English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Complex optimal design and control processes often require repeatedevaluations of expensive objective functions and consist of large designspaces. Data-driven surrogates such as neural networks and Gaussian processesprovide an attractive alternative to simulations and are utilized frequently torepresent these objective functions in optimization. However, pure data-drivenmodels, due to a lack of adherence to basic physics laws and constraints, areoften poor at generalizing and extrapolating. This is particularly the case,when training occurs over sparse high-fidelity datasets. A class ofPhysics-infused machine learning (PIML) models integrate ML models withlow-fidelity partial physics models to improve generalization performance whileretaining computational efficiency. This paper presents two potentialapproaches for Physics infused modelling of aircraft aerodynamics whichincorporate Artificial Neural Networks with a low-fidelity Vortex LatticeMethod model with blown wing effects (BLOFI) to improve prediction performancewhile also keeping the computational cost tractable. This paper also developsan end-to-end auto differentiable open-source framework that enables efficienttraining of such hybrid models. These two PIML modelling approaches are thenused to predict the aerodynamic coefficients of a 6 rotor eVTOL aircraft givenits control parameters and flight conditions. The models are trained on asparse high-fidelity dataset generated using a CHARM model. The trained modelsare then compared against the vanilla low-fidelity model and a standard puredata-driven ANN. Our results show that one of the proposed architecturesoutperforms all the other models at a nominal increase in run time. Theseresults are promising and pave way for PIML frameworks which can generalizeover different aircraft and configurations thereby significantly reducing costsof design and control.

Physics-Infused Machine Learning Based Prediction of VTOL Aerodynamics with Sparse Datasets