Application of Machine Learning and Multi-Disciplinary/Multi-Objective Optimization Techniques for Conceptual Aircraft Design

Mattos, Bento Silva de; Bortolete, Felipe A.; Fregnani, José A. T. G.; Jorge, Ariosto Bretanha; Alves, William M.; Cruz, Ronaldo V.

doi:10.4322/978-65-86503-71-5.c05

Cited by 1 publication

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…By a slight reduction of the wing aspect ratio of the Design ID 303CE (Table 22), keeping the remained parameters but the root airfoil section of the wing unchanged, a baseline airplane was created for this last optimization task (Table 24). Mattos et al [69] applied a flutter speed constraint to an MDO framework for airliner conceptual design. As optimization results, a noticeable reduction of wing aspect ratio was observed when compared with an optimization where no flutter speed constraint was applied.…”

Section: Robust Optimization Of a Chosen Individual Of A Previous Opt...mentioning

confidence: 99%

Airplane Optimal Robust Design Combined with an Entropy-Statistics Unsupervised Learning Classification Algorithm

Mattos¹,

Fregnani²,

Jorge³

2022

Uncertainty Modeling: Fundamental Concepts and Models

Self Cite

View full text Add to dashboard Cite

The present work describes an airliner design platform that has been developed over several years and its application to design airliners based on an efficiency index developed by the authors of the present work. Among other characteristics, the design tool can handle airplane geometries with detail and count on a unique and accurate neural network system to predict aerodynamic coefficients. A new feature for product classification was incorporated into the framework. It is based on entropy statistics and was calibrated for jet transport airplanes and the results of classification are analyzed in the present work. The design of a Middle of Market airplane was then carried out with the classification provided by entropy statistics set as a constraint. Middle of the Market is a term that was established to address the commercial aircraft market segment that encompasses long-haul sectors from 3500 nm to 5000 nm, and airplanes with single-class passenger capacity between 220 and 250. Until now, this segment was not extensively explored, and some airlines have been showing interest to purchase airplanes better suited to it = this category. Multi-objective optimizations were carried out to determine the best aircraft design that fits into this concept's range and capacity specifications. The unsupervised learning algorithm based on entropy statistics was applied for airplane classification into four labels, classic, niche, failure, and breakthrough designs. A higher fidelity engine model optimization task was also carried out to search for optimal classic designs. In addition, a robust optimization considering a fuel price variation on Direct Operating Cost was carried out. All results raised an important discussion about the benefits and drawbacks of a trijet configuration for transport airplanes with a capacity of over 200 passengers.

show abstract

Section: Robust Optimization Of a Chosen Individual Of A Previous Opt...mentioning

confidence: 99%

Airplane Optimal Robust Design Combined with an Entropy-Statistics Unsupervised Learning Classification Algorithm

Mattos¹,

Fregnani²,

Jorge³

2022

Uncertainty Modeling: Fundamental Concepts and Models

Self Cite

View full text Add to dashboard Cite

show abstract

Application of Machine Learning and Multi-Disciplinary/Multi-Objective Optimization Techniques for Conceptual Aircraft Design

Cited by 1 publication

References 0 publications

Airplane Optimal Robust Design Combined with an Entropy-Statistics Unsupervised Learning Classification Algorithm

Airplane Optimal Robust Design Combined with an Entropy-Statistics Unsupervised Learning Classification Algorithm

Contact Info

Product

Resources

About