 500 injections;

* A steady-state operation phase to acquire the Siemens SITRANS MASSFLO Coriolis meter signal for mean injected volume--typically 1000 injections.

The AH RDU flash boiling testing was performed under ambient air injection conditions, however at two fluid temperatures--Table 1.

It is noted here that the fluid was heated prior to entry at the dINJ to a temperature of 40 [degrees]C for the elevated fluid temperature as this was felt to be more representative of conditions when flash boiling may be taking place in an actual engine application.


Opening Time

Figure 9 compares the injector opening event for the temperature conditions of 30 [degrees]C and 130 [degrees]C.

According to the dINJ flow meter measurement, the effective opening time--i.e. the time between the actuation signal and the first movement of the injector needle (flow initiation)--at 130 [degrees]C was on the order of 0.840 ms, compared to 0.574 ms at the lower temperature, or a delay of 0.266 ms. This was consistent with findings (not published publicly) during a previous analysis of flow reductions versus temperature using standard flowmeters whereby a large portion of the flow reduction was attributed to the reduction in magnetic force brought about by presumably hotter injector coil. This had an effect on the measured opening time of the injector as illustrated in figure 10. (NB: the opening time parameter in this figure adds the flight time of the needle to the opening time described above from the dINJ data trace - thus the absolute values are not identical).

Leading Edge Appearance

Figure 11 compares the moment of first appearance of the spray for the temperature conditions of 30 [degrees]C and 130 [degrees]C.

At the 130 [degrees]C condition, the initial appearance of droplets was a little more difficult to detect than at 30 [degrees]C--figure 12 shows a more contrasted and zoomed in view to illustrate the selection of the image.

The conclusion drawn from these images was that in both the temperature cases, the spray required 1.1 ms to traverse the 10 mm from the orifice disk to the exit plane of the RDU - see figure 13.

Maximum Flow Condition

Figure 14 shows the condition of the 130 [degrees]C spray 1.2 ms after the flow maximum detected by the dINJ, compared with the same condition 1.2 ms after the detected flow minimum.

At this level of magnification, no significant difference was found in the condition of the sprays at these two presumed opposite mass flow conditions. This was also true when inspecting the images subsequent to the supposed minimum flow condition, under the hypothesis that the minimal pressure at the minimum flow condition may have induced a reduced exit velocity from the orifice, delaying the transit to the visible exit plane of the RDU--figure 15 shows the images at 100 and 200 [micro]s after the minimum flow section was expected to issue from the RDU.

At a qualitative level with these macro images, there was no detection of the substantial variation in flow measured by the dINJ.

Thus, it appeared that the large oscillations in measured flow during the injection pulsewidth did not have a noticeable (macro-level) impact on the liquid portion of the spray exiting the RDU.

Discussion-Injector Opening

With the current test campaign, no flow-related impact was measurable on the liquid fraction of the spray visible with the backlighting. During this testing, no information was available on the vapor fraction of the spray.

The simultaneous analysis of flow and spray did however contribute to improved understanding of how a dynamic flow increase was occurring at the elevated temperatures, with one possible sequence of events described by the following:

1. An initial outrush of fluid and flash boiling gas into the sac volume (which was evacuated of fluid after the prior injection)-corresponded to the initial flow overshoot due to the high delta pressure gradient across the sealing band; this would be accompanied by the displacement of gas out of the sac volume;

2. The two-phase flash-boiling flow leads to a buildup of liquid in the sac volume, similar to the charging of an accumulator, and progressively increases the hydraulic pressure behind the orifice holes and reduces the instantaneous flow across the sealing band--the reduction of flow into the sac volume and the lagging pressure buildup for flow across the orifice holes result in a strong undershoot in the measured flow;

3. The flow ultimately stabilizes at the injector static flow defined by the orifice holes with pressurized liquid in the sac volume - this flow is of course much lower than the static flow defined by the seal band gap area which defined the initial flow overshoot; the two-phase flash-boiling flow now occurs exclusively outside the injector.

Thus the dINJ measures the additional flow charging the sac volume, but not the effective flow that occurs when the sac volume is evacuated after injection closing.

Injector Closing--Sac Volume Evacuation

Figure 16 shows the condition of the 30 [degrees]C and 130 [degrees]C sprays at the very end of the injection (still flowing) and at the point where flow has stopped.

In figure 17, the same comparison is presented at a time delay of 1.2 ms after the dINJ indicated the flow had stopped.

The spray exiting the RDU at 130 [degrees]C has previously been attributed to the evacuation of the sac volume. This seemed to be confirmed in the form of the flow measurement signal which was exhibiting a higher amplitude oscillation compared to 30 [degrees]C, but without any indication that a real exit flow was being measured. The absence of a flow signal was consistent with the evacuation which was occurring after closing of the injector and isolation of the sac volume from the dINJ detection volume.


Simultaneous instantaneous flow rate measurements and backlit high speed spray videos were performed on fully assembled RDUs using a new approach based on the Zeuch method.

Measurements were made with an actively heated RDU under liquid spray conditions ([T.sub.tipfluid] = 30 [degrees]C) and flash boiling conditions ([T.sub.tipfluid] = 130 [degrees]C).

A comparison of the video imaging and the flow measurements permitted the evaluation of the following hypothesis concerning the mechanism of a previously identified rich flow shift occurring during the injector opening event under flash boiling conditions, and specifically within the first 0.6 ms of the opening:

1. An initial outrush of fluid and flash boiling gas into the sac volume (which was evacuated of fluid after the prior injection) --corresponding to the initial flow overshoot due to the delta pressure gradient across the sealing band and also to the displacement of gas out of the sac volume;

2. A buildup of liquid in the sac volume, similar to the charging of an accumulator, building up hydraulic pressure behind the orifice holes and reducing the instantaneous flow across the sealing band--the reduction of flow into the sac volume and the lagging pressure buildup for flow across the orifice holes result in a strong undershoot in the measured flow;

3. The flow ultimately stabilizes at the injector static flow defined by the orifice holes with pressurized liquid in the sac volume - this flow is of course much lower than the static flow defined by the seal band gap area which defined the initial flow overshoot.

During this process, the imaging indicated that there was minimal impact on the behavior of the liquid spray fraction exiting the RDU. The video imaging seemed to provide further confirmation of a sac volume evacuation after injector closing under flash boiling conditions (as opposed to injector bounce).


[1.] Liickert, P., Arndt, S., Duvinage, F., Kemmner, M., Binz, R., Storz, O., Reusch, M., Braun, T., Ellwanger, S., "The New Mercedes-Benz 4-Cylinder Diesel Engine OM654--The Innovative Base Engine of the New Diesel Generation", 24th Aachen Colloquium Automobile and Engine Technology, 2015

[2.] Koebel, M., Elsener, M., and Kleemann, M., "Urea-SCR: A promising technique to reduce NOx emissions from automotive diesel engines." Catalysis Today, 335-345, 2000.

[3.] Schaber, P., Colson, J., Higgins, S., Dietz, E., Thielen, D., Anspach, B., and Brauer, J., "Study of the urea thermal decomposition (pyrolysis) reaction and importance to cyanuric acid production." American Laboratory, August 1999, pp. 13-21.

[4.] van Vuuren, N. and Sayar, H., "High Speed Video Measurements of a Heated Tip Urea Injector Spray," SAE Technical Paper 2012-01-1747, 2012, doi:10.4271/2012-01-1747.

[5.] van Vuuren, N., "High Speed Video Measurements of a High Temperature Urea Injector Spray--Comparison of Spray Evolution in Water and AUS-32," SAE Technical Paper 2013-01-2527, 2013, doi:10.4271/2013-01-2527.

[6.] van Vuuren, N., Brizi, G., Buitoni, G., Postrioti, L. et al., "Experimental Analysis of the Urea-Water Solution Temperature Effect on the Spray Characteristics in SCR Systems," SAE Technical Paper 2015-24-2500, 2015, doi:10.4271/2015-24-2500.

[7.] Takamura, A., Fukushima, S., Omori, Y., and Kamimoto, T., "Development of a New Measurement Tool for Fuel Injection Rate in Diesel Engines," SAE Technical Paper 890317, 1989, doi: 10.4271/890317.

[8.] Postrioti, L., Caponeri, G., Buitoni, G., and van Vuuren, N., "Injection Rate Measurement of a SCR Injector Operating in Flash-Boiling Conditions," FISITA Technical Paper F2016-ESYH-013, 2016


AUS-32 - Aqueous Urea Solution - 32%

AC RDU - Air-Cooled Reductant Dosing Unit

AH RDU - Actively-Heated Reductant Dosing Unit

CFD - Computational Fluid Dynamics

DEF - Diesel Exhaust Fluid

[H.sub.2]O - Water

HNCO - Iso-Cyanic Acid

[N.sub.2] - Nitrogen (Molecular)

N[H.sub.3] - Ammonia

N[O.sub.x] - Nitrogen Oxides

RDU - Reductant Delivery Unit

SCR - Selective Catalytic Reduction

SVS - Selectively Variable Spray

Nic Van Vuuren

Continental Automotive Systems US Inc.

Lucio Postrioti, Gabriele Brizi, and Federico Picchiotti

Universita degli Studi di Perugia

(1.) Numbers in brackets refer to references listed at the end of this report.

Table 1. Test conditions

Injection event duration     [t.sub.inj]         5.0     ms
Injection frequency          [f.sub.inj]        10       Hz
Fluid                                           AUS-32
Fluid pressure               [P.sub.fluid]     600       kPag
Acquisition time resolution                      1       [micro]s
Fluid temperature            [T.sub.tipfluid]   30, 130  [degrees]C
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
Title Annotation:selective catalytic reduction
Author:Van Vuuren, Nic; Postrioti, Lucio; Brizi, Gabriele; Picchiotti, Federico
Publication:SAE International Journal of Engines
Article Type:Report
Date:Dec 1, 2017
Previous Article:Statistical Approach on Visualizing Multi-Variable Interactions in a Hybrid Breakup Model under ECN Spray Conditions.
Next Article:Development of an Innovative Combustion Process: Spark-Assisted Compression Ignition.

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