Imprecise P-Box Sensitivity Analysis of an Aero-Engine Combustor Performance Simulation Model Considering Correlated Variables

Uncertainties are widely present in the design and simulation of aero-engine combustion systems. Common non-probabilistic convex models only capable processing independent or correlated uncertainty variables, while conventional precise probabilistic sensitivity analysis based on ideal conditions also fails due to presence uncertainties. Given above-described problem, an imprecise p-box method is proposed this study accordance with a multi-dimensional parallelepiped model, comprising variables unified framework effectively address complex hybrid problems where two co-exist. The concepts correlation angle coefficient any parameters defined. A model built as domain marginal intervals characteristics all parameters. initial parameter space converted into affine by introducing coordinate system. Significant minor filtered out through using pinching methods characterization. feasibility accuracy verified numerical example outlet temperature distribution factor. As indicated results, coupling between can be significantly characterized notable difference exists ranking compared considering independence which input (e.g., inlet pressure, density, reference flow rate) highly sensitive changes Furthermore, structural flame cylinder exert secondary effect.