Improvement of measurement accuracy using Bayesian inference - Reduction of instrumentation effort in an axial compressor

Abstract As the next generation of turbomachinery components becomes more sensitive to instrumentation intrusiveness, a reduction number measurement devices required for evaluation performance is possible and cost-effective way mitigate arising non-mastered experimental errors. A hybrid methodology that couples techniques with modeling through Bayesian data-driven framework employed reduce effort. numerical model provide an initial belief flow, which then updated based on undersampled observations by inference algorithm. The goal present work showcase proposed demonstrate its partial application Gaussian Process regression in reducing effort testing time at outlet low aspect ratio axial compressor stage representative last high-pressure compressor. Preliminary results show accurate reconstruction mean flow field direct uncertainty quantification provided approach.