Machine Learning in Aerodynamic Shape Optimization

The deployment of self-learning computer algorithms that can automatically enhance their performance via experience is referred to as machine learning in ecommerce and is a crucial trend of the retail digital transformation. Machine learning algorithms can be unambiguously trained by analysing big datasets, identifying repeating patterns, relationships, and anomalies among all of this data, and creating mathematical models resembling such associations. These models are improved when the algorithms analyse ever-increasing amounts of data, providing us with useful insights into specific ecommerce-related events and the links between all the variables that underlie them. A tool that has been quite effective in studying current affairs, predicting future trends, and making data-driven decisions. The present work investigates the implementation of machine learning algorithms to predict the user intention for purchasing a product on a specific store's website. An Online Shoppers Purchasing Intention data set from the UC Irvine Machine Learning Repository was used for this investigation. In this study, two classification-based machine learning algorithms i.e. Stochastic Gradient Descent (SGD) algorithm and Random Forest algorithm were used. SGD algorithm was used for first time in prediction of the online user intention. The results showed that the Random Forest resulted in the highest F1-Score of 0.90 in contrast to the Stochastic Gradient Descent algorithm.

show abstract

“…Machine learning is helpful in a variety of fields and has the ability to develop throughout time. [11][12][13][14][15].…”

Section: Introductionmentioning

confidence: 99%

Machine Learning Approach for Prediction of the Online User Intention for a Product Purchase

Kumar¹,

Macedo

et al. 2023

IJRITCC

View full text Add to dashboard Cite

show abstract

“…Peng et al [9] presented an ANN that can learn physical laws and aerodynamic equation while providing accurate lift coefficient predictions. More literature can be found in [10], which provides an overview of the wide application of machine learning methods in aerodynamic shape optimization. However, the generalization performance of these surrogate models is insufficient, and the fact that they are essentially black-box makes the designers less confident in the results.…”

Section: Introductionmentioning

confidence: 99%

Knowledge-embedded meta-learning model for lift coefficient prediction of airfoils

Xie¹,

Wang²,

Zhang³

2023

Preprint

View full text Add to dashboard Cite

Aerodynamic performance evaluation is an important part of the aircraft aerodynamic design optimization process; however, traditional methods are costly and time-consuming. Despite the fact that various machine learning methods can achieve high accuracy, their application in engineering is still difficult due to their poor generalization performance and "black box" nature. In this paper, a knowledge-embedded meta learning model, which fully integrates data with the theoretical knowledge of the lift curve, is developed to obtain the lift coefficients of an arbitrary supercritical airfoil under various angle of attacks. In the proposed model, a primary network is responsible for representing the relationship between the lift and angle of attack, while the geometry information is encoded into a hyper network to predict the unknown parameters involved in the primary network. Specifically, three models with different architectures are trained to provide various interpretations. Compared to the ordinary neural network, our proposed model can exhibit better generalization capability with competitive prediction accuracy. Afterward, interpretable analysis is performed based on the Integrated Gradients and Saliency methods. Results show that the proposed model can tend to assess the influence of airfoil geometry to the physical characteristics. Furthermore, the exceptions and shortcomings caused by the proposed model are analysed and discussed in detail.

show abstract

“…These applications include aerodynamic design optimization (Mengistu and Ghaly, 2008), flow modeling (J. Li et al, 2022), performance predictions (Liu and Karimi, 2020;Taghinezhad and Sheidaei, 2022), and flow control (Kamari et al, 2018;Yu et al, 2019), among others (Z. Li and Zheng, 2017).…”

Section: Introductionmentioning

confidence: 99%

Deep neural networks for predicting and analyzing the performance of air-jets for active flow control in an axial compressor

Elhawary,

Romanò,

Loiseau

et al. 2023

European Conference on Turbomachinery Fluid Dynamics and Thermodynamics

View full text Add to dashboard Cite

Active flow control (AFC) via air-jets has proved effectiveness in mitigating rotating stall in axial compressors. However, achieving robust and reliable control strategies remains challenging. In this study, a deep neural network (DNN) is proposed to investigate the effect of air-jets control parameters on the performance of an axial compressor (CME2). We consider a control system consisting of 20 pairs of air-jets positioned circumferentially upstream of the leading edge of the rotor blades. Three control parameters are investigated: the absolute injection angle (α inj ), the number of injector pairs (N inj ), and the injection velocity (V inj ). The performances of the compressor are evaluated in terms of the power balance (PB) and the surge margin improvement (SMI tot ) at three rotational velocities Ω = 3200 RPM, 4500 RPM and 6000 RPM. We build on a set of experiments previously carried-out in our lab and use it to train and validate the DNN model. The results show that the optimal air-jet parameters giving high PB can be obtained at a velocity ratio (V R ) and an angle of attack (β att ) varying from 1.0 m/s to 1.7 m/s and −14 • to 12 • , respectively. High SMI tot values are obtained at V R and β att ranging from 1.4 m/s to 1.9 m/s and −7 • to 5 • , respectively. It is also shown that the surge margin improvement based on mass flow rate (SMI Q ) increases as the injection mass flow rate (Q inj ) or the injection velocity (V inj ) increases, while the total-to-total pressure rise (∆p tt ) tends to decrease at the same conditions. These outcomes open up to the possibility of generalizing our findings by interpreting the underlying physics of the control process, with the ultimate goal of proposing an AFC strategy applicable to other compressor geometries.

show abstract

Machine Learning in Aerodynamic Shape Optimization

Cited by 6 publications

References 418 publications

Machine Learning Approach for Prediction of the Online User Intention for a Product Purchase

Machine Learning Approach for Prediction of the Online User Intention for a Product Purchase

Knowledge-embedded meta-learning model for lift coefficient prediction of airfoils

Deep neural networks for predicting and analyzing the performance of air-jets for active flow control in an axial compressor

Contact Info

Product

Resources

About