Gilad Arwatz

Dynamic calibration and modeling of a cold wire for temperature measurement

The dynamical behavior of cold wires and their supporting structure is investigated. It is shown that conventional cold wires have a much more limited frequency response than previously believed, which can cause substantial inaccuracies in temperature data. Here, a lumped parameter model that accounts for the effects of end conduction and wire response is developed. The model is verified by comparing to the experimental data, where the frequency response of the wire is investigated under different heating conditions, with convincing agreement. All the parameters in the model can be found from the geometry of the sensor and the properties of the materials used in its construction. The new model can either be used as a sensor design and optimization tool, in order to design better sensors, or it can be used to correct the data acquired with conventional cold wire sensors, with non-negligible end-conduction effects, so that accurate measurements can be obtained.

Development and characterization of a nano-scale temperature sensor (T-NSTAP) for turbulent temperature measurements

A new nano-scale temperature sensor (T-NSTAP) is presented. The T-NSTAP is a sub-miniature, free-standing, platinum wire suspended between silicon supports, designed for temperature measurements at high frequencies. The new sensor is designed to have a bandwidth far superior to that of conventional cold-wires, which in combination with its small size, minimizes the effect of temporal and spatial filtering on the data. Unfiltered data allows for unique investigations of the scalar turbulence, including the dissipation range. Temperature measurements were conducted with the T-NSTAP in a heated grid-turbulence setup, with constant mean temperature gradient, and compared to data acquired with a conventional cold-wire. It is shown that the cold-wire signal is significantly attenuated over a broad range of frequencies, including lower frequencies. The attenuation has a direct and substantial effect on the measured variances and the estimated scalar rate of dissipation. The observed differences between the conventional cold-wire and the T-NSTAP are compared to what is predicted by a cold-wire model with convincing agreement over the entire spectrum. A close to perfect agreement between the two sensors is shown when the cold-wire data is corrected for end-conduction effects using the cold-wire model. In addition, a frequency content analysis of the probability density function of the temperature and its derivative is performed with direct comparison between the cold-wire and the T-NSTAP.