An examination of hot-wire length corrections

Abstract

A comprehensive experimental examination is presented of the effect of hot-wire length for a single-wire probe. Specially constructed hot wires, with 2.5 μm and 5.0 μm nominal diameters and lengths from 0.3 mm to 2.8 mm, were used to measure the one-dimensional spectrum of the streamwise fluctuating velocity component u as well as the skewness and flatness of both u and du/dt. All measurements were made in air at Y+=150 (corresponding to Y/δ=0.1) in a boundary layer growing inside a pipe. The Reynolds number was 37 000 based on the velocity at the centerline and on the boundary-layer thickness. The longest wires exhibited some attenuation at the highest wavenumbers. This attenuation was found to roughly confirm the correction to the spectrum due to Wyngaard. A correction for the microscale λ, and thus dissipation, was examined and found to be correct to a fair degree (within 8%). This correction is based on the early work of Skramstad and of Frankiel, but with the integral scale of R11 (0, 0, r) assumed to equal the λ scale. The measurements of skewness and flatness did not indicate an experimentally discernible effect of the length of the hot wire.