DATA-DRIVEN OPTIMIZATION AND PERFORMANCE ANALYSIS OF AIRCRAFT STRUCTURAL DESIGN USING MATERIAL PROPOERTIES AND COMPUTATIONAL EFFICIENCY METRICS

Abstract

Purpose: This study aims to develop a data-driven framework for optimizing aircraft structural design by

balancing weight reduction, structural strength, and computational efficiency, addressing the limitations of

traditional physics-based approaches in large-scale optimization. Methodology: A structured dataset of

300 samples with 22 variables were used, incorporating material properties (Young's modulus, density,

tensile strength), structural parameters, and computational efficiency metrics. Machine learning models,

namelyLogisticRegressionandRandomForest,wereappliedtopredictweightefficiency.Correlation

analysisandfeatureimportancetechniqueswereemployedtoidentifykeyinfluencingvariables,along

with evaluation of computational performance through optimization time and iteration count. Findings:

The RandomForest model had a higher ability to model nonlinear relationships with an accuracy of 61.67% as

compared to 58.33% to the Logistic Regression one. The most important factors were identified to be

the material properties, especially the density and tensile strength, which affected the structural

performance and structural parameters increased the design flexibility. Computational studies showed that

There was variability in optimization time, and thus efficient algorithms are important. Conclusion: The

The proposed framework supports efficient and scalable aircraft structural optimization by enabling

identification of critical design parameters and improving decision-making in aerospace engineering. This

paper offers a new combination of machine learning and aircraft structural design with a data-driven

solution that will increase optimization efficiency and lead to scalable solutions in aerospace design.