Data-Driven Machine Learning Model for Aircraft Icing Severity Evaluation

Li, Sibo; Qin, Jingkun; Paoli, Roberto

doi:10.2514/1.i010978

Cited by 8 publications

(8 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As described in the previous section, the aircraft icing process is a complex interaction of multiple variables and explicitly modeling the icing formation process often requires computationally expensive and/or cumbersome treatments to calculate the ice displacement and accretion along the wing, such as remeshing [22]. Data-driven methods can help alleviate this constraint by applying regression analysis and machine learning models to predict the aircraft icing based on icing data collected in experimental campaigns and/or numerical simulations [24]. Within the domain of aircraft icing, ML is applied in two main areas: ice shape prediction and icing severity evaluation.…”

Section: Data-driven Modeling For Aircraft Icingmentioning

confidence: 99%

“…Since the ice accretion process shows strong nonlinearity, linear algorithms such as linear regression [44] and logistic regression [45] are not suitable [24]. Common nonlinear algorithms include the classification and regression trees (CART) [46], Naïve Bayes (NB) [47], k-Nearest Neighbors (KNN) [48] and Support Vector Machines (SVM) [49].…”

Section: Machine Learningmentioning

confidence: 99%

“…Besides the conventional ML models mentioned above, ensemble machine learning algorithms are also widely used in aircraft icing applications [24,26]. Ensemble ML methods create multiple models and then combine them to generate improved prediction results [50].…”

Section: Machine Learningmentioning

confidence: 99%

“…In aircraft icing severity evaluation, the use of ML methods has been shown to aid in reducing the computational time, enabling aircraft safety improvement. Li et al [24] proposed a method for aircraft icing severity prediction at different flight conditions based on machine learning model XGBoost. Based on the numerical modeling, six flight conditions (flight speed, angle of attack, exposure time, LWC, MVD and freestream temperature) are considered as input to the ML model.…”

Section: Machine Learning For Icing Severity Predictionmentioning

confidence: 99%

“…Based on the training data, ML models are capable of addressing strong nonlinearity with the aid of constructing black-box input-output mapping [23]. Due to the complex interaction of multiple flight conditions, the mapping relationship between the input flight conditions and the output aircraft icing severity features is likely to be strongly nonlinear [24]; thus, ML has been implemented in several applications in aircraft icing to predict ice shape [25], icing area, maximum ice thickness, icing severity level [24,26] and the effect of ice on the aircraft aerodynamic performance [27]. The details will be given in Section 4.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Aircraft Icing Severity Evaluation

Paoli

2022

Encyclopedia

Self Cite

View full text Add to dashboard Cite

Aircraft icing refers to the ice buildup on the surface of an aircraft flying in icing conditions. The ice accretion on the aircraft alters the original aerodynamic configuration and degrades the aerodynamic performances and may lead to unsafe flight conditions. Evaluating the flow structure, icing mechanism and consequences is of great importance to the development of an anti/deicing technique. Studies have shown computational fluid dynamics (CFD) and machine learning (ML) to be effective in predicting the ice shape and icing severity under different flight conditions. CFD solves a set of partial differential equations to obtain the air flow fields, water droplets trajectories and ice shape. ML is a branch of artificial intelligence and, based on the data, the self-improved computer algorithms can be effective in finding the nonlinear mapping relationship between the input flight conditions and the output aircraft icing severity features.

show abstract

Section: Data-driven Modeling For Aircraft Icingmentioning

confidence: 99%

Section: Machine Learningmentioning

confidence: 99%

Section: Machine Learningmentioning

confidence: 99%

Section: Machine Learning For Icing Severity Predictionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Aircraft Icing Severity Evaluation

Paoli

2022

Encyclopedia

Self Cite

View full text Add to dashboard Cite

show abstract

A machine learning study to predict wind-driven water runback characteristics

Wang,

Hu,

et al. 2023

Physics of Fluids

View full text Add to dashboard Cite

The unsteady runback behavior of wind-driven runback water film (WDRWF) flows over aircraft surfaces has a significant impact on the aircraft icing process, one of the most significant aviation hazards in cold weather. The limited understanding of the complex multiphase interactions between freestream airflow, water film motion, and solid airframe surface makes conventional theoretical/numerical methods unable to precisely simulate WDRWF flow. Machine learning-based techniques can accurately capture complex physics using data, making it an attractive alternative to conventional methods. In this study, machine learning methods are used to predict the evolution of the front contact point (FCP) of WDRWF flow and film thickness distribution (FTD) of WDRWF flow. For FCP prediction, the performance of the Light Gradient-Boosting Machine (LightGBM) and Multi-Layer Perceptron is compared quantitatively. They perform well in capturing intermittent and smooth features, respectively. For the prediction of the spatial-temporal evolution of FTD, a computationally efficient deep neural network architecture named ConvLSTM-AutoEncoder was developed, which predicts a future FTD based on a sequence of FTDs in the past. The robustness of the ConvLSTM-AutoEncoder model to noisy input FTD is demonstrated. The generalizability of the three models is evaluated by applying the trained models to unexplored datasets. Based on the proposed techniques' generalizability, robustness, and computational efficiency, machine learning-based methods are demonstrated to be powerful tools in predicting the complex unsteady characteristics of the multiphase WDRWF flows.

show abstract

Development process for preliminary aircraft sizing methods with regard to new technologies

Schneider

Strohmayer

2023

AEAT

View full text Add to dashboard Cite

Purpose The purpose of this study is to develop and describe a process which can be applied to develop new methods in the context of preliminary aircraft sizing in a successful and efficient way. Design/methodology/approach The tasks to development new aircraft sizing methods are systematically analyzed. In particular, repeating and nonrepeating tasks and common or unique tasks. Then ordered in a sequence and described generically. Findings A development process for new aircraft design methods which are necessary for new technologies or configurations is introduced and explained step by step. Practical implications Introducing the capability to deal with new technologies or configurations, aircraft design tools or aircraft concepts requires new sizing methods. Originality/value The paper presents a systematic approach which can be used to develop a great amount of new sizing methods with a comparable usability and quality standard in an efficient and effective way.

show abstract

Data-Driven Machine Learning Model for Aircraft Icing Severity Evaluation

Cited by 8 publications

References 15 publications

Aircraft Icing Severity Evaluation

Aircraft Icing Severity Evaluation

A machine learning study to predict wind-driven water runback characteristics

Development process for preliminary aircraft sizing methods with regard to new technologies

Contact Info

Product

Resources

About