Effects of strain rate variation on the shear adhesion strength of impact ice

https://doi.org/10.1016/j.coldregions.2020.103168Get rights and content

Highlights

  • Accreted ice on rotating structures can be shed naturally through angular acceleration (increasing centrifugal force).

  • A relationship between interfacial strain rate and ice adhesion strength is sought.

  • In-flight impact ice appears to follow Glen's power law previously documented in studies on glacier ice in compression.

  • The effects of strain rate and temperature on adhesion strength are statistically significant for centrifugally-loaded ice.

Abstract

In-flight ice accretion on fixed-wing aircraft and rotorcraft can be catastrophic if not mitigated. Most modern ice protection systems are active systems, which require electrical or mechanical power to remove accreted ice, thereby increasing weight, cost, and complexity. Scientists and engineers now seek passive, erosion-resistant materials and coatings with low ice adhesion strength. Ideally, such materials, when applied to vulnerable components of an aircraft, would cause any ice to shed off the surface under normal aerodynamic loading.

To aid in the development of low-ice-adhesion-strength materials, the growth and structural behavior of impact ice in a wide range of atmospheric conditions must be characterized. The structural behavior of ice has been examined under pure shear, tension, compression, and mixed-mode loading. However, one important loading consideration that has not been widely investigated on atmospheric ice is strain rate.

Knowledge of the relationship between impact ice adhesion strength and strain rate is important because it can be used to design future ice protection systems, and it may dictate the appropriate course of action for a pilot flying through icing conditions—for instance, whether a helicopter pilot should increase the rotor speed rapidly or slowly to induce shedding of the ice.

NASA Glenn Research Center funded the design and construction of a new centrifuge-style ice adhesion test rig (“AJ2”) by the Penn State AERTS lab. The ice is accreted dynamically by spinning flat metal test coupons at high speed inside a simulated icing cloud environment. The design and analysis of the AJ2 rig is described in detail in this paper.

Experiments were performed using AJ2 to investigate how the adhesion strength of impact ice related to the strain rate applied to it. Stainless steel test coupons of known surface roughness were tested in a range of environmental temperatures. The strain rates applied to the ice ranged between 5E−8 and 5E-5 s−1. It was discovered that a similar power function exists between strain rate and adhesion strength as found in the freezer-ice studies described in the literature. Despite scatter in the data, regression analysis determined the relationship between strain rate, temperature, and adhesion strength to be statistically significant. The power “1/n” for a coupon roughness of 64 nm (Sa) was greater than that of the 80-nm coupon; this was the case for both tested temperatures. However, for the relatively smooth surfaces tested, regression analysis suggested that the surface roughness had negligible effect on adhesion strength. Lower temperatures caused a higher power “1/n” and coefficient “c” in the power function. The variation of the coefficient with temperature is consistent with Glen's power law for the creep of glacier ice in compression. However, Glen did not observe a variation of the power with temperature. The value of “n” in the current study ranged from 2.6 for the smoothest sample at the coldest temperature, to 8.8 for the roughest sample at the warmest temperature. In most cases of the current study, “n” was within the range of previously-reported values in literature (1.5 to 6).

Introduction

In-flight ice accretion threatens private, commercial, and military aircraft alike, and both fixed-wing and rotary-wing vehicles. According to the European Aviation Safety Agency, encounters with in-flight icing conditions caused 8% of “serious incidents” and 20% of accidents in Europe between 2009 and 2014 (2014 Annual Safety Review, 2015). Similarly, between 2010 and 2014, the National Transportation Safety Board (NTSB) reported 52 accidents and 78 fatalities in the United States caused by airframe icing (Eick, 2015). Traditionally, aircraft icing has been addressed with “active” de-icing systems such as pneumatic boots, resistive heating, and bleed air, among other methods (DiPlacido et al., 2016; Wright, 2004; Whitfield, 2016; Werner, 1975; TKS Ice Protection, 2020). Passive materials with low ice adhesion strength may provide a lighter, less expensive, and less complex alternative to active systems. Companies strive to develop so-called “ice-phobic” polymer coatings that could be applied to aircraft components and turbine blades so that any accreted ice would easily shed off under shear loading. The shear forces acting on the ice typically come either from centrifugal force (in the case of helicopter blades, engine rotors, and wind turbines) or from aerodynamic loading on fixed structures. The effort to develop passive, erosion-resistant, ice-protective coatings is crippled by a fundamental lack of understanding of how impact ice interacts with substrates under different environmental and loading conditions. Ice accretion is a complex transient engineering problem that requires multiscale physics modeling. Current adhesion models are still in their infancy and lack validation data (Kreeger et al., 2016). Integration of low ice adhesion strength coating with mechanical active ice protection systems is being considered. Existing coatings that prevent ice accretion are not robust enough to survive representative aircraft environments.

Before ice-protective materials/coatings can be designed (maybe even optimized) and integrated to air vehicles, researchers must be able to reliably quantify and predict impact ice adhesion strength in a wide range of environmental and loading conditions. Previous studies have established the effects of temperature and surface roughness on the shear adhesion strength of “impact ice,” or ice impacting a surface at a known velocity (Soltis et al., 2015). However, research into strain rate effects on ice adhesion has been severely lacking, especially concerning atmospheric impact ice (Work and Lian, 2018).

The meaning of the term “strain rate” depends on the context. Geological studies of glacier ice typically define “strain rate” (or “creep rate”) as the rate of displacement of a given point in the glacier ice as it experiences compressive stress. For impact ice adhesion purposes, “strain rate” is the time rate of change of strain experienced at the interface between the ice and the substrate, which occurs as a result of transient loading on the ice. Atmospheric ice on fixed wings and airframe surfaces is unlikely to experience high strain rate conditions in steady flight because the loading on the ice is relatively constant. However, on rotating frames such as helicopter blades, engine fans, and wind turbines, any change in angular speed causes a significant change in the centrifugal force acting on the ice, thereby changing the strain experienced at the interface. Additionally, the process of ice accretion on a rotating body in and of itself increases the centrifugal force that strains the interface.

The nature of the relationship between impact ice adhesion strength and strain rate is important because it may dictate the appropriate course of action for a pilot flying through icing conditions—for instance, whether the engine fan speed should be increased slowly or rapidly to induce shedding of accreted ice. This information would also be useful to implement into wind turbine control programs as part of a safe de-icing procedure. Knowing how strain rate affects adhesion strength could be a powerful tool for constructing and validating comprehensive adhesion prediction models, as well as developing future anti-icing and de-icing strategies. To this end, the current study seeks new data relating the adhesion strength of impact ice on a surface to the applied strain rate.

Although most previous ice strain rate studies were performed in compression and/or used lab-grown “freezer ice” instead of dynamically-accreted impact ice, the trends established in literature may provide a baseline understanding of the creep behavior of ice under loading. Additionally, the literature provides insight into the benefits and drawbacks of different experimental setups. The following sections describe past efforts to examine the relationship between stress and strain rate of ice.

For the purposes of adhesion modeling and development of ice-protective coatings, data pertaining to ice loaded in shear is preferred, since for ice, shear strength is less than tensile strength. However, since it is difficult to uniformly load ice in shear, many of the studies in literature concerning strain rate and adhesion strength of ice have been performed in compression. The key findings include:

  • The creep rate (or strain rate) of ice is related to the compressive stress by a power function, for which the coefficient depends on the environmental temperature (Glen (Glen, 1955), Goldsby (Goldsby and Kohlstedt, 2001)). The results of Glen's glacier ice experiments are shown in Fig. 1.

The relation:ε̇=kσnwhere k decreases with decreasing temperature came to be known as “Glen's law.”

According to Goldsby (Goldsby and Kohlstedt, 2001), numerous later authors obtained a similar trend as Glen, with power “n” ranging between 1.5 and 6.

  • For strain rates above a certain threshold, the ice undergoes a ductile-to-brittle transition, and the stress becomes independent of strain rate (Gold (Gold, 1970), Cole (Cole, 1987)). Both Cole and Gold observed a region where the logarithmic slope between stress and strain rate flattened out. Each study found a different value for the transition strain rate, likely because Cole used “granular” ice with randomly-oriented grains and Gold used “columnar” ice. It is worth noting that, to date, the orientation and sizes of grains inside impact ice is unknown. If these parameters affect the strain rate-adhesion strength relationship, it is imperative that impact ice is used for adhesion modeling.

  • The creep behavior of the ice is influenced by the ice grain size and orientation (Cole (Cole, 1987), Hewitt (Hewitt, 2017)). Cole found that for ice loaded with a given strain rate, the peak stress (or strength) of the ice decreased with increasing grain size. Hewitt theorized that macroscopic creep rate of ice in compression is dictated by the “rate-limiting process,” which depends on grain size, grain orientation relative to the loading direction, and temperature.

While these findings cannot be directly applied to shear adhesion modeling of atmospheric impact ice without further validation, these trends will serve as initial guidelines for impact ice adhesion strength testing.

The numerous studies of ice creep under compressive loads provide a starting point for understanding how ice deforms and how such processes are impacted by temperature, crystal structure, and other parameters. It is unclear whether these principles are directly applicable to a pure-shear loading conditions. Relatively few published studies concerning ice adhesion vs. strain rate have been performed in shear, compared with the number performed in compression. Two of the most notable shear experiments are from Parameswaran (Parameswaran, 1981) and Susoff et al. (Susoff et al., 2013). It should be noted that these studies were performed on ice grown in molds, not atmospheric ice. The key findings for lab-grown ice in shear are as follows:

  • Parameswaran (Parameswaran, 1981) used a tensile tester to pull steel beams out of ice blocks at specified strain rates. He found a power function similar to Glen's law, with n = 5.77. This suggests that Glen's law may be applicable to shear de-icing research.

  • In a similar test on coated aluminum piles frozen in ice, Susoff (Susoff et al., 2013) found that for high strain rates (2.8 × 10−3 to 5.6 × 10−1 s−1) the adhesion strength of ice on bare aluminum is independent of the applied strain rate. However, silicone rubber-coated aluminum exhibited the normal power function behavior, as shown in Fig. 2.

The strain rate independence region for bare aluminum is reminiscent of Cole and Gold's findings in compression. This also suggests that the (visco-)elasticity of ice-protective materials and coatings plays a significant role in the strain rate-adhesion strength relationship at the interface. Therefore, when optimizing ice-protective coating chemistries, special attention should be paid to the visco-elastic properties of the coating.

Section snippets

Test apparatus

The aforementioned experimental studies from literature have two things in common: they all used freezer ice, and they all involved direct mechanical testing, meaning that the ice was loaded via direct contact with a moving object, often in a tension/compression testing machine. External interaction with accreted ice could introduce unwanted energies that could influence the debonding behavior. Ice adhesion strength testing of atmospheric ice must be performed using dynamically-accreted impact

Mathematical modeling

Strain rate adhesion tests on AJ2 were performed in a similar fashion to fixed-speed adhesion tests (Soltis et al., 2015), but with some key differences. In a fixed-speed adhesion test, the icing cloud is sprayed until a shedding event occurs. In a strain rate test, the cloud is purposely stopped before the ice can shed, so that the effects of rotational acceleration can be examined. After a certain amount of time, the spray is deactivated; then the motor speed is ramped up at a known linear

Experimental results

The relationship between strain rate and shear adhesion strength observed in the AJ2 experiments in the current study followed a power function similar to the trends described in literature (Glen, 1955; Cole, 1987; Parameswaran, 1981). Coupons of three different surface roughness values were tested. They included two polished with a 320-grit polishing pad and one polished with a 600-grit polishing pad. The surface roughness values obtained from the polishing process were quantified with optical

Conclusions

A small-scale ICAT-type ice adhesion test rig, dubbed “AJ2,” was successfully designed and fabricated at Penn State to conduct strain rate testing of impact ice. A NASA-grade icing nozzle was used to simulate a cloud of supercooled droplets inside a walk-in freezer. Adhesion tests on AJ2 can be performed at rotor speeds up to 5000 rpm, corresponding to a droplet impact speed of 180 kts, which is on the order of actual droplet impact speeds for large-scale wind turbine blades and inboard areas

Future work

Shortly after this study was completed, some modifications were made to the AJ2 test rig. The strain gauge assembly on the beams was replaced with PCB piezoelectric strain sensors, mounted on the inner root of each beam. The new gauges measure dynamic strain instead of static strain, so the voltage hovers around zero until an instantaneous shed event occurs. The decision to change the strain gauges was motivated by the need to frequently replace the old strain gauge assembly. The protective

Declaration of Competing Interest

This material is based upon work supported by NASA under Award No. 80NSSC17P1049. Any opinions, findings, and conclusions or recommendations expressed in this publication are those of the authors and do not necessarily reflect the views of NASA.

References (22)

  • D.L. Goldsby et al.

    Superplastic deformation of ice: Experimental observations

    J. Geophys. Res.

    (2001)
  • Cited by (8)

    • Formation mechanism of freezing interface strain and effect of different factors on freezing interface strain

      2023, International Journal of Thermal Sciences
      Citation Excerpt :

      For example, a freezing disaster in southern China in 2008 paralyzed the power network, the direct economic loss of which was over 150 billion RMB [1,4]. Once the accumulated ice adheres to the surfaces of aircraft wings, sensors, and other components, the aerodynamic performance, handling, and stability of aircraft can be affected, compromising flight safety [5,6]. In cold regions, a mixture of ice, snow, and other materials adhering to the chassis component surfaces of high-speed trains can affect their operational safety, the service life of components, and passenger comfort [7,8].

    • Impact ice adhesion strength of stainless steel 304 as determined on a centrifuge test stand

      2022, Cold Regions Science and Technology
      Citation Excerpt :

      Further discussion of this liquid like water layer is beyond the scope of this work. A Scheffe Test, a post ANOVA test, was performed on the data in Table 2 since it was reported by Douglass (Douglass and Palacios, 2021) that Sa was not a statistically relevant predictor of τ. The Sa evaluated in that work ranged from 64 to 80 nm.

    View all citing articles on Scopus
    View full text