Surrogate-based parameter exploration for the design of MEMS calorimetric wall shear stress sensors by means of CFD

Abstract

A family of MEMS calorimetric wall shear stress sensors is experimentally and numerically investigated to determine their most significant design parameters. Fifteen sensor prototypes are first calibrated in a range of ± 2 Pa to generate an experimental database for validation of the subsequent numerical investigation. Then, a fully parameterized numerical setup is used to investigate the effect of three geometric design parameters, namely the cavity height, the cavity width, and the inter-beam distance, on amplitude and sensitivity of the sensor. This is done by building a surrogate model based on Gaussian Process interpolation (Kriging) in the four-dimensional space consisting of the three design parameters and the shear velocity in the flow. Thanks to this methodology, the calibration curves of all possible sensor designs in the investigated range can be estimated with an error of less than 2 %. A detailed study of this model reveals that the most significant design parameters are the inter-beam distance and the cavity width, while the cavity height is found to be of minor importance. • CFD Simulation of MEMS calorimetric wall shear stress sensor. • Static calibration of a family of 15 sensors in the range ±2 Pa. • Surrogate-based parameter space exploration of three design parameter. • Interpolation of sensor’s signal and sensitivity for all sensors in defined limits.