Ice Accretion Prediction Research Articles

Wing ice accretion prediction based on conditional generation adversarial network

The ice accretion on the aircraft's surface under low temperatures and high humidity is crucial for flight safety. With respect to the limitation of traditional icing simulation methods, it is very hard to predict exact ice profiles, which can extremely affect the flight performance of an aircraft. A conditional generative adversarial network (CGAN) is utilized to rapidly predict ice accretion and reconstruct three-dimensional ice patterns along the leading edge of a wing. The CGAN is trained using experimental data obtained from a wing with varying sweep angles. The results indicate that the CGAN achieves a high level of accuracy, specifically 97.5%, in predicting the similarity of ice shapes in the test set. When assessing the sample feature capture and prediction capability of the predictive model, it is shown that the CGAN exhibits superior predictive performance across different sample sizes.

Aircraft ice accretion prediction based on geometrical constraints enhancement neural networks

PurposeThe purpose of this study is to establish a novel airfoil icing prediction model using deep learning with geometrical constraints, called geometrical constraints enhancement neural networks, to improve the prediction accuracy compared to the non-geometrical constraints model.Design/methodology/approachThe model is developed with flight velocity, ambient temperature, liquid water content, median volumetric diameter and icing time taken as inputs and icing thickness given as outputs. To enhance the icing prediction accuracy, the model involves geometrical constraints into the loss function. Then the model is trained according to icing samples of 2D NACA0012 airfoil acquired by numerical simulation.FindingsThe results show that the involvement of geometrical constraints effectively enhances the prediction accuracy of ice shape, by weakening the appearance of fluctuation features. After training, the airfoil icing prediction model can be used for quickly predicting airfoil icing.Originality/valueThis work involves geometrical constraints in airfoil icing prediction model. The proposed model has reasonable capability in the fast assessment of aircraft icing.

Surrogate modelling for ice accretion prediction on wind turbine blades based on support vector regression

AbstractIce accretion poses a significant challenge, threatening the safety and operational stability of wind turbines in extreme cold weather. Accurate estimation of ice accretion is essential in minimizing this concern. Data‐driven techniques typically rely on extensive historical datasets, and simulation‐based methods may be time‐consuming. To address these challenges, surrogate modelling for ice accretion prediction is proposed, utilizing support vector regression. This surrogate model undergoes rigorous evaluation to ensure reliability. The surrogate model reduces time consumption by four orders of magnitude compared to simulation, with only a 2.5% decrease in accuracy. The analysis reveals that real‐time pitch angle adjustments, informed by both the surrogate model and meteorological data, can effectively reduce ice accretion on wind turbine blades in cold climates.

Surface Roughness in RANS Applied to Aircraft Ice Accretion Simulation: A Review

Experimental and numerical fluid dynamics studies highlight a change of flow structure in the presence of surface roughness. The changes involve both wall heat transfer and skin friction, and are mainly restricted to the inner region of the boundary layer. Aircraft in-flight icing is a typical application where rough surfaces play an important role in the airflow structure and the subsequent ice growth. The objective of this work is to investigate how surface roughness is tackled in RANS with wall resolved boundary layers for aeronautics applications, with a focus on ice-induced roughness. The literature review shows that semi-empirical correlations were calibrated on experimental data to model flow changes in the presence of roughness. The correlations for RANS do not explicitly resolve the individual roughness. They principally involve turbulence model modifications to account for changes in the velocity and temperature profiles in the near-wall region. The equivalent sand grain roughness (ESGR) approach emerges as a popular metric to characterize roughness and is employed as a length scale for the RANS model. For in-flight icing, correlations were developed, accounting for both surface geometry and atmospheric conditions. Despite these research efforts, uncertainties are present in some specific conditions, where space and time roughness variations make the simulations difficult to calibrate. Research that addresses this gap could help improve ice accretion predictions.

A framework for simulating snow accumulation and ice accretion on high-speed trains

A simulation framework for three-dimensional particle trajectory tracking, snow accumulation and ice accretion modelling on solid bodies in high-speed air flow is presented. The framework, named HADICE, solves the aerodynamic flow field as Reynolds averaged Navier-Stokes (RANS), with optional hybrid RANS/LES capabilities for modelling complex turbulent flows. Particle trajectory tracking is performed in an Eulerian-Lagrangian approach and wall collision is modelled using a hard collision model with reflect condition and a momentum based trap condition. A novel accurate and robust iterative evaluation method for the local collection efficiency is proposed for arbitrary 3D geometry and flow. Ice accretion is modelled in a multi-step approach and iced geometry is updated through mesh deformation using a radial basis function (RBF) interpolation method. Snow accumulation and ice accretion predictions based on the framework are validated against climatic wind tunnel experimental measurements. A full 3D simulation is demonstrated for snow accumulation and ice accretion on a 1:8 scaled high-speed train model. Using the presented framework, snow accumulation and icing simulation for high-speed trains can be conducted in an accurate and efficient manner, which is of great importance for physical investigations and the design of anti-snow and ice protection systems.

Prediction of ice accretion and aerodynamic performance analysis of NACA 2412 aerofoil

AbstractAdverse meteorological conditions often contribute to the formation of ice on aircraft wing section, engine nacelle and other parts leading to the loss of lift coefficient and increase in drag coefficient affecting aircraft control and stability. This paper addresses the problem of in-flight icing on an asymmetric aerofoil under three different ambient and cloud conditions. The study involves prediction of the leading-edge ice thickness using a numerical model developed from the mass and energy conservation law and Messinger freezing fraction model at the same Reynolds number. Later on, degradation in the aerodynamic performance of the iced aerofoil was also investigated using the computational fluid dynamics (CFD) technique, taking the flow field around a 2D aerofoil geometry into account. The aerodynamic study indicates that cumulus clouds embedded with stratified clouds contribute to the formation of mixed ice on aerofoil leading edge and causes the worst icing scenario reducing the lift coefficient to 90% and increasing the drag coefficient to 800% for the same ambient conditions.

Estimating and Mapping Extreme Ice Accretion Hazard and Load Due to Freezing Rain at Canadian Sites

The ice accretion load in Canadian structural design codes is developed based on an operational ice accretion prediction model. In the present study, three models are employed to predict the ice accretion amount on a flat surface and horizontal wire at Canadian sites. The results confirm that the model used by Canadian practice for predicting ice accretion leads to a conservative estimate as compared to the remaining two models. The results also indicate that the use of the Gumbel distribution for the annual maximum ice accretion is adequate for regions prone to ice accretion and that the lognormal distribution may be considered for regions with a moderate or negligible amount of ice accretion. Maps of the ice accretion hazard at five selected Canadian sites are developed. Statistical analysis of an equivalent wind speed that is concurrent with the iced wire is carried out, showing that the concurrent wind speed for the 50-year return period value of the annual maximum ice accretion amount is smaller than the 50-year return period value of the annual maximum wind speed. It is shown that the statistical characteristics of the annual maximum concurrent wind speed on iced wire differ from that of the annual maximum wind speed.

A mesh-free formalism for ice accretion prediction due to large drop impingement

This paper presents a novel mesh-free framework, using the smoothed particle hydrodynamics method (SPH), for the analysis of large drop impingement and ice accretion prediction on land engineering structures. The SPH method is an efficient and effective approach for the analysis of dynamic fluid flow problems involving free-surface interfaces. In this paper, it has been extended to include the thermodynamics of the freezing process. This allows not only the prediction of drop impact, splashing, and retraction, but also drop freezing and the formation of ice accretion. This mesh-free approach solves a set of non-linear transient dynamic continuum equations. The code also employs an enthalpy-based energy evolution algorithm, coupled with a phase transition equation. Surface tension and adhesion are handled with an approach that is shown to provide an accurate prediction of impinging water drop behavior. Numerical results of drop impact and freezing have been validated against a combination of experimental data and simple ice accretion models. In addition, ice accretions resulting from multi-drop impingement and freezing have been predicted. It is also shown that the developed method can simulate the influence of surface wettability on ice accretion shapes. The presented mesh-free formalism provides a next-generation approach to ice accretion prediction.

Numerical Modelling of Droplets and Beads Behavior over Super-Hydrophobic and Hydrophilic Coatings under in-Flight Icing Conditions

Current technology has produced a wide range of advanced micro-structured surfaces, designed for achieving the best wettability and adhesion performances for each specific application. In the context of in-flight icing simulations, this opens new challenges since the current most popular and successful ice accretion prediction tools neglect the details of the droplet behavior opting for a continuous film model. Here, a phenomenological model, following, in a Lagrangian approach, the evolution of the single droplets from the impinging to the onset of rivulets, is developed to simulate the performances of super-hydrophobic surfaces in icing application. Possible rebound and droplet spread on the impact, coalescence, single ice bead formation and droplet to rivulet transition are taken into account. The first validation shows how the models are able to predict the anti-icing capability of a super-hydrophobic surface coupled with a heating system.

A multi-autoencoder fusion network for fast image prediction of aircraft ice accretion

Icing scaling is of great importance for ice wind tunnel experiments and airworthiness certification. In this study, MAEFN (Multi-Autoencoder Fusion Network) method is developed to achieve fast icing image prediction. The novelty of MAEFN lies in its multi-module network structure, which enables the control of the training process. Such a structure can solve the problem that icing image generation networks are difficult to train directly and effectively adjust the accuracy of icing images. MAEFN is divided into four modules, i.e., feature extraction, mapping relationship construction, image generation, and image post-processing. Four neural networks are used to complete different steps, and all models are finally fused into the MAEFN model. MAEFN generates a 120 × 120 icing image through a seven-dimensional icing condition vector and achieves an average pixel accuracy of 98.85% on the supercritical wing. The standard deviation on the validation set is 0.0117, and the pixel accuracy of 680 out of 777 validation images exceeds 98%. It takes 2000 s to train 7000 icing images on V100 GPU. Furthermore, the efficiency of the fast icing algorithm is improved significantly by 25 000 times compared with the traditional icing algorithm, which only costs for a single case. Based on MAEFN, the exhaustive icing scaling method is realized, and similar icing results of four special ice shapes are found from 1 × 106 icing images in a few minutes. The proposed image generation method MAEFN can be further developed for highly complex icing images such as long-time icing results or other anti-icing tasks such as critical icing analysis.

Numerical modelization of contact angle hysteresis of falling droplet under enhanced lubrication approximation

Moving contact lines are involved in several engineering applications: in in-flight icing phenomenon, the eventual transition from droplet to rivulet or continuous film regime is crucial for the prediction of ice accretion over the aircraft surface; absorption process through structured packing is also characterized by a thin film flowing over the corrugated sheets. Disjoining pressure together with the assumption of a thin precursor film is largely used in numerical simulations of thin films and moving droplets in order to model the dynamics of moving contact lines and the surface wettability properties, in terms of imposed static contact angle. The disjoining pressure model was largely validated in case of falling films with the well known Voinov-Hoffman-Tanner law. On the other side, the capability of the disjoining pressure to model the contact angle hysteresis, which is a crucial parameter for predicting moving droplets behavior, has not been discussed yet. Here, numerical simulations of both falling films and moving droplets under lubrication approximation are conducted and the disjoining pressure model is used to predict the contact line dynamics. After verification of the full curvature implementation for a 1D falling film, the effective contact angle hysteresis is estimated for a moving droplet under different flow conditions and the transition from droplet to rivulet regime detected.

Neural Network Prediction for Ice Shapes on Airfoils Using iceFoam Simulations

In this article the procedure and method for the ice accretion prediction for different airfoils using artificial neural networks (ANNs) are discussed. A dataset for the neural network is based on the numerical experiment results—obtained through iceFoam solver—with four airfoils (NACA0012, General Aviation, Business Jet, and Commercial Transport). Input data for neural networks include airfoil and ice geometries, transformed into a set of parameters using a parabolic coordinate system and Fourier series expansion. Besides input features include physical parameters of flow (velocity, temperature, droplets diameter, liquid water content, time of ice accretion) and angle of attack. The novelty of this work is in that the neural network dataset includes various airfoils and the data augmentation technique being a combination of all time slices. Several artificial neural networks (ANNs), fully connected networks (FCNNs), and convolutional networks (CNNs) were trained to predict airfoil ice shapes. Two different loss functions were considered. In order to improve performance of models, batch normalization and dropout layers were used. The most accurate results of ice shape prediction were obtained using CNN and FCNN that applied batch normalization and dropout layers to output neurons of each layer.

Immersed Boundary Methodology for Multistep Ice Accretion Using a Level Set

The numerical prediction of in-flight ice accretion involves a sequential call to different modules including mesh generation, aerodynamics, droplet trajectories, wall heat transfer, ice accretion, and geometry update. Automation of this process is critical as these solvers are embedded in a time loop that is repeated several times to obtain an accurate ice shape prediction. The robustness of ice accretion tools is often limited by the difficulty of generating meshes on complex ice shapes and also by the geometry update that can exhibit overlaps in concave regions if not treated properly. An immersed boundary method combined with a level set has the potential to alleviate these issues. The objective of this paper is twofold: confirm the potential of this methodology and assess its accuracy against the usual body-fitted approach. The proposed methodology is tested on two-dimensional (2D) rime and glaze ice cases from the 1st AIAA Ice Prediction Workshop, showing good correspondence with the body-fitted approach. The new methodology also performs well for a 2D three-element airfoil configuration when a proper mesh refinement is used. The immersed boundary method combined with the level-set ice accretion provides a viable alternative to the body-fitted approach.

A Mesh-Free Formalism for Ice Accretion Prediction Due to Supercooled Large Drop Impingement

A Mesh-Free Formalism for Ice Accretion Prediction Due to Supercooled Large Drop Impingement

Penalization Method for Eulerian Droplet Impingement Simulations Toward Icing Applications

The numerical prediction of in-flight ice accretion generally involves geometry updates and remeshing as the ice builds up. However, the generation of body-fitted meshes around complex ice shapes is not trivial and can be repeated several times to obtain the final ice shape. The use of an immersed boundary method can simplify the mesh generation and help in the automation of the ice accretion process. This paper studies the application of an immersed boundary method to Eulerian droplet impingement simulations. A penalization method is suggested, requiring only the addition of source terms in the continuous form of the equations. The wall boundary condition must be treated with care to avoid droplets reinjection in the computational domain from a solid boundary. This is solved by the introduction of a droplet mask function in addition to the usual solid mask, providing an automatic detection of the wall boundary condition and therefore avoiding droplet reinjection. The approach is tested on canonical cylinder cases and on more realistic NACA0012 airfoil and ice horn cases. The results show that the solution from a body-fitted simulation can be reproduced using the penalization method.

Sensitivity Study of Ice Accretion Simulation to Roughness Thermal Correction Model

The effects of atmospheric icing can be anticipated by Computational Fluid Dynamics (CFD). Past studies show that the convective heat transfer influences the ice accretion and is itself a function of surface roughness. Uncertainty quantification (UQ) could help quantify the impact of surface roughness parameters on the reliability of ice accretion prediction. This paper aims to quantify ice accretion uncertainties and identify the key surface roughness correction parameters contributing the most to the uncertainties in a Reynolds-Averaged Navier-Stokes (RANS) formulation. Ice accretion simulations over a rough flat plate using two thermal correction models are used to construct a RANS database. Non-Intrusive Polynomial Chaos Expansion (NIPCE) metamodels are developed to predict the convective heat transfer and icing characteristics of the RANS database. The metamodels allow for the computation of the 95% confidence intervals of the output probability distribution (PDF) and of the sensitivity indexes of the roughness parameters according to their level of influence on the outputs. For one of the thermal correction models, the most influential parameter is the roughness height, whereas for the second model it is the surface correction coefficient. In addition, the uncertainty on the freestream temperature has a minor impact on the ice accretion sensitivity compared to the uncertainty on the roughness parameters.

Aerodynamic shape optimization of wind turbine blades for minimizing power production losses due to icing

Ice formation on a wind turbine alters the airfoil profiles of the blades and causes degradation in the aerodynamic performance of the wind turbine and the resulting power production losses. Since the blade profile plays a significant role in the icing of a blade, power production losses due to icing can be minimized by optimizing the blade profile against icing. In this study, blade profiles are optimized in order to minimize power production losses. A Gradient based aerodynamic shape optimization method is developed together with the Blade Element Momentum method and an ice accretion prediction tool in order to minimize the power production losses of horizontal axis wind turbines under various icing conditions. In an optimization study performed for the Aeolos-H 30 kW and NREL 5 MW wind turbines exposed to icing conditions up to 1 h, the power loss due to icing is reduced by about 4%.

Predicting Rotor Heat Transfer Using the Viscous Blade Element Momentum Theory and Unsteady Vortex Lattice Method

Calculating the unsteady convective heat transfer on helicopter blades is the first step in the prediction of ice accretion and the design of ice-protection systems. Simulations using Computational Fluid Dynamics (CFD) successfully model the complex aerodynamics of rotors as well as the heat transfer on blade surfaces, but for a conceptual design, faster calculation methods may be favorable. In the recent literature, classical methods such as the blade element momentum theory (BEMT) and the unsteady vortex lattice method (UVLM) were used to produce higher fidelity aerodynamic results by coupling them to viscous CFD databases. The novelty of this research originates from the introduction of an added layer of the coupling technique to predict rotor blade heat transfer using the BEMT and UVLM. The new approach implements the viscous coupling of the two methods from one hand and introduces a link to a new airfoil CFD-determined heat transfer correlation. This way, the convective heat transfer on ice-clean rotor blades is estimated while benefiting from the viscous extension of the BEMT and UVLM. The CFD heat transfer prediction is verified using existing correlations for a flat plate test case. Thrust predictions by the implemented UVLM and BEMT agree within 2% and 80% compared to experimental data. Tip vortex locations by the UVLM are predicted within 90% but fail in extreme ground effect. The end results present as an estimate of the heat transfer for a typical lightweight helicopter tail rotor for four test cases in hover, ground effect, axial, and forward flight.

Aerodynamic validation studies on the performance analysis of iced wind turbine blades

Ice accretion on wind turbine blades distorts blade profiles and causes degradation in the aerodynamic characteristic of the blades. In this study ice accretion on turbine blades are simulated under various icing conditions, and the resulting power losses are estimated. The Blade Element Momentum method is employed together with an ice accretion prediction methodology based on the Extended Messinger model in a parallel computing environment. The predicted iced profiles are first validated with the experimental and numerical data available in the literature. 2D flow solutions and aerodynamic loads over iced blade profiles are obtained with 3 different flow solvers of increasing fidelity; XFOIL, an open-source panel code coupled with a turbulent boundary layer model, SU2, an open-source RANS solver, and METUDES, an in-house DDES solver. The power production losses of a 30 kW wind turbine operating with iced blades are then investigated in detail. It is shown that the XFOIL-based tool developed for the performance analysis of iced wind turbines successfully predicts ice profiles on turbine blades under various icing conditions and the consequent power losses. About 20% power loss is predicted for a 30 kW wind turbine exposed to icing conditions for an hour.

An Assessment of LEWICE Roughness and Convection Enhancement Models

<div class="section abstract"><div class="htmlview paragraph">During aircraft design and certification, in-flight ice accretions are simulated using ice prediction codes. LEWICE, the ice accretion prediction code developed by NASA, employs a time-stepping procedure coupled with a thermodynamic model to calculate the location, size and shape of an ice accretion. LEWICE has been extensively validated for a wide range of icing conditions. However, continuing improvements to LEWICE predictive capabilities require better understandings of 1) the fundamental physics of turbulent flow generated by ice accretion roughness during an icing event and 2) the mechanisms responsible for convective enhancement of real ice accretion roughness. Recent experiments in the Icing Research Tunnel (IRT) at NASA Glenn Research Center have enabled significant insights into the nature of ice accretion roughness spatial and temporal variations. Other recent investigations have employed scans from the IRT to generate scaled test surfaces to investigate convection enhancement and skin friction interactions of flow over the surfaces with real ice roughness. For this investigation, the measurements of ice roughness characteristics, skin friction, and convection heat transfer enhancement are compared directly to predictions calculated using the LEWICE models. Issues associated with the LEWICE model for predicting laminar-to-turbulent transition are also addressed. The comparisons show that while the LEWICE models significantly overestimate equivalent sand-grain roughness heights and the local skin friction, the resulting LEWICE predictions of convection heat transfer generally agree within 25% of the experimental values for surfaces constructed with low thermal conductivity materials. At each step of the LEWICE model, methods to improve the predictions are provided.</div></div>