Melih YıldızEnglish

Kongre bildirisi · 16th EASN International Conference · 27–30 Ekim 2026

Machine Learning-Assisted Multidisciplinary Design Optimization in Aircraft Conceptual Design Using Digital Twin Technology

Elif Zeyneb Tekin, Melih Yıldız (sunan)

Department of Aeronautical Engineering and ERHAM Aviation Application and Research Center, Erciyes University, Kayseri, Türkiye · ISAE-SUPAERO, Toulouse, France

✓ Sunuldu: 29 Ekim 2026, 14:40 · salon 05-114 · Digital Engineering for Aerospace Systems III oturumu · sunan Melih Yıldız

A full optimization cycle in 5 minutes instead of 1.2 hours, with each prediction's uncertainty stated up front.

One-pager (PDF, A4)One-pager (HTML)Tam metin: yayına hazırlanıyor

Özet

Conceptual design locks in most of an aircraft's life-cycle cost, yet it is the stage where physics-based analysis is used least, because each design point costs too much. Hybrid-electric and hydrogen configurations have no historical data to fall back on, so their design space has to be explored with physics.

We built a digital twin for this stage: a physics-based sizing loop for a 70-seat hybrid-electric regional aircraft, calibrated against the ATR 72-500, with Gaussian process surrogates and an NSGA-II optimizer on top. Constraints are trusted only when the surrogate is confident, and new physics runs are placed only where the Pareto front is uncertain.

Framework diagram: parametric aircraft definition, DOE sampling and a physics-based sizing loop (431 ms per point) train Gaussian process surrogates (51 µs per point); an NSGA-II optimizer queries them with uncertainty-gated constraints, and Pareto-guided infill sends new physics runs only where the front is uncertain.
One loop, two speeds: the design iterates at surrogate speed, and physics runs only where the twin is uncertain.

Üç bulgu

  1. Judge a surrogate where the optimizer uses it. Active learning gives no global accuracy gain over an equal-budget static sample (0.50 % vs 0.48 %), but at the Pareto front its error is 29 times lower (0.03 % vs 0.84 %).
  2. −10.5 % block energy, fuel and CO₂. Relative to an ATR 72-500 reference that the model reproduces to within 1 %. The gain comes from an aspect ratio of 15.9 and a cruise L/D rising from 14.9 to 16.9.
  3. Battery specific energy is the gate. At 350 Wh/kg (pack level) the energy optimum is a conventional turboprop. Hybridisation starts to pay above 560 Wh/kg and saves 13.7 % at 1500 Wh/kg.

Measured cost, 70-seat hybrid-electric regional aircraft

Physics modelThis framework
One design evaluation431 ms51 µs (×8 400)
Full optimization cycle, all training charged1.16 h5.0 min (×14)
Re-run after a changed requirement1.16 h44 s
Mean error, 120 independent designs—0.50 %
Error at the Pareto front0.84 % (static)0.03 % (×29)

The physics layer is mid-fidelity, so these speed-up ratios apply to it only. With RANS/FEM the per-point ratio grows, while the number of samples one can afford shrinks. The framework itself is fidelity-agnostic.

Tam metin: Engineering Proceedings (MDPI), EASN 2026 volume · yayına hazırlanıyor

Aynı kongredeki kardeş bildiri: Experimental Validation of an Aircraft Digital Twin: From Concept to Validation (Ayşe Nur Dişlitaş, Melih Yıldız)

Bu kongredeki tüm bildiriler →