Printer Friendly

Development of SLD Capabilities in the RTA Icing Wind Tunnel.


Safety is naturally a top priority in the aviation industry. When an aircraft flies in upper cloud layers, supercooled water droplets suddenly freeze and, within minutes, can form a centimeter thick layer of ice on critical components such as the engines, rotors, wings and sensors. In order to obtain certification for these components and the whole aircraft, manufacturers must demonstrate that sufficient protective measures have been implemented and that the formation of ice will not endanger flight safety in any way.

In order to develop, test and certify products, the aviation industry therefore urgently requires high-performance facilities where suitable icing tests can be carried out. Rail Tec Arsenal (RTA), an internationally active independent research and test institute for rail and road vehicles, aircraft and technical equipment, has taken on this challenge and built up new icing capabilities with funding from the Austrian Ministry for Transport, Innovation and Technology (BMVIT) via the Austrian Research Promotion Agency (FFG)

The large Climatic Wind Tunnel (CWT) depicted in the Appendix can be upgraded to an Icing Wind Tunnel (IWT) using an additional icing rig mounted on the existing CWT contraction section (see Figure 1). The IWT speed can be increased by integrating an additional contraction nozzle (see Figure 2). This new equipment allows cloud simulations to be carried out according to EASA CS-25 [1, Appendix C] or FAA Title 14 CFR Part 25 [2, Appendix C].

According to the EURICE Final Summary Report [3], the current EASA CS-25 [1, Appendix C] does not include all possible conditions present in the real atmosphere. Freezing rain in particular has long been underestimated as a potential hazard condition (see [4, p. 1]). The simulation of Supercooled Large Droplet (SLD) icing conditions such as freezing rain and freezing drizzle in IWTs has thus become more important in recent years. Since these facilities offer the possibility to test components under reproducible conditions, RTA started to develop a method to simulate SLDs as freezing rain and freezing drizzle according to EASA CS-25 [1, Appendix O] or FAA Title 14 CFR Part 25 [2, Appendix O] in 2015.


The definition of SLD icing conditions in the form of freezing drizzle and freezing rain can be found in EASA CS-25 [1, Appendix O]. The droplet distribution and Liquid Water Content (LWC) for SLDs are defined as shown in Figures 3, 4, 5, 6.


Two modifications were required for running SLD simulations in the RTA IWT: suitable nozzles and modifications of the water supply system compatible with the existing system.

Nozzle Selection

Before selecting the most suitable nozzles it was important to define two main parameters: the Median Volumetric Diameter (MVD) and the flow rate which was used for the theoretical Total Water Content (TWC) in the IWT. The following approach was used for the theoretical analysis of the MVD. The distributions given in Figures 3 and 4 have bimodal characteristics. Each distribution may be closely represented using a sum of two Gaussian distributions. This results in five parameters for each type of rain: MVD and standard deviation of small droplets, MVD and standard deviation of large droplets and mass fraction of small droplets related to total mass. Table 1 below shows the parameters for the types of rain as defined in EASA CS-25 [1, Appendix O].

The LWC according to EASA CS-25 [1, Appendix O] is based on the water flow rate versus the IWT cross-section (theoretical TWC). The flow rate can thus be calculated as follows (see Equations 1. 2, 3, 4):

[[m.sub.Nozzle_S] = [LWC x ][phi] x Vt x A / n x [epsilon]] (1)

[[m.sub.Nozzle_S] = [LWC x (1-[phi]) x Vt x A / n x [epsilon]] (2) with

[epsilon] = [LWC/TWC]] (3) and

[phi] = LW[C.sub.S] / LW[C.sub.S] + LW[C.sub.L] (4)

This approach requires homogeneous LWC distribution in the test section. Furthermore, e was estimated at 0.9 based on the small droplet (see EASA CS-25 [1, Appendix C]) calibration of the IWT according to SAE ARP5905 [5], which was carried out in 2013/2014. A value of 1.0 was assumed for large droplets because of the missing effects of air influence (expansion and evaporation) when using a single-medium nozzle. The values (shown in Table 1) provided the basis for selecting the most suitable nozzles for the existing icing system. The selection process produced the following results:

* For MVD < 200 [micro]m: pneumatic atomizer nozzles

* For MVD > 200 [micro]m: single-medium nozzles

A feasibility study [6] carried out at the RTA IWT with the new configuration showed that conditions for "Freezing Drizzle MVD < 40 [micro]m and "Freezing Rain MVD < 40 [micro]m" cannot be met in the current IWT setup. The reason for this is that there are no nozzles available that can produce droplets of approx. 560 [micro]m or 100 [micro]m MVD at the extremely low mass flow rate required. Research was therefore focused on the SLD simulation of freezing rain > 40 [micro]m MVD and freezing drizzle > 40 [micro]m MVD because of the promising results achieved during the mentioned feasibility study 2015.

Water Supply System Modification

The IWT was built to produce droplets according to EASA CS-25 [1, Appendix C] and thus required a modification of the water supply system. In the basic design the water circuits (see Figure 7) were all connected to one water supply so that the nozzles could only be operated with one temperature setting.

With the SLD modification, a second water supply was added to provide capability for two different temperature settings in parallel. The bimodal droplet spray needs two different nozzle types with different air and water settings for achieving various MVDs (a single-medium nozzle needs only a water pressure set point). On the one hand, the water supplied to the pneumatic atomizer nozzle must be preheated because smaller droplets tend to freeze before reaching the test section. On the other hand, the water supplied to the single-medium nozzle must be precooled in order to ensure that the larger drops in particular are supercooled when reaching the test section.

Future research will be necessary to determine the correct water temperature for freezing rain > 40 [micro]m at certain ambient temperature, wind speed and residence time.


The challenge in simulating SLD icing conditions is the different behavior of large droplets (> 100 [micro]m) compared to droplets < 50 [micro]m. Gravitational influences as well as droplet breakup and high inertia are some of the major characteristics of SLDs [7, p. 1].

Droplet Trajectory in the IWT

A 2D CFD analysis of droplet trajectories was conducted in collaboration with the Institute of Aviation at the FH JOANNEUM (University of Applied Sciences) in Graz. This analysis was designed to estimate the trajectories of particles with different diameters and to determine the air velocity needed to transport large droplets (100 [micro]m to 1200 [micro]m) to the test section of the IWT. The particle trajectories of droplets with variable diameters were thus simulated at varying air velocities and two different temperatures. The analysis was done using ANSYS CFX15.0 [8]. The IWT at RTA has an asymmetrical contraction nozzle with a length of about 10 m, a height of 2.5 m and a width of 3.5 m at the outlet. The 2D simulation was performed for the geometrical center of the nozzle. The velocity field in the contraction nozzle and the particle tracks were exported in order to estimate potential droplet breakup. Figure 8 shows the mesh for the CFD analysis of the contraction nozzle and the region where the water droplets were injected.

The simulation was conducted at temperatures of -2[degrees]C and -18[degrees]C. Gravity and the buoyancy force were considered and the shear stress transport turbulence model was used. Heat transfer was neglected.

Air velocity in the inlet varied from 15 m/s to 41 m/s in order to obtain a velocity range from 30 m/s to 80 m/s in the test section, which is the upper wind speed limit of the RTA's IWT in Vienna.

Particle injection was simulated using a point cone with a cone angle of 20[degrees]. Injection velocity was measured with a high speed camera (Phantom v711 [9]) at velocities of 7 m/s up to 10 m/s (see Figure 9). The CFD calculation involved simulating 1000 particles with a specific particle diameter distribution (minimum diameter 100 [micro]m, maximum diameter 1200 [micro]m, mean diameter 650 [micro]m) for each of the eleven nozzles. A particle mass flow rate of 0.07 1/min was used based on flow-meter measurements. The following assumptions were made: no droplet breakup, no heat transfer, no evaporation of droplets, drops do not interact with each other, one-way coupling between the air and the droplets, and when a droplet hits a wall it stays where it impinged.

The CFD analysis showed that the test section temperature has a negligibly small influence on the particle tracks. The trajectories of water droplets with diameters greater than 1000 [micro]m at wind speeds of 30 m/s and 60 m/s are shown in Figures 10 and 11, respectively, where the color represents the velocity of the droplets. At velocities below 40 m/s, the larger droplets (> 800 [micro]m) from the lower injection nozzles impinge on the bottom of the contraction nozzle. The particle tracks of droplets with a diameter range of 1000 to 1200 [micro]m at 80 m/s are displayed in Figure 12. It can be concluded that the droplets from the injection spray located at the top impinge on the wall of the contraction nozzle. This phenomenon was also observed during tests in the IWT.

Secondary Droplet Breakup in the IWT

Another focus of the CFD calculation was to analyze potential droplet breakup under the given flow conditions. Droplets with diameters from 100 to 1200 [micro]m were considered to break up as they become sufficiently distorted due to aerodynamic forces [7, p. 1]. The critical condition where breakup occurs is when the aerodynamic drag force is equal to the surface tension force. For an estimation of the breakup, effects of heat and mass exchange as well as transient effects were neglected [10, p. 45]. Hence the Weber number (see Equation 5), which describes the ratio of pressure force to the surface tension force, and the

Reynolds number (see Equation 6), which represents the ratio of inertia force to viscous force, both describe the droplet conditions adequately [7, p. 2]. In a horizontal wind tunnel the Weber number is calculated with the initial droplet diameter and the relative velocity between the droplet and main airflow (see Equation 5) [11. p. 60]. As the Weber number does not include fluid viscosity which can influence the mode of breakup, the Ohnesorge number (see Equation 7) was also considered [12, p. 1]. According to Wierzba [11, p. 59] the bag breakup type starts to occur at Weber numbers of 11 < We < 14, at relatively small Ohnesorge numbers of < 0.1.

[mathematical expression not reproducible] (5)

[mathematical expression not reproducible] (6)

[mathematical expression not reproducible] (7)

The Ohnesorge numbers calculated for droplet diameters from 100 to 1200 urn were in a range of 0.006 up to 0.021. According to Majithia et al., the experimentally determined critical Weber number for the bag breakup regime in constant and pulsed air flows is even higher than 14 for such small Ohnesorge numbers [12, pp.5-6]. An analysis of a single particle with a diameter of 1160 [micro]m is shown in Figure 13. It displays the trajectory of a particle injected at spray bar 5 (height of y = 0.11 m above the geometrical center) at a main airflow velocity of about 80 m/s. It also indicates that the droplet velocity does not reach the velocity of the airflow when it passes the test section (the plotted fluid velocity describes the velocity of the main airflow along the droplet trajectory). The Weber number increases to about 10 directly after the droplet is injected. Subsequently, it declines to about 5 and then again rises to 11 near the end of the contraction of the IWT. The jagged area in the plot results from numerical inaccuracies from the calculation of the relative velocity (airspeed relative to drop speed).

The Weber number of a 1160 urn droplet injected at spray bar 5 is plotted as a function of time in Figure 14.

The maximum Weber numbers obtained for several particle tracks with diameters from 100 to 1200 urn at various velocities are displayed in Figure 15 for spray bar 3 (y= -0.66 m below the geometrical center, spray bar 5 (y = 0.11 m above the geometrical center) and spray bar 7 (y = 0.87 m above the geometrical center). At wind speeds of 60 m/s and less the particles do not reach the critical area of We > 11. The Weber number is also dependent on where the droplet is injected (height from the center and injection angle) and its exact trajectory.

The influence of the injection angle on the Weber number can be seen in Figure 16, the critical values are only exceeded at larger injection angles (red area).

In conclusion, the trajectory simulation showed that all droplets reach the test section at velocities higher than 60 m/s and that droplet break up is possible according to the break up analysis. It also showed that the breakup depends on where the droplet is injected and its injection angle.


The new SLD capabilities of the RTA IWT were validated through an investigation of the droplet diameter distribution and the LWC. First investigations on cloud uniformity were also performed, but did not produce satisfactory results. Nevertheless, improvements of cloud uniformity are essential for accurate SLD simulations and further research is in progress.

Measurements of Droplet Diameter Distribution

A homogenous LWC uniformity is very challenging to achieve for bimodal droplet diameter distributions [13, p. 39]. In a first step, research was focused on achieving the necessary bimodal distribution. Further steps to improve LWC uniformity will also be necessary. The droplet diameter distribution was measured using two different instruments: the Spraytec system from Malvern Instruments [14] and the Cloud Imaging Probe (CIP) from Droplet Measurement Technologies [15].

The Spraytec system is able to measure droplet diameters in a range of 1-2500 [micro]m. It uses the technique of laser diffraction by detecting the intensity of light scattered as a laser beam passes through a spray. A great advantage of the instrument was the long distance between laser and detector (1 m), resulting in a large measuring zone. But this setup did not allow measurements beyond a main air velocity of 50 m/s.

The CIP is capable of detecting particles in a size range of 12.5-1550 [micro]m with a resolution of 25 [micro]m. Because of this only droplets larger than 100 [micro]m were taken into account. The arm width of the instrument used is 70 mm, which limits the area of observation. It uses a collimated laser beam and projects shadow images of particles passing through onto a linear array of 64 photodetectors to create particle images. Capability for air velocities up to 300 m/s is the main advantage of the CIP.

Freezing Rain MVD > 40 [micro]m

The Malvern device was set up at the geometrical center of the test section and approx. 13 m from the spray bars. Droplet distribution was measured with all spray bars active (192 normal sprays and 72 SLD sprays). A velocity range of 30-47 m/s was tested. The measurements revealed a change in droplet size distribution with increasing wind speed. From 30-47 m/s the MVD increased from about 220 [micro]m to about 550 [micro]m (see Figure 17) because of changes in the trajectories of the large drops (see also Figures 10, 11, 12). Between 45 m/s and 47 m/s, the error between the droplet diameter distributions seems to be negligible. Figure 18 shows the measurement result at 47 m/s compared to the EASA CS-25 [1, Appendix O] requirements. It should be noted that droplets with a diameter of 1550 [micro]m were detected.

The CIP was used to compare the measurements with the Malvern device and for measurements at higher velocities. The CIP was used to measure droplets with diameters of more than 1200 [micro]m up to a speed of 80 m/s. One disadvantage of the CIP was that droplets were excluded by the evaluation software when the laser beam of an end diode was covered. Hence some of the droplets > 1200 [micro]m were not taken into account. The end diode rejection can be turned off but this bears the risk of detecting a wrong diameter. The results of the measurement are shown in Figure 20, which displays the particle images at flow velocities of 60 m/s and 80 m/s. The detected droplet sizes are very similar. A comparison of the droplet diameter distributions obtained from the Spraytec and CIP (only > 100 [micro]m) measurements is displayed in Figure 19 for various velocities.

The droplets had an almost ideal spherical shape in all particle images. The images obtained by the CIP showed that at certain velocities only a particular range of droplet diameters passed the measurement section, which demonstrates the poor dispersion and the small effect of turbulence on the large droplets (see Figure 21). Here only spray bar 7 with single medium nozzles (SLD) was in operation. At 40 m/s and 50 m/s the small drops were not detected because their drop trajectories were above the sample area of the instrument.

Moreover, droplet breakup seems to be negligible, since droplet diameter distribution does not significantly change at higher velocities. In fact, particle images obtained from these measurements also showed a spherical shape with no indications of deformations typical of bag breakup [16, pp. 893-894].

Freezing Drizzle MVD > 40 [micro]m

The droplet diameter distribution for freezing drizzle MVD > 40 [micro]m was measured during the technical feasibility study carried out in 2015 [6] using the Malvern instrument. It was produced using the same pneumatic atomizer spray nozzles as used for EASA CS-25 [1, Appendix C]. Figure 22 shows the results from the measurement in comparison to the EASA CS-25 [1, Appendix O] requirements. As the maximum droplet diameters are smaller than those for freezing rain, droplet breakup will be precluded.

Measurements of Freezing Rain Particle Trajectory

Three different measurements were performed to validate the particle tracks computed. A vertical bar with a paper strip (which changes color when wet) was mounted in the test section. This measurement visualized the height at which the droplets leave the nozzle.

The computed trajectories were also validated using the CIP from Droplet Measurement Technologies. The instrument was placed 2.05 m from the contraction nozzle exit, at varying heights. With single rows of injection nozzles (at the same height) activated, the wind speed was increased until no droplets passed the measuring area. The trajectories obtained from the CFD simulation and the points where the droplets were measured (marked with crosses) are shown in Figure 23. The results from the measurements demonstrate that the CFD analysis is in good conformity with the actual trajectories of the droplets.

A third measurement was carried out under actual icing conditions at a temperature of -13[degrees]C and a main airflow velocity of 80 m/s. One circuit with 12 injection nozzles (single-medium nozzles only) of a spray bar (0.11 m below the geometrical center) was activated. An accretion grid was placed at a distance of 1.5 m from the contraction nozzle outlet. The detected ice accretion is shown in Figure 24.

The results of the measured peaks of ice accretion on the grid (black points) and the calculated impingement height (500 [micro]m to 800 [micro]m droplets; gray line) are shown in Figure 25. This can also be seen from the 2D calculation shown in Figure 10, 11, 12.


RTA began its research project aimed at achieving SLD conditions by performing appropriate calculations and carrying out a feasibility study in June 2015. The promising results from the feasibility study led to further theoretical investigations and adaptations of the existing icing rig. An analysis of compatible nozzles available on the market showed restrictions in terms of creating SLD conditions for freezing rain and freezing drizzle < 40 [micro]m MVD (very small quantity of water versus necessary droplet sizes up to 1000 [micro]m). These two SLD conditions were not investigated in more detail due to time restrictions. Both theoretical and experimental investigations were carried out for freezing rain and freezing drizzle > 40 [micro]m MVD, with the focus being placed on the simulation of freezing rain. The conditions for freezing drizzle were already achieved with the nozzles used for EASA CS-25 Appendix C. A complete calibration has not been performed yet, but is expected to be possible with reasonable effort due to the use of well-known nozzles.

The theoretical investigations and measurements using different instruments and methods showed that it is possible to transport large droplets with a diameter of up to 1500 [micro]m to the test section in RTA's horizontal icing wind tunnel at a distance of [greater than or equal to]11 m from the nozzle outlet to simulate freezing rain. In addition to determining the minimum velocity of approx. 50 m/s required to transport droplets > 1000 [micro]m, the project also involved investigating the phenomenon of secondary droplet breakup up to 80 m/s. Measurements using two different instruments have confirmed that the droplet distribution required by EASA CS-25 Appendix O can be achieved with acceptable accuracy at the airspeed above 45 m/s.

Investigations were also carried out on the issue of cloud uniformity, but did not produce satisfactory results. Further relevant considerations and research are planned for the future. Validation of the theoretical analysis of the LWC is planned for the end of 2016. There have also been theoretical and experimental investigations concerning the supercooling of SLDs in the RTA icing wind tunnel, the results of which will be published in a separate document.


[1.] EASA, "Certification Specifications and Acceptable Means of Compliance for Large Aeroplanes CS-25, Appendices C and O," Amendment 18, 12 Jun 2016.

[2.] FAA, "Title 14 - Aeronautics and Space, Chapter 1 - FAA, Department of Transportation, Subchapter C - Aircraft, Part 25 - Airworthiness Standards: Transport Category Airplanes, Special Federal Aviation Regulation No. 109, Subpart 1, Appendices C and O," GPO - Electronic CFR, 2016.

[3.] CIRA et al., "European Research on Aircraft Icing Certification - EURICE," Final Summary Report, CORDIS, Sep. 2008, URL:

[4.] Bernstein, B., Ratvasky, T., Miller, D., and Mc Donough, F., "Freezing rain as an In-Flight Icing Hazard," NASA/TM-2000-210058, USA, Jun. 2000.

[5.] SAE International Aerospace Recommended Practice, "Calibration and Acceptance of Icing Wind Tunnels," SAE Standard ARP5905[TM], Reaf. Sept. 2015.

[6.] Ferschitz, H., Bucek, O., and Wannemacher, M., "Capabilities of the New Icing Wind Tunnel Vienna," Presented at SAE 2015 International Conference on Icing of Aircraft, Engines, and Structures, Prague, Czech Republic, Jun. 2015.

[7.] Luxford, G., Hammond, D. and Ivey, P., "Role of Droplet Distortion and Break-Up in Large Droplet Aircraft Icing," 42nd AIAA Aerospace Sciences Meeting and Exhibit, AIAA-2004-411, Reno, NV, Jan. 2004.

[8.] ANSYS, CFX (Version 15.0), Computer Software, Canonsburg, PA, Nov. 2013.

[9.] AMETEK, "Phantom v-Series Cameras Brochure,", accessed Oct. 2016.

[10.] Wozniak, G., "Zerstaubungstechnik: Prinzipien, Verfahren, Gerate, First Edition," (Berlin, Springer, 2002), 9 and 45-50., doi: 10.1007/978-3-642-55835-1.

[11.] Wierzba, A., "Deformation and break-up of liquid drops in a gas stream at nearly critical Weber numbers," Experiments in Fluids 9: 59-64, 1990, doi :10.1007/BF00575336.

[12.] Majithia, A., Hall, S., Harper, L. and Bowen, P., "Droplet Breakup Quantification and Processes in Constant and Pulsed Air Flows," ILASS08-4-4, Como Lake, Italy, Sep. 2008.

[13.] Hervy, F., "New SLD Icing Capabilities at DGA Aero-engine Testing," SAE Technical Paper 2011-38-0086, 2011, doi: 10.4271/2011-38-0086.

[14.] Malvern Instruments, "Spraytec System Brochure," http :// accessed Oct. 2016.

[15.] Droplet Measurement Technologies, "Cloud Imaging Probe (CIP) Brochure," accessed Oct. 2016.

[16.] Chou, W.-H., Faeth, G, "Temporal properties of secondary drop breakup in the bag breakup regime," International Journal of Multiphase Flow 24 (6): 889-912, 1998, doi: 10.1016/S0301-9322(98)00015-9.


Hermann Ferschitz

+43 2568081 301

CFD -                   Computational Fluid Dynamics
CFR -                   Code of Federal Regulations
CIP -                   Cloud Imaging Probe
CWT -                   Climatic Wind Tunnel
EASA -                  European Aviation Safety Agency
FAA -                   Federal Aviation Administration
IWT -                   Icing Wind Tunnel
RIA -                   Rail Tec Arsenal
SLD -                   Supercooled Large Droplets
MVD -                   Median Volumetric Diameter [[micro]m]
LWC -                   Liquid Water Content [g/[m.sup.3]]
TWC -                   Total Water Content [g/[m.sup.3]]
Oh -                    Ohnesorge number [1]
Re -                    Reynolds number [1]
We -                    Weber number [1]
d -                     Droplet diameter [[micro]m]
p -                     Density [kg/[m.sup.3]]
[eta] -                 Fluid dynamic viscosity [Pas]
[sigma] -               Standard deviation [[micro]m]
Y -                     Surface tension [N/m]
[v.sub.ret] -           Relative droplet velocity [m/s]
[m.sub..sub.Zozzle] -   Mass flow of the nozzle [g/s]
[phi] -                 Mass fraction small droplets of the restrictive
                        nozzle [1]
Vt -                    Velocity in the test section [m/s]
A -                     Area in the test section [[m.sup..sup.2]]
[epsilon] -             Freeze-out coefficient [1]
n -                     Number of operating nozzles [1]


a - Air

w - Water

d - Droplet

S - Droplet distribution with smaller droplets (< 40 [micro]m)

L - Droplet distribution with large droplets (> 40 [micro]m)


Hermann Ferschitz, Michael Wannemacher, Otto Bucek, Florian Knobel, and Wolfgang BreitfuB RTA
Table 1. Parameters for SLDs

                         Total          Small droplets
                      MVD     [phi]     MVD      [sigma]
                  ([micro]m)   (-)   ([micro]m)

Freezing drizzle      <40     0,85      19         11
                      >40     0,31      24         14
Freezung rain         <40     0,69      14         10
                      >40     0,19      20         13

                      Large droplets
                     MVD       [sigma]

Freezing drizzle     100          73
                     182         147
Freezung rain        559         447
                     644         532
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:supercooled large droplet, Rail Tec Arsenal
Author:Ferschitz, Hermann; Wannemacher, Michael; Bucek, Otto; Knobel, Florian; BreitfuB, Wolfgang
Publication:SAE International Journal of Aerospace
Article Type:Report
Date:Sep 1, 2017
Previous Article:Designing a Hybrid Electric Powertrain for an Unmanned Aircraft with a Commercial Optimization Software.
Next Article:Towards Standardising Methods for Reporting the Embodied Energy Content of Aerospace Products.

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters