Sensitivity analysis of the Cercignani-Lampis accommodation coefficients in prototype rarefied gas flow and heat transfer problems via the Monte Carlo method (Q6592676)

In rarefied gas dynamics, the Cercignani-Lampis (CL) scattering kernel, characterized by two accommodation coefficients (ACs) -- the tangential momentum and normal energy coefficients -- is widely used to describe gas-surface interactions, particularly in non-isothermal conditions where both momentum and energy exchange occur. This study presents a detailed sensitivity analysis of how variations in these ACs influence key output quantities in several prototype flow problems: cylindrical Poiseuille flow, thermal creep flow, thermomolecular pressure difference (TPD) flow, plane Couette flow, and heat transfer (Fourier flow).\N\NIn the analysis, random uncertainties are introduced to the ACs as input parameters, and their impact on the main output quantities is evaluated through Monte Carlo propagation. By comparing the output uncertainties across different flow configurations and rarefaction regimes, the study identifies conditions where the outputs are most sensitive to the ACs. These high-sensitivity scenarios are particularly useful for estimating ACs, as they allow for larger tolerances in modeling and experimental errors.\N\NKey findings include:\N\NIn Poiseuille and Couette flows, the output uncertainties (flow rate and shear stress, respectively) are significantly influenced by the tangential momentum AC but are less affected by the normal energy AC over a broad range of gas rarefaction. For thermal creep and TPD flows, the output uncertainty (flow rate) depends on both ACs, but it generally remains smaller than the input uncertainties. In contrast, Fourier flow exhibits a higher sensitivity, where the uncertainty in heat flux can be comparable to or even exceed the input uncertainties in both ACs across a wide rarefaction range. These results suggest that combining Poiseuille (or Couette) flow with Fourier flow provides a more effective approach for estimating the CL ACs compared to the conventional combination of Poiseuille and thermal creep flows. For instance, to estimate the normal energy AC with 10\% accuracy, experimental uncertainties should be below 4\% in thermal creep or TPD flows but can be as high as 10\% in Fourier flow. This insight offers a refined strategy for characterizing gas-surface interactions in rarefied gas dynamics.