Scholarlink

<p>Aerodynamics, a main discipline in fluid mechanics that examines the behavior of air as it interacts with solid surfaces, such as wings or airfoils. This paper uses machine learning (ML) algorithms to predict drag coefficient (CD) and lift coefficient (CL) in aerodynamic modeling. The main target is to determine the best models for predicting CD and CL based on error metrics and coefficient of determination (R<sup>2</sup>). Linear, Ridge, Random Forest, Gradient Boosting, Support Vector Regressor, and XGBoost were used for CD and CL prediction. Through experiments, we noticed that XGBoost and Random Forest (RF) outperformed CD and CL forecasts. XGBoost achieved a mean squared error (MSE) of 0.0002 and a R<sup>2</sup> of 0.9916 for CD prediction, followed by RF by 0.0002 and 0.9915. XGBoost surpassed other models in CL predictions with an MSE of 0.0005 and R<sup>2</sup> of 0.9919; Random Forest was a close contender (MSE: 0.0009, R<sup>2</sup>: 0.9839). This approach enhances aerodynamic coefficient prediction.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Prediction of aerodynamic coefficients based on machine learning models

  • 期刊论文

  • 发表日期：

    2025年3月24日

查看全文
  • Ahmed M. Elshewey,
  • Mohamed A. Aziz,
  • Osama A. Gaheen,
  • Mohamed S. Sawah,
  • A. Abd ELhamid,
  • Ahmed M. Osman

摘要

Aerodynamics, a main discipline in fluid mechanics that examines the behavior of air as it interacts with solid surfaces, such as wings or airfoils. This paper uses machine learning (ML) algorithms to predict drag coefficient (CD) and lift coefficient (CL) in aerodynamic modeling. The main target is to determine the best models for predicting CD and CL based on error metrics and coefficient of determination (R2). Linear, Ridge, Random Forest, Gradient Boosting, Support Vector Regressor, and XGBoost were used for CD and CL prediction. Through experiments, we noticed that XGBoost and Random Forest (RF) outperformed CD and CL forecasts. XGBoost achieved a mean squared error (MSE) of 0.0002 and a R2 of 0.9916 for CD prediction, followed by RF by 0.0002 and 0.9915. XGBoost surpassed other models in CL predictions with an MSE of 0.0005 and R2 of 0.9919; Random Forest was a close contender (MSE: 0.0009, R2: 0.9839). This approach enhances aerodynamic coefficient prediction.