Frequency response analysis of guard-heated hot-film wall shear stress sensors for turbulent flows

• Wall shear stress fluctuations in turbulent flow are difficult to measure accurately.
• New design of thermal sensor with guard-heating proposed to remove large errors.
• Analysis: large dynamic errors of standard thermal sensors in air can be eliminated.
• Accurate thermal anemometry, with its many advantages, enabled by this new design.

Guard-heated thermal sensors were recently proposed for the measurement of wall shear stress (or “skin friction”) fluctuations in turbulent flow, to overcome the severe errors due to substrate heat conduction encountered in conventional single-element (SE) hot-film sensors. An earlier computational study of steady-state performance showed that a sensor with guard-heating in two-planes (GH2P) can eliminate errors due to spatial averaging and axial heat conduction in the fluid, both of which limit the spatial resolution of conventional SE sensors. Here we present analytical and numerical results comparing the dynamic behavior – frequency response and phase lag – of the guard-heated and conventional designs.For the water–glass fluid-substrate combination, sensor amplitude and phase errors begin only at a frequency (fcfc) near the onset of attenuation due to boundary layer thermal inertia. In this case, although the SE sensor suffers spatial averaging errors, it shows low amplitude attenuation and phase lag, close to that of the GH2P sensors, up to fcfc.For air-glass, analysis suggests and numerical results confirm, that the response of the conventional SE sensor is dominated by unwanted substrate heat transfer, with rapid signal attenuation beginning at frequencies that are five orders of magnitude smaller than fcfc. In this case, guard-heating enables strong improvement in the dynamic response, with a small drop in the amplitude response ratio from 0.95 to 0.85 (compared to 0.95 to 0.06 for the SE sensor) and negligible phase lag errors over an additional five decades of frequency. For the guard-heated design, upstream pre-heating occurs, but does not use heat drawn from the sensing element. Numerical results show that signal phase lag is zero and amplitude deviations are small, with modest variation over four decades of frequency. Guard-heated (GH2P) sensors appear to be an attractive option for wall shear stress fluctuation measurement in turbulent flows.