Announcing the 1st international workshop on icing of unmanned aircraft. The UAV icing workshop will be held in Trondheim (Norway) and online during November 29-30, 2022. The workshop will be hosted by the UAV Icing Lab at the Norwegian University of Science and Technology (NTNU). The objective of this workshop is to provide a platform for stakeholders in science and industry to discuss the challenges and technical solutions of icing for unmanned aircraft. We highly encourage all participants to contribute with a presentation relevant to the workshop objective. Find more information about the workshop online: www.uavicingworkshop.com.
Location/travel: The workshop is held in the Scandic Nidelven hotel in Trondheim. Trondheim has an international airport with direct flights to London, Amsterdam, Olso, Stockholm, Helsinki, Copenhagen, and more. Green transportation options are available by train via Oslo.
Cost: NTNU will cover fees for the conference for up to 50 participants on a first-come-first-served basis. This includes two lunches and a dinner. Travel and accommodation costs will need to be covered by each participant. We will offer special rates for the conference hotel (Scandic Nidelven).
Registration: You can register online here: https://register.uavicingworkshop.com/. Registration for presentations closes 1st October and for participation 21st October 2022.
Written by Bogdan Løw-Hansen, PhD candidate. Today, harsh weather conditions and especially icing are a big problem for uncrewed aerial vehicles (UAVs). Several solutions exist, but many require substantial amounts of energy to operate them, which are not always available on smaller UAV platforms. To this end, the UAV Icing Lab at NTNU is currently conducting research on the optimization of an electrically heated de-icing system. The researched solution is based on an ice shedding detection algorithm presented here.
The problem with icing
Research on in-flight icing for uncrewed aerial vehicles (UAVs) is a new topic that has only recently started to gain momentum. This is driven by several factors. For example, a survey on civil applications of UAVs reports that small to mid-sized UAVs with a wingspan up to a couple of meters have significant potential to succeed in many commercial applications [1]. Other factors are based on the fact that UAVs have already been shown to be effective in critical missions such as search and rescue, human organ transport, and surveillance [2,3]. Furthermore, UAVs have been able to provide crucial capabilities in modern warfare [4]. For instance, in the ongoing conflict in Ukraine, the Ukrainian forces have been so successful at deploying their UAVs that it has led to the creation of a patriotic song about them [5]. Together, the high utility of current UAVs and the projected growth of the global UAV market, estimated to reach over $25.13 billion by 2027 [6], have opened opportunities for further funding of UAV research. One of the research topics presently receiving attention and funding is the operation of UAVs in harsh weather conditions. Among the challenging weather conditions, the icing conditions, which cause a build-up of ice on the wings of the UAV during flights, are considered to be especially problematic. In-flight icing is a critical issue to solve because it has been recognized as a severe hazard for UAVs, leading to problems ranging from reduced flight performance to complete loss of the vehicle in extreme situations [7]. Furthermore, icing conditions are relatively common phenomena, especially in cold climate regions such as northern Europe and northern America. In fact, if icing conditions weren’t a problem, the time window of when it is safe to operate a UAV could have been more than doubled in certain locations [8].
The electric ice protection system
One of the developed solutions to make UAV operations in icing conditions safe is based on an electric heating system. The heating system uses power stored in batteries on board the UAV to heat the UAV wings when the vehicle is experiencing icing. This makes it difficult for the built-up ice to stick to the wings’ surface and leads to ice shedding. Figure 1 shows a schematic of such a heating system inside a UAV wing. The particular system displayed in Figure 1 is developed by UBIQ Aerospace in collaboration with the NTNU UAV Icing Lab. The IPS has four heating panels and one heating wire per wing. All of the heating elements stretch across the length of the wing. In addition, the IPS includes five temperature sensors used to measure the effectiveness of the applied heat.
Optimizing the ice protection system with an ice shedding detection algorithm
The electrically-heated IPS is used to initiate ice shedding when a critical amount of ice has been accumulated. A successful ice shedding is presented in Figure 2. The process shows a de-icing test performed in an icing wind tunnel, in which a UAV wing section is exposed to artificial icing conditions. The de-icing process can be described as follows. When enough ice has accumulated, the heating elements are used to melt some of the ice nearest to the wings’ surface. This loosens the ice, letting the incoming air shed the ice off the wings. By applying heat and shedding the ice, the IPS makes it possible to ensure that the amount of build-up ice never reaches the point where it becomes dangerous for the UAV. However, there are a number of problems associated with the use of IPSs in UAVs. One of them is that such electrically-heated de-icing systems use a lot of energy that otherwise could have been used to extend the flight time of the vehicle. In this article, an ice shedding detection algorithm is presented as a solution to the energy consumption issue. It works because an ice shedding detection system makes it possible for the heating to be turned off shortly after ice shedding has occurred. Thus, significantly reducing the amount of energy needed to operate the heating system. In contrast, without such detection systems, the heating cycles operate irrespective of the ice shedding status based on conservative estimates of how long the heat must stay on before the ice is removed.
Ice shedding detection
The ice shedding detection algorithm works by using two sources of information. The temperature sensors embedded in the wings and a model that relates the temperature measurements to the input from the heating panels. By comparing the expected temperature to the measured temperature, one obtains an error signal called the “innovation sequence.” The properties of the innovation sequence are such that it stays close to zero when the model and the measurements agree and grows large when they disagree. An example where the model and the measurements agree for the first 15 seconds is shown in Figure 3.
Now assume one has obtained a model of how the temperature in a UAV wing should behave when a layer of ice is surrounding it. With such a model, one can set up a detection threshold, based on the innovation sequence, to detect when the wing goes from being iced to being ice-free. By comparing data from several experiments, a threshold was found that could quickly and reliably identify ice shedding through the change of state in the wing from iced to ice-free. Figure 4 shows the four such detections in different de-icing experiments. The value αPU in the Figure 4 plots is the threshold at which the ice shedding detections are made.
The results show that ice shedding detection on a UAV is achievable. Furthermore, the average detection time of the presented method is only 2 seconds, allowing for efficient use of the IPS. The next step for the developed ice shedding detection system is to apply it in an actual flight, not only in an icing wind tunnel. Furthermore, it would be interesting to test the system together with an ice detection system which is supposed to initiate the de-icing process. To sum up, this research resulted in a method that can significantly reduce the energy requirements for UAV ice protection systems. By using this method in the future, UAVs will be able to operate safely and conduct longer missions inside icing conditions.
References
[1] Shakhatreh, H., Sawalmeh, A. H., Al-Fuqaha, A., Dou, Z., Almaita, E., Khalil, I., Othman, N. S., Khreishah, A., and Guizani, M., “Unmanned aerial vehicles (UAVs): A survey on civil applications and key research challenges,” IEEE Access, Vol. 7, 2019, pp. 48572–48634.
[2] Scalea, J. R., Restaino, S., Scassero, M., Blankenship, G., Bartlett, S. T., and Wereley, N., “An initial investigation of unmanned aircraft systems (UAS) and real-time organ status measurement for transporting human organs,” IEEE Journal of translational engineering in health and medicine, Vol. 6, 2018, pp. 1–7.
[3] Girard, A. R., Howell, A. S., and Hedrick, J. K., “Border patrol and surveillance missions using multiple unmanned air vehicles,” 2004 43rd IEEE conference on decision and control (CDC)(IEEE Cat. No. 04CH37601), Vol. 1, IEEE, 2004, pp. 620–625.
[4] M. Burgess, “Small drones are giving Ukraine an unprecedented edge,” Wired, 06-May-2022. [Online]. [Accessed: 10-May-2022].
[5] CNN, “Turkish drone is so effective, Ukrainian troops are singing about it,” YouTube, 07-Apr-2022. [Online]. Available: https://www.youtube.com/watch?v=S4qUsPCFV28. [Accessed: 18-May-2022].
[7] Hann, R., and Johansen, T. A., “UAV icing: the influence of airspeed and chord length on performance degradation,” Aircraft Engineering and Aerospace Technology, Vol. 93, 2021, pp. 832–841.
[8] Sørensen, K. L., Borup, K. T., Hann, R., Bernstein, B. C., and Hansbø, M., “Atmospheric icing limitations, climate report for Norway and surrounding regions,” Tech. rep., UBIQ Aerospace, 2021.
In one of our recent icing wind tunnel experiments, the NTNU UAV Icing Lab investigated if ice fragments are a hazard to UAV propellers. The resulting high-speed video footage shows the moment ice impacts the propeller blades. Here is a small preview of the footage.
When an aircraft flies into special meteorological weather conditions, ice can form on the airframe. When fragments of ice break off the surface, they can pose a hazard to the propeller. Such ice fragments can be created after long icing periods or in combination with de-icing systems. When an ice fragment hits the blades, there is a risk that the propeller might be damaged or destroyed – which may lead to the entire aircraft crashing. Especially for UAV propellers, which are often small and lightweight, this is a very relevant risk that can severely affect the safety of the drone.
The NTNU UAV Icing Lab is currently working together with UBIQ Aerospace and Mejzlik Propellers to study this topic. We are currently preparing a scientific paper about the issue.
The UAV Icing Lab has been featured in the online magazine NorwegianSciTechNews.com and its Norwegian counterpart, Gemini.no. The article is using a lot of multi-media elements and is aimed at the general public!
The Norwegian University of Science and Technology (NTNU) has signed a collaboration agreement with the German Aerospace Center (DLR). The scope of the agreement is to conduct joint research in the field of icing on unmanned aerial systems (UAVs). The collaboration is going to be pivoting around the NTNU UAV Icing Lab the the DLR Institute of Flight Systems
The DLR is the national aeronautics and space research centre of Germany. It has extensive research and development experience in aeronautic and other research fields. The DLR Institute of Flight Systems is part of several national and international projects and research cooperations to develop new high-quality simulation models as well as new ice detection methodologies. Profiting from its long tradition in aircraft model identification the Institute of Flight Systems has achieved outstanding results in the modelling of icing effects on manned aircraft during the last years.
Together with the NTNU UAV Icing Lab, the two partners will venture further into the field of icing on drones, with potential applications also for urban air mobility (UAM). The focus of the work will be to conduct model identification on iced unmanned aircraft provided by the UAV Lab. The DLR will contribute their experience in system identification and NTNU their operational experience and drones.
**NEW Publication** A recent paper investigates the sensitivity of unmanned aircraft to icing in comparison to manned aircraft. Differences in airframe size and airspeed can lead to substantially different icing performance penalties.
Think of a “normal” aircraft. Chances are that you are thinking of something that is officially called a “large transport aircraft” – a jet or turboprop aircraft that is mainly used for passenger transportation. Now think of an unmanned aircraft. Maybe you think of a small quadcopter, the type of which are sold commercially. Or maybe of a large military fixed-wing aircraft that is often seen in the news. Or maybe you think of something completely different.
Typical airliners. Photo: Richard Hann
In fact, while most manned aircraft are kind of similar to each other (with exceptions of course), UAVs come in a very large variety of shapes and sizes. The smallest unmanned aircraft can have a size in the order of centimeters, while the largest have wingspans that are comparable to “normal” manned aircraft.
When it comes to the risk of in-flight icing, airframe size and airspeed are two important parameters. Research has shown that these two parameters have a large influence on how much ice is accumulating on airfoils and wings. In this new paper, we are taking this research a step further. We investigated the effect of airframe size and airspeed on the icing performance penalties that result from the ice accretions. To do this, we ran computational fluid dynamic (CFD) simulations with a software called ANSYS FENSAP-ICE.
The simulation showed that smaller airframes lead to an increase of the area-specific ice accumulation. In other words, smaller airframes collect more ice relative to their size. FENSAP-ICE also indicated that larger ice shapes tend to lead to larger performance penalties in lift, drag, and stall behaviour. In summary, this means that in identical icing conditions, a smaller airframe will accumulate more ice and generate larger icing penalties compared to a larger airframe. Small airframes are therefore more sensitive to icing.
The figure shows how size affects the ice accretion on a glaze airfoil. The smaller the airframe size (here: chord lenght), the larger the ice shape relative to the airfoil. Larger ice shapes usually lead to higher aerodynamic penalties.
The simulation results also show how the speed of the UAV has a large effect on aerodynamic heating. The faster an aircraft flies, the more friction it has with the air. This friction generates heat, which affects the ice accretion process. For fast aircraft, this heat may prevent ice accretion at temperatures close to the freezing point. Since UAVs are typically much slower than manned aircraft, they can encounter severe icing conditions at near freezing temperatures.
The figure shows how airspeed affects the ice accretion on a glaze airfoil. The faster the aircraft is, the more friction is generates and the more “wet” the ice shapes become.
In summary, this paper show that unmanned aircraft can be considered more sensitive to icing. Icing conditions that a large and fast manned aircraft experiences as “light” without any significant aerodynamic penalties may be very hazardous to small and slow unmanned aircraft.
This summer, the UAV Icing Lab went on a joint campaign with UBIQ Aerospace at the icing wind tunnel facility of the Technical Research Centre of Finland (VTT). The main goal of the campaigns was to do experiments on ice protection systems for UAVs and to study the effects of icing on propellers. Among other things we used a high-speed camera system to capture ice shedding events.
Icing of unmanned aerial vehicles (UAVs) occurs during flight through clouds that contain supercooled liquid droplets. These cloud droplets are in a liquid phase but have a temperature below the freezing point. When these droplets collide with a surface, like the wing or the propeller of a UAV, they freeze and can accumulate into large ice accretions. This ice disturbs the airflow and can lead to severe performance penalties and lead to the crash of the UAV.
One of the most reliable tools to research icing on UAVs are icing wind tunnel tests. An icing wind tunnel is a special facility that can generate the same type of icing conditions that a UAV would encounter in real flight – but under controlled laboratory conditions. An icing wind tunnel typically has three main components: the wind tunnel, the cooling system, and the spray system.
A wind tunnel is the core of every icing wind tunnel. It is the part that generates the high wind speeds that match the flight velocity of UAVs. The spray system injects small droplets into the air stream of the wind tunnel. These droplets are equivalent in size and number to the conditions of icing clouds. The cooling system is required to lower the temperature of the air and droplets to sub-zero conditions. There are multiple ways of how this can be achieved. In our case, the entire wind tunnel was located inside of a large climate chamber that can be cooled down to -25°C. The combination of these three systems, wind tunnel, spray system, and climate chamber generates the same type of supercooled liquid droplets that a UAV encounters in flight.
During the test campaign this summer we used a high-speed camera to record ice shedding. Ice shedding refers to the process when the ice that has accumulated on a surface is removed. On a wing, this happens when an ice protection system is used that heats the wing and therefore breaks the bond between the ice and surface. On a propeller, ice shedding occurs when the ice accretion becomes so big, that shedding occurs due to centrifugal forces. We used a high-speed camera system to capture both of these events.
Ice shedding from a de-icing system on a UAV wing. Video: Nicolas Müller.
The high frame rate videos allow us to see exactly how the ice shedding process occurs. We can use them to identify different shedding processes that can occur. For example, on the ice protection systems we have cases where most of the ice melts and other cases where the ice breaks away due to the aerodynamic forces. Understanding these processes helps to develop efficient ice protection systems.
On the propellers, ice shedding is a dangerous mechanism, especially when ice sheds asymmetrically on the blades. When ice sheds form on one blade but not the other, it can lead to very strong vibrations that can damage the propeller and motor.
Ice shedding from a de-icing system on a UAV propeller. Video: Nicolas Müller.
The data and knowledge gathered during these icing wind tunnel tests will be used in the UAV Icing Lab to advance our understanding of the effects of icing on unmanned vehicles. The data will also assist in developing better models and methods to mitigate the dangers of icing. This is an important contribution to help enable unmanned aircraft technologies for future applications like urban air mobility or advanced air mobility.
Ice build-up at the leading-edges of unmanned aerial vehicle (UAV) wings has been identified to severely limit the aerodynamic performance of mid-sized fixed-wing UAVs like the Maritime Robotics “Falk”. This UAV type with wing spans up to several meters usually flies long-range and long-endurance reconnaissance missions. When the flight path of the UAV is passing through clouds or precipitations, the UAV can encounter icing conditions and is prone to ice shapes accretion on its fuselage. This is why the aerodynamic performance losses caused by three different ice shapes has been investigated experimentally and numerically.
Icing on a fixed-wing UAV. Artistic illustration: UBIQ Aerospace.
The experimental campaign was conducted in Belgium at the von Karman Institute for Fluid Dynamics. The institute provided its largest wind tunnel facility to perform the aerodynamic performance experiments on a RG-15 airfoil. To determine the deteriorations from ice build-up at the leading-edge of an UAV wing, a clean airfoil reference case was compared to an iced airfoil case. Therefore, the clean airfoil tests served as the baseline comparison case for the iced airfoil tests. The RG-15 airfoil was artificially iced by taping 3D-printed ice shapes to the wing’s leading edge. Throughout the experimental campaign, the baseline clean airfoil was tested at multiple angles of attack and wind speeds. Due to some experimental challenges, only the glaze ice shape could be tested at multiple angles of attack and two different wind speeds. The other two considered ice shape geometries (rime and mixed) were only tested at one angle of attack at multiple wind speeds. The comparison of the experimental data revealed that the glaze ice shape introduces severe aerodynamic penalties compared to the clean RG-15 airfoil.
RG-15 airfoil mounted inside the wind tunnel with artifical ice shapes attached. Photo: Johannes Oswald.
The minimum aerodynamic drag of the glazed RG-15 airfoil was increased with up to 130% compared to the clean airfoil. The lift coefficient was slightly decreased by 0.03-0.07. Furthermore, the ice shape introduced a destabilizing “nose-down” moment to the baseline RG-15 airfoil. Since experimental campaigns are in general expensive and time consuming, it is also of great interest to virtually simulate these aerodynamic performance penalties. These numerical computational fluid dynamic (CFD) simulations also require validation experiments.
Different ice shapes on the RG-15 airfoil. Figure: Johannes Oswald.
Consequently, the data of the conducted experiments was used to compare to CFD simulations with FENSAP-ICE. The CFD simulations were generally in-line with the measured aerodynamic drag propagation. Limitations occurred in the prediction of the maximum aerodynamic lift of the iced RG-15 airfoil. This led to the conclusion, that the chosen CFD simulation approach of the iced RG-15 airfoil may be adequate to estimate general aerodynamic penalties but may be limited in the flight-stability critical onset of stall region.
We often hear the following question: “How often do drones actually encounter icing conditions?”. This is a very good question – and until recently we had only vague answers. Usually, we would refer to two reports in the literature that discuss icing frequencies for the US and the rest of the world. These two reports have been written for manned aviation – which means that they focus on icing at scales and altitudes at which normal passenger transport airplanes travel. Drones typically travel at much lower altitudes and operate in much smaller areas, which has somewhat limited the significance of these reports.
A new answer to this question has been provided by UBIQ Aerospace. UBIQ is a company building ice protection solutions that enhance the capabilities of unmanned aircraft and one of the key collaborators of the UAV Icing Lab. Today, UBIQ released a climatological report about the risk of icing in Norway and surrounding regions.
The report investigates the frequency at which atmospheric icing is expected to occur over Norway and neighbouring countries. The study is based on a climatologcal study conducted by one of the leading icing meteorologists, Ben Bernstein, for UBIQ Aerospace. The datasets covers 10-year period and evaluates the risk for icing at various altitudes.
The report highlights the risk of atmospheric icing for aircraft and proves that icing is a year-round risk. It also shows seasonal effects with the highest icing risks occuring in Norway during the winter months. Furthermore, it also shows topological effects where the risk of icing is higher in mountainous regions compared to costal areas.
Potential icing condition frequencies over the year – based on a climate analysis spanning data from 2010-2019. Source: UBIQ Aerospace.
Unmanned aircraft are particularly at risk for icing conditions [read more]. Icing conditions which are experienced as “light” are having much more severe effects on UAVs which are typically smaller and travel at lower velocities. In addition, unmanned aircraft often travel at lower altitudes compared to manned aircraft. As the report shows, very high icing frequencies can occur near the ground during the winter months – making this a particularly severe hazard for UAVs.
Potential icing condition frequencies near Oslo airport at different altitudes for each month – based on a climate analysis spanning data from 2010-2019. Source: UBIQ Aerospace.
In summary, the report offers important insights into the overall risk that unmanned aircraft are facing in Norway and the surrounding regions. Our mission at the UAV Icing Lab is to explore these hazards scientifically and to provide the knowledge and solutions to overcome the “icing barrier”. In this, are working in close cooperation with UBIQ Aerospace who are an industrial supplier of holistic ice protection systems for unmanned aircraft.
Reference: Sørensen, K.L., Borup, K.T., Hann, R., Bernstein, B., Hansbø, M.: “UAV Atmospheric Icing Limitations: Climate Report for Norway and Surrounding Regions”. UBIQ Aerospace. 2021. Available at: http://www.ubiqaerospace.com/climate-report
**NEW PUBLICATION** In-flight icing on UAVs is a severe hazard that prevents drone operations in bad weather conditions. In order to overcome this limitation, UAVs can be equipped with ice protection systems that mitigate the negative effects of icing. One solution is to heat the exposed surfaces of the aircraft electrically with carbon fiber materials. The big question is how much (or rather how little) energy is required to prevent ice from building on the aircraft.
In this paper, we investigated several designs of such electrothermal ice protection systems in collaboration with UBIQ Aerospace. Our goal was to find out how different designs and operation modes affect the required energy for ice protection. We tested two systems in an icing wind tunnel at the Technical Research Centre of Finland (VTT) in Helsinki. An icing tunnel is a facility where icing conditions can be simulated under laboratory conditions.
The ice protection system in action.
The experiments showed that de-icing with parting strip gives the most energy-efficient ice protection. De-icing means that the ice protection system is not operating continuously (which is called anti-icing). Instead, the system allows for uncritical amounts of ice to build-up on the airframe, which is then periodically removed by heating the surfaces. In practice, we allow ice to build up over a few minutes (typically 2-4min) and then heat for about 10-30s. In addition, we use a parting strip. A parting strip is essentially a continuously heated wire that creates a “gap” in the ice that builds on the leading edge of the wings. Instead of having a continuous ice accumulation, the parting strip separates them into an upper and lower part. The advantage of this gap is that it gives the aerodynamic forces a larger area and longer leverage for acting on the ice. This means that the incoming airflow now “helps” to shed ice by exerting an additional shedding force on the ice accumulations. In practice, parting strip ice protection systems de-icing a wing substantially faster compared to de-icing without parting strip.
The left image shows the cross-section of an ice accretion on the airfoil leading edge after 6 minutes of icing. The right images shows how the parting strip separets the ice accretion with a “gap”.
Schematic of how the parting strip helps to remove ice from the airfoil by increasing the area and leverage of aerodynamic forces acting on the ice.
In summary, these experiments will help us improve the design of ice protection systems. The results will be used to increase the energy-efficiency of ice protection systems. As a consequence, UAVs can safely fly for longer durations in icing conditions. This is an important step for unlocking many new applications of unmanned aircraft such as package deliveries, search & rescue operations, or even passenger transports in urban air mobility.
This is the main results from the paper. They show that de-icing with a parting strip requires less energy than de-icing without a parting strip.
UAV Icing Lab 