One of the challenges to the safe and reliable operation of aircraft is the hazard of in-flight icing. Ice accretions on the aircraft happen when supercooled droplets – droplets that are liquid although their temperature is below freezing – hit the surface. Depending on the position where the ice grows, different problems can arise. Ice accretions on wings reduce the lift and increase the drag of the aircraft. Ice on propellers can reduce the thrust and increase the required power significantly within one or two minutes. When sensors or antennas ice up, their output might be faulty or stop completely. Overall, icing on aircraft has multiple negative effects that can cause a crash in the worst case.

Hence, icing research has been an important field within the aircraft community for more than seven decades. Two different main areas of research can be distinguished. First, understanding the physics behind icing and how ice accretions affect different parts of the aircraft. Second, developing and testing solutions to protect aircraft that fly in icing conditions. Different methods are used to conduct the research. The foundation has been laid by performing theoretical research about the path of droplets in the air, where they impinge, and what factors influence the rate of droplet freezing. This research has been supported by performing experiments. A second, more recent method is the use of numerical simulations.

While numerical simulations allow investigating a wide range of different conditions, they cannot serve as a standalone tool. This is because the equations that are used to do the numerical calculations are often not exact replications of the true processes but require modeling of some parameters. The model tuning is typically done by comparing numerical simulations to experimental results. Hence, experimental tests play a very significant role in the field of icing research. Experiments can have different levels of complexity and similarity to the conditions found in real flights. Especially some of the foundational work has been done in conditions that are different from real applications, for example by using flat plates as the geometry. These experiments still allow gaining insight into the theory behind icing. However, when we aim to understand in more detail how different parts of aircraft are affected or can be protected, the experimental conditions must be closer to the conditions in real flights (real geometries, sizes, ambient conditions, etc.).

The highest similarity to real flights can be achieved by performing flight tests. The goal is to operate the aircraft in real icing conditions and investigate the effect of icing or the performance of the protection solutions. Two big challenges related to flight tests in icing conditions exist. First, it is extremely difficult to find icing conditions and it can take a very long time to conduct enough flight tests to cover the wide range of potential cloud conditions. Second, measuring all important variables, related to both the ambient conditions and the influence of ice accretions, during the flight is challenging or even impossible. Thus, experiments in icing wind tunnels are a popular alternative since they allow testing at different conditions in a controlled manner.

Like conventional wind tunnels, icing wind tunnels have fans to create an airflow through the tunnel and towards a test object, for example, a wing or a propeller. The special feature of icing wind tunnels is that they can recreate conditions as aircraft could find them in icing clouds during a flight. To do so, icing wind tunnels are equipped with a spray system to spray droplets into the airflow. Additionally, the air in the tunnels can be cooled down below freezing to supercool the droplets before they hit the test object. This can be achieved by having a heat exchanger that only cools down the air that flows through the closed return tunnel, or by having an open wind tunnel and cooling down the whole room, see figures above. Hence, testing in icing wind tunnels allows performing experiments that are close to real icing conditions in a controlled environment. This is an important tool since it allows testing at specified conditions without having to wait or search for them for a long time. Additionally, the physics does not have to be modeled, as is the case for numerical simulations.

But unfortunately, there are also shortcomings of experiments in icing wind tunnels. First, the number of facilities in the world is small. This also means that time in icing wind tunnels is limited and expensive. Second, all icing wind tunnels come with limitations in the range of conditions they can test. Except for the largest facilities in the world, the size of the test section is typically only in the range of one to three meters. Hence, testing full aircraft or even full wings for manned aircraft is typically not possible. This is especially challenging because scaling is a very difficult aspect of icing. To meet all similarity requirements of icing conditions, many parameters must be matched, often resulting in contrasting constraints. Hence, scaling of results is not commonly done. Additionally, also the range of ambient parameters that can be tested in icing wind tunnels depends on the facilities. Most facilities can only spray limited amounts of water in the air, only have limited speed variation, or are limited regarding the coldest temperatures they can keep during experiments. Last, but not least, the conditions in icing wind tunnels are no exact replicates of icing conditions in real flights. For example, the experiments in icing wind tunnels are typically performed at constant conditions during the whole run. However, in real clouds, the conditions can change significantly within a few hundred meters.

Summarized, icing wind tunnels are neither a universal tool for icing research nor a niche tool. Yes, icing wind tunnels have some shortcomings that reduce their capabilities. However, they are still a very important tool for performing icing research because they allow testing without the limitations of numerical models and without the very expensive and time-consuming task of finding icing conditions for real flights. Also, the other methods have their strengths and weaknesses. Hence, it is probably the combination of the different methods that leads to the best results. Using numerical simulations to cover wide ranges of icing conditions, verifying the results using icing wind tunnels, and finally confirming these findings in real flights should be the best use of all the different techniques and maximize the learnings.

Text: Joachim Wallisch

**NEW PUBLICATION**  In-flight icing is a severe risk for unmanned aerial vehicles (UAVs). In these conditions, ice accumulates on the wings and propellers which disturbs the airfoil. As a consequence, the aerodynamic performance of the wings and propeller is reduced. Ice also reduces the effectiveness of the control surfaces. Most UAVs use an autopilot system for flying which typically cannot cope well with these icing performance losses. In the worst case, the autopilot may even steer the UAV into a situation where it cannot maintain stable flight – and crash.

In our recent publication, we explore the use of a more robust autopilot, that can deal with the icing effects in a safe manner. We compare a model reference adaptive control (MRAC) scheme to PID controllers to maintain stable flight in icing conditions. The findings show that MRAC control scheme and the PID controller demonstrate similar qualities in tracking performance, with the MRAC performing better under certain conditions.

Reference: Högnadottir, S., Gryte, G., Hann, R., Johansen, T.A. (2023). Inner-Loop Control of Fixed-Wing Unmanned Aerial Vehicles in Icing Conditions. AIAA Sci-Tech Conference. DOI: 10.2514/6.2023-1049

Text: Richard Hann

By Bogdan Løw-Hansen, Joachim Wallisch, Markus Lindner, Michael Cheung and Nicolas Müller

The 1^st International Workshop on Unmanned Aircraft Icing took place in Trondheim on 29^th-30^th of November. The purpose of the workshop was to gather different stakeholders interested in unmanned aerial vehicle (UAV) icing, including – scientists, engineers, manufacturers, investors, operators, and authorities – and learn what other stakeholders are working on and where potential collaborations can be found.

Keynote presentations

The opening talk, given by Richard Hann, highlighted the significant growth of UAVs in the current commercial and defense markets, as well as the limitations that hinder even faster growth, one of them being the atmospheric icing.

In the first keynote speech, Kim Sørensen – CEO of UBIQ Aerospace – presented the commercial opportunities related to the UAV icing challenge. Since the current solution is to ground the UAVs if there is a potential for icing, and icing conditions are likely to occur in the range of 19 to 78 days a year in the US, there is a large potential market for ice protection systems (IPS), especially light-weight and energy efficient IPS that do not reduce the operational capabilities of the UAVs.

On the second day, Professor Peter Webley from the University of Fairbanks of Alaska highlighted the challenges of operating a UAV in the Arctic. An important takeaway from the presentation is that it is not sufficient to only protect the aircraft, as the other parts of the system, like the payload, the communication links, and the pilots on the ground are highly affected by cold weather and icing as well. Therefore, a holistic view on the operations is required to ensure safe operation of UAVs in cold climates.

Research topics

Even though UAV icing might seem like a very narrow field, based on the topics presented at the workshop, it is clear that it is a challenge that has engaged a multidisciplinary effort towards a solution. The UAV icing topic has attracted several specialists from the UAV industry, research institutions, and academia that work on solutions to enable UAV operations in icing conditions. The research presented during the workshop can be categorized into the following groups:

• Meteorology analysis and prediction of icing conditions
• Efficiency optimization of heat-based ice protection systems
• Efficient and lightweight ice detection sensors and methods for small UAVs
• Development of testing facilities for atmospheric icing conditions
• Research on icephobic materials and coatings
• Regulations and certification of systems for unmanned flight in icing conditions

Highlights

Many interesting facts were presented during the two-day workshop; some of the most interesting highlights are mentioned below.

Today, UAV operations are severely limited by icing. As Kim Sørensen quoted in his keynote, the current solution to UAV icing is to not fly in icing conditions:

Standard procedures for UAS operations: avoid icing at all costs. Ground the aircraft in the event of potential icing. – Boeing Insitu

Another important finding was that the main challenge of developing a viable IPS for small UAVs is to make it efficient. The current solutions require large amounts of power and are therefore highly limited by the low battery capacity of small UAVs.

Finally on the point of frequency and severity of icing conditions, VTT (Technical Research Centre of Finland), has performed a high-resolution meteorological analysis of atmospheric icing conditions in the altitude 0-1000 m, relevant for small UAVs. The results indicate that severe icing conditions will on average occur 12.8% or 47 full days per year in the region of northern Europe.

Concluding remarks

Based on the feedback, the 1^st International Workshop on Unmanned Aircraft Icing was a success. The shared interest and different backgrounds of the presenters created good opportunities to find synergies and build new connections. Several social events were organized to accompany the workshop, including an informal icebreaker event held the evening before the workshop, lunch and dinner served at Scandic Nidelven Hotel, as well as a visit to a bar, which all provided an excellent opportunity to socialize.

The 2^nd International Workshop on Unmanned Aircraft Icing is planned to take place in 2024. Subscribe to our newsletter to stay updated about the research done at NTNU UAV Icing Lab and the upcoming events. More information will follow on: www.uavicingworkshop.com.

One week after the 1st UAV Icing Workshop in Trondheim, we thank all participants for joining! With your support, we have made this event into a successful forum for discussing UAV icing-related challenges in research, industry, and regulations. More than 80 participants were registered for the event with about half of them physically present in Trondheim. We had representatives from 16 countries, spanning all around the globe. Many European countries, but also Canada, USA, South Korea, China, and New Zealand – resulting in the working being a truly international forum! Furthermore, about 60% of the participants had a background in research, 15% in industry, 10% in government, and 5% in defense.

We are happy to announce that the UAV Icing Workshop will return again in 2024! More information will follow on: www.uavicingworkshop.com.

**NEW PUBLICATION**  When flying in icing conditions, uncrewed aerial vehicles (UAVs) face severe dangers. The propeller of the UAV, which is creating the thrust for the UAV is especially sensitive to the accumulation of ice during flight in atmospheric icing conditions. To estimate the impact of icing on the propeller, we conducted experimental tests in an icing wind tunnel. The icing wind tunnel is a special facility whose main feature is a spray system, that can emulate the conditions in a cloud. The propeller was mounted on a test-bench that recorded the forces on the propeller. Also, a high-speed camera captured images of the ice on the propeller to get further details on the ice accumulation process. The propeller was tested across a wide range of icing conditions to separate the different influence factors on the icing on the propeller.

The ice shapes start to grow on the leading edge of the propeller. This ice shape disturbs the airflow around the propeller. The disturbed airflow makes the propeller less efficient. If the propeller is no longer producing enough thrust, the UAV will no longer be able to continue to fly.

If the mass of the accumulated ice has grown to a critical points, parts of the ice will break off and thrown away from the propeller. The removal of ice improves the aerodynamic performance of the propeller in the briefly, until the ice has been formed again. But the blocks of ice breaking are also a risk to the UAV. The ice fragments can hit other parts of the UAV, like the empennage. More importantly, the ice shedding leads to an imbalanced mass between the propeller blades. This leads to very strong vibrations. In our testing campaign those vibrations exceeded the 10G measurement rage of the used sensor. This is the reason why focusing on the ice shedding process is very important to understand the risk of ice on the propeller of a UAV.

The analysis of the ice shedding has shown that the amount of ice that can form on the propeller is strongly dependent on the temperature. Lower temperatures lead to larger amounts of ice on the propeller, while at higher temperatures the amount of ice on the propeller is lower. The size of ice fragments that shed from the propeller were the largest at temperatures of -10 °C.

Another focus of the study was the aerodynamic performance of the propeller with ice accumulated on the propeller. Ice on the propeller reduces the thrust of the propeller and its efficiency. This could be measured by our test setup to see how different temperatures influence the propeller. Here it is clear that the loss in performance is the worst for temperatures of about -10°C. This hits a sweet spot between higher temperatures which have less amount of ice freezing on the propeller and lower temperatures where the ice shapes on the propeller are very streamlined and do not influence the performance of the propeller as much.

The whole range of measurements has been taken and merged into an algorithm to predict the performance of the propeller across a range of conditions. This can be used as a tool in flight simulators to predict the influence icing has on the flight of a UAV. This enabled the development of advanced path planning methods to optimize flight routes to reduce the impact of icing and to train autopilots to cope with icing on propellers.

Reference: Müller, N., Hann, R. (2022). UAV Icing: A Performance Model for a UAV Propeller in Icing Conditions. AIAA Atmospheric and Space Environments Conference. DOI: 10.2514/6.2022-3903

**NEW PUBLICATION** Aircraft without ice protection systems will face severe issues when flying in icing conditions. To reduce the risk of losing the aircraft, the aircraft must either remain grounded when the risk of icing exists, or it must be equipped with an ice protection system. As of today, no mature ice protection system exists for small unmanned aircraft. To fill this gap, a new ice protection system was developed and thoroughly tested in an icing wind tunnel.

The occurrence of in-flight icing can significantly influence the performance of aircraft. Hence, every aircraft that is supposed to fly in or through icing conditions must have an ice protection system (a system that protects the aircraft from the adverse consequences of icing). This can be particularly important for unmanned aircraft. Drones, uncrewed aerial vehicles (UAVs), unmanned aerial systems (UAS), and urban air mobility (UAM) are often used as synonyms to describe unmanned aircraft – an area that is getting more and more popular.

Unmanned aircraft are typically more prone to icing than manned aircraft. Hence, ice protection systems are extremely important for unmanned aircraft to allow operation in all weather conditions. However, as of today, unmanned aircraft are not commercially available with an ice protection system. This means that unmanned aircraft will typically remain grounded when potential icing conditions are present (liquid droplets are in the air at temperatures below the freezing point).

An important parameter for the quality of an ice protection system for unmanned aircraft is the energy-efficiency. The more energy the ice protection system requires, the less the aircraft’s possible range. Since the available energy on board the aircraft is significantly lower for unmanned aircraft than for manned aircraft, this is a bigger issue for unmanned aircraft than for manned aircraft.

An example of a possible ice protection system for unmanned aircraft is D•ICE. D•ICE was originally developed at the Norwegian University of Science and Technology (NTNU) and is now commercially developed by UBIQ Aerospace. It is an electrothermal system that is designed to work best in de-icing mode. This means that after some time of ice accretion, the wing is heated by powering carbon fiber layers inside the wing. The heat melts some ice at the interface between the wing and the ice. When enough ice is melted, the ice will shed from the wing, as can be seen in the picture below.

To examine the performance of different settings of the ice protection system, tests were performed in the icing wind tunnel at the Technical Research Centre of Finland (VTT). An icing wind tunnel is a special wind tunnel that is equipped with a spray system and can be operated at constant temperatures below freezing. The wing was placed in the airstream that contains water droplets. After allowing ice to accumulate on the wing for four minutes, the heating zones are activated, and the de-icing procedure starts.

One of the key findings was that the heat flux provided to the heating zones is one of the operational settings of an electrothermal ice protection system that can be adjusted. Increasing the heat flux results in higher temperatures at the wing’s surface; hence, the ice starts melting faster. As a result, ice shedding happens faster when higher heat fluxes are used. While one might think that faster shedding also results in a more energy-efficient de-icing process, this is not always true. On the contrary, the results showed that – averaged over time – more energy was required for the de-icing when higher heat fluxes are used.

The heat flux provided to the heating zones is one of the operational settings of an electrothermal ice protection system that can be adjusted. Increasing the heat flux results in higher temperatures at the wing’s surface; hence, the ice starts melting faster. As a result, ice shedding happens faster when higher heat fluxes are used. While one might think that faster shedding also results in a more energy-efficient de-icing process, this is not always true. On the contrary, the results showed that – averaged over time – more energy was required for the de-icing when higher heat fluxes are used.

This means that for an encounter with a given duration, using lower heat fluxes for de-icing results in less energy used although the individual ice shedding times are faster.

While it should be said that the results might be different for different geometries and internal structures of ice protection systems, the study showed some potential ways to reduce the energy needs of ice protection systems. Using the results of the study to improve ice protection system operation modes will hopefully enable the flight of unmanned aircraft in icing conditions soon.

Reference: Wallisch, J., Hann, R. (2022). UAV Icing: Experimental Investigation of Ice Shedding Times with an Electrothermal De-Icing System. AIAA Atmospheric and Space Environments Conference. DOI: 10.2514/6.2022-3905

**NEW Publication** Icing severity indices are ameasure for the in-flight icing risk and are an important factor for path and mission planning. Small aircraft, such as unmanned aerial vehicles (UAV) and urban air mobility (UAM) vehicles, are particularly sensitive to icing and thus proper assessment of weather risks is critical.

Icing severity is a metric for how quickly in-flight icing degrades the performance of the aircraft in a certain area. Performance means the ability of the aircraft to generate lift and sufficient thrust to stay airborne. In icing conditions, lift generation of the wings decreases while drag increases and thrust of the propeller decreases. This ice accumulation is very dangerous and is likely to lead the UAV to crash. For these UAVs, knowing the icing severity in an area becomes crucial, as they are much more sensitive to icing compared to their larger manned counterparts. Hence, having an accurate index is vital for planning purposes as to minimize hazardous conditions and to enable increased mission success rates.

Our recently published article proposes a new method to calculate an icing index that is better suited for unmanned aircraft. Icing indices today – as defined by the Aviation Weather Center (AWC) – categorize icing severity levels according to how fast 6 mm of ice accumulates on an airfoil. The issue is that this fixed value of ice thickness will have different degrees of impact, depending on the size of the aircraft. The amount of ice build-up that is deemed severe is relative to the size of the aircraft. One can think of it in terms of ice relative to the airfoil size, and the larger the fraction the worse it is. A small aircraft will experience severe icing effects with 6 mm, while a larger aircraft it might experience just light icing effects. The AWC icing index also does not account for the type of ice encountered which has been shown to degrade performance differently. Hence, the AWC definition is lacking in scalability and does not differentiate between different meteorological icing conditions. Our new proposed method compensates these shortcomings.

There are two main types of ice that are of interest, rime ice which preserves the airfoil shape, and glaze ice which leads to more uneven surfaces leading to worse performance in general. One more aspect of ice types is that it affects different aircraft surfaces differently. Rime ice is less severe on the wings while glaze is worse. On the propeller, it is the opposite. The main reason is that the propeller can shed ice due to rotational forces. Glaze ice has little effect on the propellers as it can shed this ice type completely. Rime ice has a higher adhesion force which allows it to hold on more. It can also lead to uneven shedding as rime ice has less cohesion with itself which leads to increased degradation.

Our proposed method assumes exposing an aircraft to different ice for 20 minutes, then uses lift and drag curves to calculate the performance degradation directly. The increase in power consumption is compared to the baseline no ice condition to calculate the index. By calculating the index this way, scalability is not an issue as the performance curves are directly linked to the aircraft and key variables, such as wingspan, are included in the calculation. The different icing conditions that are tested also consider specific icing conditions. It does mean, however, that the proposed method is aircraft specific and will require more testing in general to get the iced performance curves.

Some of the results are shown below with icing severity index values at different median volume diameters (MVDs) of droplets and it shows that rime ice is overall worst. This is because the propeller is what is keeping the aircraft in the air as it is the propeller that generates thrust and hence dominates the index. A decrease in propeller efficiency means it compensates by an increased consumption in power to deliver the same thrust. With longer icing durations, this might change because the wing does not shed ice naturally as the propeller does. Hence, with longer durations, the glaze will ultimately overtake rime as the worst ice type.

To summarize, icing indices today have been found to be inadequate for planning purposes for typical UAVs and UAMs. UAVs and UAMs tend to be much smaller than the commercial airliners and icing indices today do not take size into account. Our proposed method uses icing performance degradation metrics directly to calculate an icing index. This new method accounts for the problems with scaling and also accounts for different icing conditions. It is, however, more complex to calculate as it is aircraft specific.

To improve upon what is investigated in the paper, including 3D effects, and expanding the investigated icing conditions would be beneficial for a more accurate index.

Reference:  Cheung, M., Hann, R., Johansen, T.A. (2022). UAV Icing: A Unified Icing Severity Index Derived from Performance Degradation. AIAA Atmospheric and Space Environments Conference. DOI: 10.2514/6.2022-3906

**NEW PUBLICATION** Computational fluid dynamics (CFD) simulation methods are one of the most important design and development tools for aircraft – manned or unmanned. For the challenge of in-flight icing, special icing CFD codes have been developed such as FENSAP-ICE or LEWICE3D. These tools can simulate and predict the behavior of airfoils and aircraft in icing conditions. A common characteristic of these icing CFD tools is that they have been designed for manned aircraft — typically large passenger airplanes. When it comes to unmanned aircraft — unmanned aerial vehicles (UAVs), unmanned aerial systems (UAS), drones, or urban air mobility (UAM) vehicles — the existing icing CFD codes have significant limitations.

In a recent publication, we explore these limitations and gaps. The objective is to highlight the main challenges of icing CFD on unmanned aircraft and to suggest research steps to overcome them. In short, the main challenge is related to the low Reynolds numbers at which unmanned aircraft operate. Existing icing tools often use models that are based on experiments or are validated at high Reynolds numbers at which manned aircraft operate. The application of such models is thus limited when applied to unmanned aircraft. In addition, there are several physical flow phenomena that occur more frequently at low Reynolds numbers and which have gotten little attention for manned aircraft. The main issues are the following:

• Laminar-turbulent transition;
• Surface roughness modeling;
• Laminar separation bubbles;
• Turbulence models for low Reynolds numbers;
• Ice shedding model for rotors and propellers;
• Electrothermal ice protection systems;
• and lack of experimental validation data.

Based on the findings in the paper, we suggest the following steps to advance icing CFD on unmanned aircraft in order to unlock the full potential of digital twin development methods:

• Apply advanced turbulence models in existing icing CFD codes to improve the capabilities to capture low Reynolds number flow effects. In particular, LSB, laminar-turbulent transition, and surface roughness interactions.
• Generate validation datasets from experiments with conditions and geometries specifically for unmanned aircraft. Conduct validation of existing icing CFD tools for ice accretion, aerodynamic performance degradation, and ice protection systems.
• Develop or adapt existing models for ice shedding, surface roughness, and ice density that are valid at low Reynolds numbers.

Reference: Hann, R. (2022). UAV Icing: Challenges for computational fluid dynamic (CFD) tools. International Conference on Computational Fluid Dynamics, ICCFD11.

Announcing the 1^st 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 1^st October and for participation 21^st 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].

[6] Fortune Business Insights, “Unmanned Aerial Vehicle Market Size: UAV industry report 2027,” Unmanned Aerial Vehicle Market Size | UAV Industry Report 2027. [Online]. Available: https://www.fortunebusinessinsights.com/industry-reports/unmanned-aerial-vehicle-uav-market-101603. [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.