A Data-Driven Probabilistic Learning approach for the Prediction of Controllable Pitch Propellers Performance

Themulti-fidelity machine learning framework proposed in this paper leverages a probabilistic approach based on Gaussian Process modeling for the formulation of stochastic response surfaces capable of describing propeller performance for different mission profiles. The proposed multi-fidelity techniques will help coping with the scarcity of high-fidelity measurements by using lower-fidelity numerical predictions. The existing correlation of the multi-fidelity datasets is used to infer high-fidelity measurements from lower fidelity numerical predictions. TheprobabilisticformulationsembeddedinGaussianProcessregressionsgivestheuniqueop-portunitytolearnthetargetfunctionsdescribingpropellerperformanceatdifferentoperatingconditions,whilequantifyingtheuncertaintyassociatedtothatspecificprediction.Whilethemulti-fidelityautoregressiveschemeallowstoconstructhighaccurateresponsesurfacesusingonlyfewexperimentaldata,UncertaintyQuantification(UQ)providesanimportantmetrictoassesthequalityofthelearningprocess.Wedemonstratethecapabilityoftheproposedframe-worktopredicttheperformanceofacontrollablepitchpropellerusingfewexperimentaldatacomingfromtowingtankexperimentsandmanymedium-fidelitypredictionsobtainedusinganin-housedevelopedBEM,validatedandverifiedinmanypreviousstudies