We review a set of velocity calibration methods for one popular configuration of a four-hot-wire probe with the aim of finding a method of estimating with 10% accuracy (based on full scale) the mean flow vector, the rms of the turbulent velocity component and the associated linearized anisotropy invariants. We performed full 3D velocity calibrations of a classic and a sub-miniature probe. With a calibrated King/Jørgensen model and with an interpolation scheme in the real calibrations, we simulated calibrations and (anisotropic) turbulence measurements. Analysis of the idealized and of the realistic simulation data showed that the lookup-table method is the only good tool for processing hot-wire measurements. Polynomial models give large systematic errors. The King/Jørgensen model can only be used for qualitative ends, but it forms a good basis for a simulation-based assessment of the quality of a measured calibration set. Our results indicate that the experimental results reported in the literature, and which were based on hot-wire measurements, may require a reassessment of the calibration method.
- turbulent channel flow
- hot-wire probes