applied

sciences
Article
Assessment of Runway Surface Conditions by British
Pendulum Testing under the Global Reporting Format
Winter Conditions
Jean-Denis Brassard *, Audrey Beaulieu, Marc Mario Tremblay and Gelareh Momen

Anti-Icing Materials International Laboratory, Université du Québec à Chicoutimi,

Saguenay, QC G7H 2B1, Canada
* Correspondence: jean-denis1_brassard@uqac.ca

Abstract: Poor braking performance on runways during winter is one of the most significant factors
causing runway excursions. The International Civil Aviation Organization (ICAO) proposed a
standardized method, named the Global Reporting Format (GRF), to establish runway surface
conditions based on human observers. The GRF includes the description of most winter conditions.
This work aimed at reproducing the GRF winter conditions in the laboratory, and to monitor the effect
of the surface conditions, both without surface treatment and with surface treatment using liquid
runway de-icing products (RDP) in both de-icing and anti-icing modes, using the British Pendulum
Tester (BPT). The reproduction of the GRF winter conditions was possible in the laboratory since the
BPT results correlated well with the GRF ratings. The worst conditions retained snow on ice. Wet
conditions obtained with water and RDPs (potassium formate KFO, potassium acetate KAC and
hybrid glycol-KAC HYB) behave similarly, indicating that RDPs alone on a runway for prevention are
not any more dangerous than just water. The range from wet to dry, equivalent to a British Pendulum
Citation: Brassard, J.-D.; Beaulieu, A.; Number (BPN) from ~40 to ~50, was considered as good conditions on the runway. All the RDPs
Tremblay, M.M.; Momen, G. tested in the study improved the BPN from untreated conditions. In most of the conditions, the RDPs
Assessment of Runway Surface improved the BPN to values in the range of favorable conditions. KFO and KAC reacted almost in
Conditions by British Pendulum the same way for each condition. However, the HYB was better in conditions with ice. The obtained
Testing under the Global Reporting results demonstrated that simulated laboratory winter conditions could be used to determine the
Format Winter Conditions. Appl. Sci. efficiency of RDPs in both de-icing and anti-icing modes.
2022, 12, 9646. https://doi.org/
10.3390/app12199646 Keywords: winter maintenance; runway de-icing products; landing and takeoff; snow; ice; British
Academic Editors: Yanhu Mu and Pendulum Tester; Global Reporting Format
Pengfei He

Received: 14 September 2022

Accepted: 23 September 2022
1. Introduction
Published: 26 September 2022
One of the most significant factors causing runway excursions is poor braking per-
Publisher’s Note: MDPI stays neutral
formance. Loss of control on the runway accounted for 15.5% of all the aviation accidents
with regard to jurisdictional claims in
that occurred between 2010 and 2019 [1]. In 2019 alone, those accidents cost the air trans-
published maps and institutional affil-
portation industry $4B directly. Most of the accidents occur during winter since the runway
iations.
is contaminated by snow, slush, ice, brine or water, rendering the surfaces slippery [2].
They all have a negative effect on braking performance, significantly reducing the friction
between aircraft tires and the runway surface, inhibiting maneuverability [1–7].
Copyright: © 2022 by the authors.
Different methods exist in the literature to evaluate braking performance based on
Licensee MDPI, Basel, Switzerland. the state of the surface. Most of these are American Society for Testing and Materials
This article is an open access article (ASTM, West Conshohocken, PA, USA) standards used to measure the surface friction
distributed under the terms and characteristics. These tests can be divided in four classes: (i) fixed tests, (ii) braking tests,
conditions of the Creative Commons (iii) contact tests, and (iv) non-contact tests.
Attribution (CC BY) license (https:// Fixed tests are the most popular. There are a few apparatuses which allow the friction
creativecommons.org/licenses/by/ between the tire and runway to be measured directly. These apparatus are capable of
4.0/). performing lateral force tests (ASTM E670) [8], fixed slip tests (ASTM E2340) [9], and locked

Appl. Sci. 2022, 12, 9646. https://doi.org/10.3390/app12199646 https://www.mdpi.com/journal/applsci

Appl. Sci. 2022, 12, 9646 2 of 14

wheel tests (ASTM E274) [10]. They all work following the same principle where a host
vehicle drags a tribometer (here, a tire) maintaining a constant speed, thus allowing the
friction and load force to be measured. The friction coefficient is then determined using the
ratio of the friction and load forces.
The second class of tests consists of braking tests which are done by: (i) measuring
the stopping distance (ASTM E445) [11], or (ii) measuring the deceleration rate (ASTM
E2101) [12]. The stopping distance measurement is conducted by displacing a vehicle at
65 km/h on a contaminated road followed by locking of the wheels. The distance d is
then measured until the point where the vehicle comes to a complete stop and the friction
coefficient is then determined. The greater the distance, the lower the friction coefficient will
be. In measuring the deceleration rate, on the other hand, a vehicle is displaced at a speed
between 32 and 48 km/h followed by locking of the wheels to determine the deceleration.
The friction coefficient is then calculated as a function of the difference between the initial
and final speed. The greater the difference, the higher the friction coefficient will be. The
application of the two latter classes is aimed at real-life operational cases and may not be
considered for a laboratory test.
The third class consists of the contact tests. They are those tests which imply contact
with a limited part of the runway. Contrarily to the other two classes, these tests could be
easily performed under laboratory conditions, as well as in natural outdoor operational
conditions. The main apparatus in this category is the British Pendulum Tester (BPT, ASTM
E303) [13]. The BPT is a dynamic test which works on the principle of the Sharpy test,
whereby a rubber slider in contact with the surface allows for skid evaluation. The British
Pendulum Tester allows for the evaluation of surface texture. The greater the surface
roughness, the greater the difference will be. The result is given as the British Pendulum
Number.
More recently, non-contact tests have been proposed to the industry: (i) electro-optic
test, and (ii) the circular texture meter (ASTM E2157) [14]. The electro-optic test measures
the pavement texture properties at high-speed, using ultra high-frequency laser triangu-
lation mounted on a car. On the other hand, the circular texture meter is used at a fixed
position with a laser-displacement sensor. With this method, the surface macrotexture is
determined using the calculated roughness parameters.
Although existing ASTM tests provide relevant information regarding the texture of
the runway, which is important for the braking procedure of aircraft, more information is
still needed to confirm the nature of contaminants on the surface. Efforts have been made
to use weather and flight data to estimate surface conditions at landing using data analysis,
machine learning and fusion and model correlation [15–17]. A few sensors exist in the
market that are mounted in the runway to predict the type of contaminant present [18–20].
Hashimoto et al. developed a new sensor which operates using light scattering
principles to accurately measure the snow cover properties [18,19]. The concept was proven
using different snow samples having various grain size distribution, liquid water content,
density and thickness. The sensors are still in the development stage and could be adapted
to other types of precipitation or contaminants. Marchetti et al. demonstrated that Raman
spectroscopy could be used to differentiate the fluids (aircraft and runway products), to
identify the evolution of freezing temperatures with water dilution and the concentrations
of active compounds in each fluid [20]. The presented methodology could be mounted with
a surface texture sensor to accurately determine the nature of the contaminant and its effect
on the runway surface. Despite the existence of such tools being constantly developed and
improved, the only recognized method to-date of evaluating surface conditions remains
the human eye, as specified in the Global Reporting Format document [21].
Since November 2021, the International Civil Aviation Organization (ICAO, Montreal,
QC, Canada) has implemented the New Global Reporting Format (GRF) for runway surface
conditions at all airports. The main idea is to mitigate the risk of runway excursions by
enabling a harmonized assessment and reporting of runway surface conditions, and by an
improved flight crew assessment of takeoff and landing performance.
Appl. Sci. 2022, 12, 9646 3 of 14

Among the work carried out to implement this global standard, Transport Canada
published the Advisory Circular (AC, Ottawa, ON, Canada) No. 300-019 [22]. The GRF
principle implies that at any time when water, snow, slush, ice, or frost are present on
an operating runway, the airport operator evaluates the runway surface conditions. This
evaluation yields a RunWaY Condition Code (RWYCC) and a description of the runway
surface, which the flight crew may use to calculate airplane performance. The airport
operator’s best evaluation of the runway surface quality is based on this format, which is
based on the type, depth, and coverage of contaminants. However, other relevant factors
should be considered as well, and the changes in circumstances should be notified as soon
as possible.
The Runway Condition Assessment Matrix (RCAM) is the technique by which an
airport or aerodrome operator generates a preliminary RWYCC for each runway-third if
there is water, snow, slush, ice, or frost present on the runway surface. Descriptions of
the winter conditions are included directly in this matrix and allow their reproduction in
controlled laboratory conditions. Airport maintenance teams, as well as runway de-icer
product manufacturers need procedures to evaluate the impact of RDPs when used in those
conditions. The idea is to gather relevant information on the skid conditions of the surface
to know whether the product is being used in an anti-icing mode (before the precipitation)
or in the de-icing mode (after the precipitation).
The most relevant method used to evaluate the skid resistance of pavements and run-
ways is based on their microtexture, and hence their frictional properties can be assessed
using the BPT [23–28]. Thanks to its light weight, the BPT is portable and does not necessi-
tate surface preparation or external energy [23]. The BPT is well-documented and plenty
of research proves that this method gives sufficient information to accurately evaluate the
state of the surface. Little research has been undertaken to establish a relationship between
the British Pendulum Number (BPN) arising from the BPT and the surface coefficient of
friction (µ) [23,29]. Additionally, the BPT has been used to evaluate road and runway
microstructure by several authors. Their main results are presented in the Table 1.

Table 1. Main Results obtained using BPT with winter conditions.

Ref. Authors Condition BPN Note

Dry 70
Rough superhydrophobic
Eriskin et al. [24] Wet 60
coating on a coarse surface
Iced 48
Dry 74 Coarse surface
Rodin et al. [27] Wet 64 Complete ice cover
Iced by soaking 22
Dry 11 Coarse surface
Sajid et al. [28]
Wet with de-icer 60

Eriskin et al. [24] used the BPT to evaluate the impact of superhydrophobic coatings
on different pavements. Their main findings indicated that a superhydrophobic coating
may delay ice formation and then enhance surface friction resulting in a higher BPN [24].
Rainwater et al. [26] and Sajib et al. [28] used the BPT to evaluate the performance of chem-
ical de-icers used on roads. Both studies have obtained results during natural field testing
campaigns showing that chemical de-icers could lead to effective results by significantly
increasing the surface BPN compared to the untreated surface condition. However, since
the tests were conducted in natural conditions, it was hard to obtain repeatable results.
Rodin et al. [27] used the BPN to evaluate the skid resistance of previously used concrete
slabs subjected to ice. They obtained results showing that in both de-icing mode and
anti-icing mode, the de-icer, namely calcium chloride, helped increase the BPN significantly.
However, the way the ice was formed on the concrete slab was not representative of how it
occurs under natural conditions, since the ice is formed by soaking the concrete substrate
Appl. Sci. 2022, 12, 9646 4 of 14

completely in water at cold temperatures leading to ice formations in the cavities and not
on the top of the substrate. Finally, in the anti-icing mode, the chemical de-icer is applied
before exposure to ice which contradicts the literature facts, suggesting that de-icers should
be applied after exposure to ice. Based on a recent literature survey on the BPT, there is no
work that includes the use of different winter contaminants in conjunction with common
runway de-icing products [25–28].
With the introduction of the new GRF and the fact that there is a gap in the evaluation
of RDPs, the main goal of this study was to reproduce the cited winter conditions in the
laboratory and on concrete cement substrates, to present a correlation between the GRF
and the BPT, and to then propose a complete testing procedure of liquid RDPs in those
conditions. The obtained results will be useful in determining preferable conditions for the
use of RDPs.

2. Materials and Methods

Experiments were carried out to evaluate the impact of winter contaminants on the
skid resistance of a runway, as well as the impact of common liquid RDPs used in anti-
icing and de-icing modes. All the measurements were conducted under controlled cold
conditions in the UQAC’s snow chamber. Concrete cement paving stones of dimensions
10 cm × 20 cm × 5 cm were used as test substrates, as shown in Figure 1a (Permacon
Heritage Universal Paver, Anjou, Canada). These paving stones were selected as substrates
since their surface roughness and absorption characteristics correlated well with the actual
concrete surfaces found at most Canadian airports. Although the surface roughness is less
than that of some airport surfaces, it serves as a point of comparison for the conditions
targeted in the study. Each stone was cleaned using tap water and dried at room temper-
ature. Figure 1b shows the BPT setup with the concrete substrate mounted for test. The
pavements and testing apparatus were conditioned at the test temperature in the cold room
Appl. Sci. 2022, 12, x FOR PEER REVIEW ◦ 5 of 15
(−15 to −5 ◦ C) for 24 h prior to testing. The skid measurements were undertaken using
the BPT according to ASTME303 [13], and the BPN was obtained.

Figure 1. (a) British Pendulum Tester (BPT) in operation in the cold room and (b) the concrete cement
test substrate.
Figure 1. (a) British Pendulum Tester (BPT) in operation in the cold room and (b) the concrete ce-
ment test substrate.
The smaller the BPN, the more the surface will slide and is considered risky. Con-
versely, the bigger the BPN, the less the surface will slide. The BPT is primarily used to
assess microtexture. This test is commonly supposed to have a slip speed of 10 km/h, and
microtexture is thought to rule in low-speed applications [30]. Prior to each testing day, the
Appl. Sci. 2022, 12, 9646 5 of 14

apparatus was levelled and the zero verified. The contact surface was also verified using a
ruler to ensure that the slider was in contact with 125.0 ± 1.6 mm [13].
Seven conditions were excerpted from the RCAM and reproduced in the labora-
Figure
tory 1. (a)
[22]. A British Pendulum Tester
plan summarizing the(BPT)
testsinis operation
presentedin in
theFigure
cold room and (b)
2, with theschematics
their concrete ce-
ment test substrate.
shown in Figure 3.

Figure 2. Test plan of the seven GRF icing conditions [22] selected for this study.
Figure 2. Test plan of the seven GRF icing conditions [22] selected for this study.
The first test consisting of the dry “as-is” condition is considered as the reference value,
which is equivalent to a RWYCC of 6 (no treatment prior to testing).
The second condition represented the wet condition, where the concrete was fully
wetted with water at room temperature, with a RWYCC of 5. The first two conditions
were conducted at 20 ◦ C. Following the RCAM, the next conditions were reproduced using
lab-harvested natural dry snow particles.
The third test refers to the “removed snow” condition conducted at −15 ◦ C; snow was
sifted and compacted on the stone with a low pressure and then removed using a scraper.
In this condition, all the pores of the concrete were filled with snow and a thin layer of
snow remained on the surface. This reproduced a compacted-snow condition, with an
equivalent RWYCC of 4.
The fourth test condition consisted of dry snow which was obtained using a 1 mm
sift to deposit approximately 4 mm of snow. The thickness of snow was measured at
five different positions using a metallic ruler. The snow was not compacted during the
deposition process, so it was easily removed during the passage of the pendulum slider.
The equivalent RWYCC is 3.
The same methodology was used to obtain the fifth test condition, wet snow, except
that the snow was exposed to a simulated freezing drizzle precipitation for a period of six
minutes, drizzled at a targeted intensity of 8.5 g/dm2 ·h. This condition also corresponds to
a RWYCC of 3. The latter two conditions were conducted at −10 ◦ C.
Appl.
Appl.Sci.
Sci.2022,
2022,12,
12,x 9646
FOR PEER REVIEW 6 of
6 of1514

Figure 3. Schematic of the different winter conditions presented in the GRF [22] and a close-up view
Figure 3. impact
of their Schematic of the
on the different
surface winter conditions presented in the GRF [22] and a close-up view
texture.
of their impact on the surface texture.
The sixth condition consisted of ice obtained by simulating a freezing drizzle precip-
itation,
The with
first atest
nozzle of median
consisting volumetric
of the dry “as-is” diameter
condition of 115 µm (Nozzle
is considered as#650017) being
the reference
drizzled at a targeted intensity of 8.5 g/dm 2 ·h at −5 ◦ C for six minutes. After the precipita-
value, which is equivalent to a RWYCC of 6 (no treatment prior to testing).
tion,The
a waiting
secondtime of 10 min
condition was required
represented to ensure
the wet that the
condition, contaminant
where was completely
the concrete was fully
frozen. In this case, the equivalent RWYCC is 1.
wetted with water at room temperature, with a RWYCC of 5. The first two conditions
Finally, theat
were conducted seventh
20 °C. test condition
Following theconsisted
RCAM, the of snow
next on ice, reproduced
conditions using the same
were reproduced us-
icing
ing conditions as
lab-harvested used dry
natural in the sixth
snow condition with the addition of 3 mm of sifted snow.
particles.
TheTheequivalent
third testRWYCC
refers to isthe
0. “removed snow” condition conducted at −15 °C; snow was
sifted and compacted on the stone withsix
Each condition was reproduced times,
a low and for
pressure andeach
thensubstrate
removedthe BPNa was
using mea-
scraper.
sured five times, resulting in 30 data results per condition.
In this condition, all the pores of the concrete were filled with snow and a thin layer of
snow Three
remained generic liquid
on the RDPs This
surface. currently used byaairport
reproduced maintenance
compacted-snow teams were
condition, tested
with an
during the process
equivalent RWYCC of 4. [31]. All RDPs were prepared in the laboratory using deionized water
andThefromfourth
solid test
commercial
condition salts. The first
consisted of RDP
dry snow consisted
which of was w/w potassium
50% obtained using aformate
1 mm
(KFO, HCOOK). At this dilution, its freezing point is nearly − 60 ◦ C. The second RDP
sift to deposit approximately 4 mm of snow. The thickness of snow was measured at five
consisted
different of 50% w/w
positions usingpotassium
a metallicacetate (KAC,
ruler. The snowCHwas3 COOK), also with the
not compacted corresponding
during the depo-
freezing point of nearly −60 ◦ C. The third product consisted of a hybrid RDP. Its compo-
sition process, so it was easily removed during the passage of the pendulum slider. The
nents consisted of 25% w/w of liquid propylene glycol (PG, C3 H8 O2 ) and 25% w/w KAC.
equivalent RWYCC is 3.
The initial freezing point of this RDP is approximately −48 ◦ C. For each de-icing and anti-
The same methodology was used to obtain the fifth test condition, wet snow, except
icing condition, 3 mL of RDP was used, depositing 1 mL using a syringe at three different
that the snow was exposed to a simulated freezing drizzle precipitation for a period of six
positions around the centre of the pavement substrate. The testing protocols for de-icing
minutes, drizzled at a targeted intensity of 8.5 g/dm2. h. This condition also corresponds
and anti-icing are detailed in Figure 4. The de-icing conditions consisted of depositing
to a RWYCC of 3. The latter two conditions were conducted at −10 °C.
 freezing point of nearly −60 °C. The third product consisted of a hybrid RDP. Its compo-
nents consisted of 25% w/w of liquid propylene glycol (PG, C3H8O2) and 25% w/w KAC.
The initial freezing point of this RDP is approximately −48 °C. For each de-icing and anti-
icing condition, 3 mL of RDP was used, depositing 1 mL using a syringe at three different
Appl. Sci. 2022, 12, 9646 positions around the centre of the pavement substrate. The testing protocols for de-icing 7 of 14
and anti-icing are detailed in Figure 4. The de-icing conditions consisted of depositing the
RDP after the icing/snowing precipitations. The anti-icing conditions consisted of depos-
iting theRDP
the RDPafter
before
thethe icing/snowingprecipitations.
icing/snowing precipitations. The
In each condition,
anti-icing a waiting
conditions time of of
consisted
20 min ensured that the RDP melted the contaminants before measuring the BPN.
depositing the RDP before the icing/snowing precipitations. In each condition, a waiting
time of 20 min ensured that the RDP melted the contaminants before measuring the BPN.

Figure 4. De-icing and anti-icing protocols.

Figure 4. De-icing and anti-icing protocols.
3. Results
Table 2 shows the summary of the BPN results obtained for different test conditions
tested in anti-icing and de-icing modes.

Table 2. BPN results obtained in the RWYCC conditions.

Condition Product Mode BPN

Dry
none - 50 ± 3
(RWYCC6)
Water - 38 ± 1
Wet KFO - 40 ± 2
(RWYCC5) KAC - 36 ± 1
HYB - 39 ± 1
Untreated - 33 ± 3
De-icing 48 ± 1
KFO
Anti-icing 48 ± 1
Removed Snow
(RWYCC4) De-icing 40 ± 1
KAC
Anti-icing 38 ± 2
De-icing 47 ± 1
HYB
Anti-icing 41 ± 1
Appl. Sci. 2022, 12, 9646 8 of 14

Table 2. Cont.

Condition Product Mode BPN

Untreated - 23 ± 2
De-icing 32 ± 1
KFO
Snow Anti-icing 45 ± 1
(RWYCC3) De-icing 40 ± 1
KAC
Anti-icing 43 ± 1
De-icing 31 ± 1
HYB
Anti-icing 34 ± 1
Untreated - 19 ± 2
De-icing 45 ± 1
KFO
Anti-icing 35 ± 1
Wet Snow (RWYCC3) De-icing 40 ± 1
KAC
Anti-icing 33 ± 1
De-icing 32 ± 1
HYB
Anti-icing 34 ± 1
Untreated - 17 ± 1
De-icing 48 ± 2
KFO
Anti-icing 37 ± 1
Ice
(RWYCC1) De-icing 47 ± 1
KAC
Anti-icing 30 ± 1
De-icing 42 ± 1
HYB
Anti-icing 34 ± 1
Untreated - 11 ± 2
De-icing 42 ± 1
KFO
Anti-icing 49 ± 1
Snow on ice (RWYCC0) De-icing 48 ± 1
KAC
Anti-icing 39 ± 1
De-icing 42 ± 1
HYB
Anti-icing 44 ± 1

Firstly, the main winter conditions prescribed by the GRF were reproduced in the
laboratory and the obtained BPN values were correlated with the RWYCC ratings. Since
the RWYCC has been established as a function of the state of the runway, by taking into
consideration the runway friction index (RFI), it is inherent that the results should correlate.
These results are presented in Figure 5. The first condition is the dry condition, with a
RWYCC of 6, while the evaluated BPN is as high as 50 ± 3, indicating that the surface
skid is very low. Therefore, this can be considered the safest runway condition. This value
is much lower than those available in the literature [24,25,27]. This is explained well by
the difference between the surface’s macro- and microtexture. The pavement tested is less
coarse than those used in other research [25,27].
 consideration the runway friction index (RFI), it is inherent that the results should corre-
late. These results are presented in Figure 5. The first condition is the dry condition, with
a RWYCC of 6, while the evaluated BPN is as high as 50 ± 3, indicating that the surface
skid is very low. Therefore, this can be considered the safest runway condition. This value
is much lower than those available in the literature [24,25,27]. This is explained well by
Appl. Sci. 2022, 12, 9646 the difference between the surface’s macro- and microtexture. The pavement tested is9less of 14

coarse than those used in other research [25,27].

55

50

45

40
BPN 35

30

25

20

15

10

0
0 1 2 3 4 5 6
GRF rating (RWYCC)

Figure 5. BPN as a function of the GRF RWYCC.

Figure 5. BPN as a function of the GRF RWYCC.
In the wet condition (RWYCC of 5), the BPN reduced significantly to 38 ± 1. This value
corresponded
In the wettocondition
a reduction of 24% when
(RWYCC compared
of 5), the to that obtained
BPN reduced with to
significantly dry
38conditions.
± 1. This
Rodincorresponded
value et al. [27] andtoEriskin et al. [24]
a reduction of 24%observed the same decrease
when compared when the
to that obtained surface
with dry con-was
wet, with 30% lower and 25% lower values, respectively. This condition
ditions. Rodin et al. [27] and Eriskin et al. [24] observed the same decrease when the sur- remains secure on
the runway. Further, the removed snow condition corresponded
face was wet, with 30% lower and 25% lower values, respectively. This condition remains to the compacted snow
condition
secure as described
on the in the GRF,
runway. Further, the with
removeda RWYCCsnow of 4 for both,
condition and a measured
corresponded to theBPNcom- of
33 ± 3.snow
pacted The snow and as
condition wet snow conditions,
described in the GRF, corresponding
with a RWYCC to aofRWYCC
4 for both, of and
3, presented
a meas-
a BPN
ured BPNof 19
of 33± 2± 3.
with
Thewetsnow snow
andand wet23snow± 2 conditions,
for snow-only, indicating that
corresponding the wet snow
to a RWYCC of 3,
conditions lead to more slippery states. The wet snow condition
presented a BPN of 19 ± 2 with wet snow and 23 ± 2 for snow-only, indicating that the wet was close to the slush
condition
snow which lead
conditions corresponded to a RWYCC
to more slippery states.ofThe
2. Slush
wet snowrepresents snowwas
condition supersaturated
close to the
slush condition which corresponded to a RWYCC of 2. Slush represents snow In
with water that should be easily evacuated when compressed with the hand. our case,
supersatu-
the water could not be removed from the snow, unless a high
rated with water that should be easily evacuated when compressed with the hand. In pressure was applied, soour
the
condition
case, was considered
the water could not be as removed
wet snow.from The next condition
the snow, unless corresponded
a high pressure to icewas
cover, with a
applied,
RWYCC of 1, which resulted in a BPN of 17 ± 1, indicating that
so the condition was considered as wet snow. The next condition corresponded to ice the surface is quite slippery.
It corresponded
cover, with a RWYCC to a reduction
of 1, which ofresulted
66% from in athe
BPNreference.
of 17 ± 1,Rodin et al. that
indicating [27] the
observed
surfacea
reduction of approximately 70% in their ice conditions. The BPN is
is quite slippery. It corresponded to a reduction of 66% from the reference. Rodin et al. [27] even more reduced
when the asnow
observed is added
reduction to the ice layer,70%
of approximately corresponding to a RWYCCThe
in their ice conditions. of 0,
BPNandisrepresenting
even more
the worst condition to be encountered on the runway.
reduced when the snow is added to the ice layer, corresponding to a RWYCC It has been confirmed with a BPN
of 0, and
of 11 ± 2, close to the lower limit of the pendulum. In this case,
representing the worst condition to be encountered on the runway. It has been confirmed the dry snow acts as a
lubricant on ice, rendering the surface even more slippery and hence more dangerous. In
with a BPN of 11 ± 2, close to the lower limit of the pendulum. In this case, the dry snow
order to validate the results, a linear fitting has been added to the graph. Most of the data
acts as a lubricant on ice, rendering the surface even more slippery and hence more dan-
were slightly over the curves except for the snow and wet snow conditions which were
gerous. In order to validate the results, a linear fitting has been added to the graph. Most
under. The obtained equation is

BPN = (5.5 ± 0.7)RWYCC + (10.2 ± 2.4) (1)

The intercept was not forced to 0 since at the worst condition, a BPN of 11 was still
obtainable. The Pearson correlation coefficient (r) of 0.96 indicates that the relation is
strongly positive, validating the results. Finally, the coefficient of determination (R2 ) of 0.93
also indicated that more than 90% of the data fits with this linear assumption.
It is obvious that most of the laboratory-reproduced conditions resulted in BPN data
which correlated well with the RWYCC. By applying Sabey et al.’s formula, it is possible to
Appl. Sci. 2022, 12, 9646 10 of 14

estimate the value of the coefficient of friction (µ). The empirical relationship is presented
in Equation (2) as follows:
3BPN
µ= (2)
(330 − BPN)
The calculated values are presented in the Table 3.

Table 3. Comparison between the CRFI values of friction coefficient (µ) and CRFI.

Corresponding Friction In the Range

Condition RWYCC BPN CRFI Range [22]
Coefficient (µ) [29] (<10%)
Dry 6 50 0.54 >0.4 yes
Wet with water 5 38 0.39 >0.4 yes (3%)
Removed snow 4 33 0.33 0.35–0.39 yes (6%)
Snow 3 23 0.22 0.30–0.34 no (27%)
Wet snow 3 19 0.18 0.30–0.34 no (40%)
Ice 1 17 0.16 <0.19 yes
Snow on ice 0 11 0.10 <0.19 yes

The CRFI range presented in the GRF document [22] is also presented in Table 2.
A criterion of <10% of difference at the targeted range has been applied to verify if the
measured values are in the range. All values are within range except for snow and wet
snow, which are much lower than the targeted range. If the values were considered with a
RWYCC of 2, these results could be considered comparable to the range. This may signify
that the simulated conditions are more dangerous than the specified values, indicating
that the tests will be more conservative than the reality, adding an additional safety factor.
However, the Sabey’s equation is based on measurements and may also induce an error
greater than 20%. Therefore, the estimated values of µ should be only used as a comparison
and not as a direct value of µ.
In order to deepen the analysis of this correlation, data were collected in different
wet conditions which could be close to those encountered on the runway. Visually, the
wet condition should also include the presence of liquid runway de-icing products in
their initial concentration and also in their diluted concentration. So, in addition to the
wet condition with water at 4 ◦ C, data were collected when the concrete was wetted by
KFO, KAC and HYB RDPs at −5 ◦ C. It is not unusual for aircraft to encounter a runway
completely wetted by RDPs. It is the case that when the RDP is applied in anti-icing mode,
before precipitation, and also when RDPs melt the icy contaminants. The overall results
are presented in the Table 1, as well as in Figure 6. As discussed above, the corresponding
BPN to dry and wet with water conditions are 50 ± 3 and 38 ± 1, respectively. This slight
decrease in BPN indicates that water acts as a lubricant. However, the BPN is not further
decreased when RDPs are present on the concrete. KFO, KAC and HYB have a BPN of
40 ± 2, 36 ± 1 and 39 ± 1 respectively, all in the range of the wet-with-water conditions,
indicating that they can be all treated as wet conditions.
Those dry and wet conditions will then be considered as the good conditions, as stated
by the RWYCC. Further analysis using the upper and lower BPN values obtained with
those conditions will help to classify the values and to help understand if they are efficient
in the GRF winter conditions.
Knowing that the simulated winter conditions correlate well with the GRF RWYCC,
they were then reproduced in the same settings described in the previous sections, but using
three generic RDPs: KFO, KAC and HYB, in de-icing mode (D) and anti-icing mode (A).
The main results are presented in Table 1, as well as in Figure 7. In the removed snow
conditions presented in Figure 7a, all the RDPs significantly improved the BPN in both
de-icing and anti-icing mode. KFO showed the best improvement, with a BPN of 48 ± 1 in
both modes. Hybrid RDP also showed an improvement with a BPN of 47 ± 1. However,
in all the presented cases, all the RDPs improved the surface conditions to at least equal
or better than the wet conditions, indicating that the conditions are acceptable. In snow
Appl. Sci. 2022, 12, 9646 11 of 14

conditions, as presented in Figure 7b, the most efficient RDP is KAC. It allowed for an
increase in the BPN from 23 ± 2 on untreated counterparts, to 40 ± 1 and 43 ± 1, in de-icing
and anti-icing modes, respectively. In this condition, KFO was more efficient in anti-icing
mode than in de-icing mode. It had an increased BPN of 45 ± 1 in anti-icing and only 32 ± 1
in de-icing modes. The worst results were obtained with the HYB, even with a slightly
increased BPN. There is a lack of information regarding snow conditions in the literature.
However, we can observe that in anti-icing mode, the snow is melted more easily than in
de-icing mode. In anti-icing mode, the concrete allowed the RDP to spread, allowing the
Appl. Sci. 2022, 12, x FOR PEER REVIEW 11 of 15
protection of a larger part of the concrete. In de-icing mode, the RDP penetrated the snow
by gravity, partially melting only a small part of the snow.

Figure 6. BPN obtained in wet condition as compared to the dry condition.

Figure 6. BPN obtained in wet condition as compared to the dry condition.
Under the wet snow conditions, as shown in Figure 7c, the KFO and KAC are better
in de-icing dry
Those mode andthanwetin conditions will then
anti-icing mode. KFObeincreased
considered the as BPN thetogood
45 ±conditions,
1 and KAC as to
stated by the RWYCC. Further analysis using the upper and
40 ± 1, while the untreated surface is at a BPN of 19 ± 2. The main difference compared lower BPN values obtained
with
to drythose
snow conditions
is due towill the help to classify
presence the values
of a slight and toofhelp
proportion understand
liquid water which if they are
could
efficient in the GRF winter conditions.
activate melting, while in the dry snow condition the product may pass through the snow
Knowing
without melting thatit.the simulated in
Conversely, winter conditions
anti-icing mode,correlate
since thewell RDP with
is onthethe
GRF RWYCC,
concrete, in
they were then condition,
the snow-only reproduceda in theofsame
part it is settings
melted. described
However, in inthewetprevious
snow, the sections,
water thatbut
using
arisesthree
fromgeneric
the snow RDPs: KFO,the
inhibits KAC and short-term
RDP’s HYB, in de-icing modeThe
efficiency. (D) and
HYBanti-icing
remains mode stable
(A). The main results are presented in Table 1, as well as in Figure
irrespective of the kind of snow or the mode used. Regarding the ice condition presented 7. In the removed snow in
conditions presented in Figure 7a, all the RDPs significantly
Figure 7d, all three RDPs are efficient in de-icing mode. KFO, KAC and HYB give BPNs improved the BPN in bothof
de-icing
48 ± 2, 47 and± anti-icing
1 and 42 ±mode. KFO showed
1, respectively, all inthe
thebest
range improvement, with a BPN
of the wet conditions. By of 48 ± 1 in
comparing
both
their modes.
results Hybrid RDP also counterparts,
to the untreated showed an improvement
the BPNs are with a BPN of
increased to 47 ± 1.than
more However,
200%.
in all the
When presented
used in de-icing cases, all the
mode, allRDPs
threeimproved the surface
RDPs sufficiently meltedconditions
the iceto atat
theleast equal
interface,
or better than
rendering the wetand
it rougher conditions,
thereforeindicating
less slippery. thatHowever,
the conditions are acceptable.
in anti-icing mode, even In snow
if the
conditions,
time of failure as presented
is not reached,in Figure 7b, time
i.e., the the most
takenefficient
for ice to RDP
startis forming
KAC. It on allowed for an
the concrete,
all threein
increase RDPs
the BPN onlyfrom
increased
23 ± 2 ontheuntreated
BPN to values rangingtofrom
counterparts, 40 ± 30 to 37.
1 and 43 ±The
1, inbest RDP
de-icing
in anti-icing
and anti-icingmode modes, was KFO, withIna this
respectively. of 37 ± 1.KFO
BPNcondition, Thewasphenomena
more efficient wereinnot exactly
anti-icing
the same
mode thanininthe snow mode.
de-icing on ice It condition presentedBPN
had an increased in Figure
of 45 ±7e.1 inThe BPN increased
anti-icing and onlyfrom 32 ±
111in±de-icing
2 to 42 ± 1, 48 ±
modes. The1 and
worst ± 1, inwere
42results de-icing mode,
obtained andthe
with 49 ± 1,
to HYB, 39 ±
even 1 and
with 44 ± 1
a slightly
in anti-icing
increased BPN. mode ThereforisKFO,
a lack KAC and HYB, respectively.
of information regarding snow In this particular
conditions in condition,
the literature. the
addition of snow in de-icing mode promotes better melting of
However, we can observe that in anti-icing mode, the snow is melted more easily than in the ice, with this impact even
de-icing mode. In anti-icing mode, the concrete allowed the RDP to spread, allowing the
protection of a larger part of the concrete. In de-icing mode, the RDP penetrated the snow
by gravity, partially melting only a small part of the snow.
 Appl. Sci. 2022, 12, 9646 12 of 14

more pronounced in anti-icing mode. HYB showed a better efficiency in icy conditions
than in snowy conditions, while the glycol complex is more efficient on snow than on
ice. Literature is scarce in this area of study, however, a few researchers have observed
Appl. Sci. 2022, 12, x FOR PEER REVIEW similar trends of improvements in the BPN, but with using other RDP molecules 12 ofsuch
15 as
chlorine-based products [26] and corn-derived polyols (sorbitol, mannitol, and maltitol)
added to traditional salt brine de-icer [28].

Figure 7. Skid
Figure resistance
7. Skid of different
resistance RDPs
of different usedused
RDPs in de-icing mode
in de-icing (D) and
mode (D) anti-icing mode
and anti-icing (A) under
mode (A) under
various
various icing conditions (a) Removed snow (b) snow (c) wet snow (d) ice and (e) snowice.
icing conditions (a) Removed snow (b) snow (c) wet snow (d) ice and (e) snow on on ice.

Under the wet

Overall, snow
it is conditions,
obvious that theas use
shown in Figure
of RDPs 7c, the
in both KFO and
de-icing andKAC are better
anti-icing modes
in de-icing
improves the BPN and consequently, the surface conditions. KFO was the best40RDP,
mode than in anti-icing mode. KFO increased the BPN to 45 ± 1 and KAC to
± 1, especially
while the untreated
in anti-icingsurface
modeisfor
at asnow
BPNandof 19ice
± 2. The main followed
conditions, differenceclosely
compared to dry
by KAC. Once
snow is duetotothe
applied theconcrete,
presenceKFO of a and
slight
KACproportion
helped the of liquid
melting water
of thewhich could activate
contaminants and kept
melting, while in
the surface theonly.
wet dry snow condition the product may pass through the snow without
melting it. Conversely,
This study could in be
anti-icing
extended mode, since
to solid RDPs theto
RDP is ontheir
observe the concrete,
efficiencyininthe snow-
both de-icing
onlyand anti-icing
condition, modes.
a part of it isFurther
melted.study is in-progress
However, in wet snow, withtheother
waternon-contact methods
that arises from the to
snow inhibits the RDP’s short-term efficiency. The HYB remains stable irrespective of the
kind of snow or the mode used. Regarding the ice condition presented in Figure 7d, all
three RDPs are efficient in de-icing mode. KFO, KAC and HYB give BPNs of 48 ± 2, 47 ± 1
and 42 ± 1, respectively, all in the range of the wet conditions. By comparing their results
to the untreated counterparts, the BPNs are increased to more than 200%. When used in
Appl. Sci. 2022, 12, 9646 13 of 14

determine the surface macrotexture to ensure the validity of the current apparatus. These
results will be published as a follow-up article. The results presented focus on only one
type of substrate, namely concrete cement. The results are comparative, however, and
allow some conclusions to be drawn about RDPs. Further studies are planned with the use
of asphalt concrete as well as other alternative materials.

4. Conclusions
The reproduction of the GRF winter conditions were possible in the laboratory. The
results obtained using the British Pendulum Tester with concrete cement paving stones
correlated well with RWYCC ratings. The worst condition remained snow on ice. Wet
conditions obtained with water and RDPs were in the same range, indicating that the use of
RDPs-only on a runway in prevention is not any more dangerous than just water. The range
of wet to dry, equivalent in BPN from ~40 to ~50, are considered as good conditions on the
runway. All the RDPs tested in the study improved the BPN from untreated conditions to
values that are acceptable in terms of skids in most of the conditions. KFO and KAC reacted
almost similarly for each condition. KFO is considered the most efficient, reaching the best
score in 80% of the conditions. However, HYB was better in conditions involving ice. The
obtained results show that the simulated laboratory winter conditions could be used to
estimate the laboratory efficiency of RDPs in both de-icing and anti-icing modes prior to
testing in the field. Numerous products could be compared cost-effectively using the same
conditions. Work is currently in-progress to validate the adaptability of the procedure to
solid and prewet-solid RDPs, which are also often used by airport maintenance teams. The
non-contact runway surface condition sensors will also be investigated and reported in
a follow-up article. These results have great potential to be used to improve the actual
standards in this field of study and to increase safety on the runway.

Author Contributions: Conceptualization, J.-D.B.; Data curation, A.B.; Formal analysis, J.-D.B.;
Funding acquisition, J.-D.B. and G.M.; Investigation, J.-D.B. and G.M.; Methodology, J.-D.B. and
A.B.; Project administration, M.M.T. and G.M.; Resources, M.M.T.; Supervision, J.-D.B.; Visualization,
J.-D.B.; Writing—original draft, J.-D.B.; Writing—review & editing, J.-D.B., M.M.T. and G.M. All
authors have read and agreed to the published version of the manuscript.
Funding: This research was funded jointly by the CRIAQ (O2HPA) and NSERC (Grant Number
CRDPJ 537834-18).
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: Not applicable.
Acknowledgments: We acknowledge the support of the Natural Sciences and Engineering Research
Council of Canada (NSERC). This research was conducted in support of the Consortium for Research
and Innovation in Aerospace in Québec (CRIAQ) and the Ministère de l’Économie et de l’Innovation
du Québec, and the support provided by Aéroports de Montréal, WPred and Nachur Alpine Solutions.
AB acknowledges MITACS for the research grant.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Boeing. Statistical Summary of Commercial Jet Airplane Accidents. Available online: https://www.boeing.com/resources/
boeingdotcom/company/about_bca/pdf/statsum.pdf (accessed on 19 January 2022).
2. Tuncal, A.; Suat, U.; Dursun, E. A Milestone to Enhance Runway Safety: The New Global Reporting Format. Rev. Investig. Univ.
Quindío 2021, 33, 168–178. [CrossRef]
3. Kornstaedt, L.; Lignee, R. Operational Landing Distances, A new standard for in-flight landing distance assessment. Safety 2010,
10, 1–5.
4. Procházka, J.; Kameník, M. Contaminated Runway Operations-Adverse weather. MAD-Mag. Aviat. Dev. 2013, 1, 3–7. [CrossRef]
5. Niu, Y.; Jiang, X.; Meng, F.; Wang, R.; Ju, G.; Zhang, S.; Menga, Z. Techniques and Methods for Runway Friction Measurement: A
Review of State-of-the-Art. IEEE Trans. Instrum. Meas. 2021, 70, 9510717. [CrossRef]
Appl. Sci. 2022, 12, 9646 14 of 14

6. Klein-Paste, A. Airplane braking friction on dry snow, wet snow or slush contaminated runways. Cold Reg. Sci. Technol. 2018, 150,
70–74. [CrossRef]
7. Brassard, J.-D.; Laforte, C.; Tremblay, M.M.; Volat, C. Runway Deicing Product Anti/Deicing Performance Assessment: Review and
Future Directions; 0148-7191; SAE Technical Paper; SAE: Warrendale, PA, USA, 2019.
8. ASTM E670-09(2020); Standard Test Method for Testing Side Force Friction on Paved Surfaces Using the Mu-Meter. ASTM: West
Conshohocken, PA, USA, 2015.
9. ASTM E2340/E2340M-11(2021); Standard Test Method for Measuring the Skid Resistance of Pavements and Other Trafficked
Surfaces Using a Continuous Reading, Fixed-Slip Technique. ASTM: West Conshohocken, PA, USA, 2021.
10. ASTM E274/E274M-15(2020); Standard Test Method for Skid Resistance of Paved Surfaces Using a Full-Scale Tire. ASTM: West
Conshohocken, PA, USA, 2020.
11. ASTM E445/E445M-88(1996); Standard Test Method for Stopping Distance on Paved Surfaces Using a Passenger Vehicle Equipped
with Full-Scale Tires. ASTM: West Conshohocken, PA, USA, 2019.
12. ASTM E2101-15(2020); Standard Test Method for Measuring the Frictional Properties of Winter Contaminated Pavement Surfaces
Using an Averaging-Type Spot Measuring Decelerometer. ASTM: West Conshohocken, PA, USA, 2015.
13. ASTM E303-22; Standard Test Method for Measuring Surface Frictional Properties Using the British Pendulum Tester. ASTM:
West Conshohocken, PA, USA, 1993; p. 5.
14. ASTM E2157-15(2019); Standard Test Method for Measuring Pavement Macrotexture Properties Using the Circular Track Meter.
ASTM: West Conshohocken, PA, USA, 2019.
15. Midtfjord, A.; Huseby, A.B. Estimating runway friction using flight data. In Proceedings of E-Proceedings of the 30th European
Safety and Reliability Conference and 15th Probabilistic Safety Assessment and Management Conference (ESREL2020 PSAM15),
Research Publishing Services, Venice, Italy, 1–5 November 2020.
16. Niu, Y.; Zhang, S.; Tian, G.; Zhu, H.; Zhou, W. Estimation for Runway Friction Coefficient Based on Multi-Sensor Information
Fusion and Model Correlation. Sensors 2020, 20, 3886. [CrossRef] [PubMed]
17. Midtfjord, A.D.; De Bin, R.; Huseby, A.B. A Machine Learning Approach to Safer Airplane Landings: Predicting Runway
Conditions using Weather and Flight Data. arXiv 2021, arXiv:2107.04010.
18. Hoshino, S.; Hashimoto, K.; Tateyama, K.; Harada, Y.; Sato, Y.; Ikeda, Y.; Uchikata, I.; Ohmae, H.; Miyake, T.; Kanda, A. Snow and
ice monitoring technique for the contaminated runway. In Proceedings of the AIAA Scitech 2020 Forum, Orlando, FL, USA, 6–10
January 2020; p. 1685.
19. Hashimoto, K.; Yamaguchi, S.; Hoshino, S.; Kanda, A. Light-scattering sensor for monitoring properties of snow. Cold Reg. Sci.
Technol. 2020, 178, 103131. [CrossRef]
20. Marchetti, M.; Bourson, P.; Fontana, M.D.; Jobard, C.; Saintot, B.; Casteran, G. Spectroscopic and chemiometrics supported
studies on discrimination, phase transition and concentration identifications of 1,2-propylene glycol solutions, and of a mixture
of potassium acetate with 1,3-propanediol solutions as anti-icing fluids. Cold Reg. Sci. Technol. 2017, 142, 34–41. [CrossRef]
21. ICAO. The New Global Reporting Format for Runway Surface Conditions. Available online: https://www.icao.int/safety/
Pages/GRF.aspx (accessed on 5 January 2022).
22. Transport Canada. Advisory Circular (AC) No. 300-019: Global Reporting Format (GRF) for Runway Surface Conditions; Transport
Canada: Ottawa, ON, Canada, 2021.
23. Chu, L.; Guo, W.; Fwa, T. Theoretical and practical engineering significance of British pendulum test. Int. J. Pavement Eng. 2020,
23, 1–8. [CrossRef]
24. Eriskin, E.; Karahancer, S.; Terzi, S.; Saltan, M. Examination of the effect of superhydrophobic coated pavement under wet
conditions. Procedia Eng. 2017, 187, 532–537. [CrossRef]
25. Hurtado Mayen, A.; Farfán Cabrera, L.I.; Garza Montes-de-Oca, N.F.; Gallardo Hernández, E.A.; Moreno Ríos, M. Sand as a
friction coefficient improver on asphalt ice layers. Proc. Inst. Mech. Eng. Part J J. Eng. Tribol. 2022, 236, 13506501221074779.
[CrossRef]
26. Rainwater, K.A.; Lawson, W.D.; Surles, J.G.; Estrada, F.J.; Jackson, W.A. Side-by-side field comparison of snow and ice control
chemicals for anti-icing applications. Cold Reg. Sci. Technol. 2021, 184, 103230. [CrossRef]
27. Rodin Iii, H.; Nassiri, S.; AlShareedah, O.; Yekkalar, M.; Haselbach, L. Evaluation of skid resistance of pervious concrete slabs
under various winter conditions for driver and pedestrian users. Road Mater. Pavement Des. 2021, 22, 1350–1368. [CrossRef]
28. Sajid, H.U.; Naik, D.L.; Kiran, R. Improving the ice-melting capacity of traditional deicers. Constr. Build. Mater. 2021, 271, 121527.
[CrossRef]
29. Sabey, B.E.; Lupton, G. Friction on wet surfaces of tire-tread-type vulcanizates. Rubber Chem. Technol. 1964, 37, 878–893. [CrossRef]
30. Abdelaal, A.; Mirto, C.; Nims, D.; Ng, T.; Jones, K.; Ryerson, C.; Helmicki, A.; Hunt, V. Investigation of using icephobic coatings
on a cable stayed bridge. In Proceedings of the 16th International Workshop on Atmospheric Icing of Structures (IWAIS), Uppsala,
Sweden, 28 June–3 July 2015; pp. 133–138.
31. AMS1435I; Fluid, Generic, Deicing/Anti-Icing Runways and Taxiways. SAE G-12 RDF; SAE International: Warrendale, PA, USA,
2012; p. 16.