An investigation into the use of Cross Correlation Velocimetry

This study analyses the frequency and spatial dependence of Cross Correlation Velocimetry (CCV) towards the measurement of fire induced flows. CCV uses temperature-time records from a pair of thermocouples, one downstream of the other, cross-correlated to determine the flow's velocity and is based in principle on the “frozen eddy” concept in turbulent flows. In between 1975 and 1980 Cox et al. (Cox) performed a series of experiments that showed that spatial and temporally resolved velocity measurements could be achieved by means of CCV. These types of velocity measurements are crucial in understanding ceiling jets, the role of sprinkler activation, and also in micro-gravity fire induced flows that conventional techniques cannot measure. However, the high cost associated with expensive analogue correlators available those days caused the CCV technique to gradually phase out after the advent of the bidirectional probe which was significantly cheaper and more robust in design. There have been vast improvements in data acquisition techniques, digital signal conditioning, filtration of random noise, as well post processing statistical packages which allow better and faster cross correlation of two random signals. This study is a first step towards applying these technological advantages to this outdated technique. The CCV probe’s accuracy is most sensitive to the thermocouple wire diameter, separation distance, and speed of data acquisition (sampling frequency). This study presents a parameter sensitivity analysis that includes the measurement of axial components of velocity in a heated turbulent jet with a velocity of 1.1 m/s with the sampling frequency, and probe separation distance adjusted independently (thermocouple wire diameter is kept constant). Nomenclature: A = Surface Area (m 2 ) Cp = Specific Heat (J/(kg °C)) d = Distance between Thermocouples (mm) f = Sampling Frequency (Hz) h = Average Convective Heat Transfer Coefficient (kW/m 2 ) Rxy = Cross correlation function of x(t) and y(t) (-) t = Time (s) T = Averaging Time (s) Tmax = Maximum temperature (°C) Tave = Average Temperature (°C) v = Velocity (m/s) V = Volume (m 3 ) x(t) = Temperature profile (°C) y(t) = Temperature profile (°C) ρ = Density (kg/m 3 ) θ = Non Dimensional Time (ND) τ = Time lag (s) τs = Spacing Lag (-) τR = Response Time (s) 1 Graduate Research Assistant (Goddard Fellow), Department of Fire Protection Engineering, WPI 2 Professor, Department of Fire Protection Engineering, WPI * Corresponding Author: [email protected] Proceedings of the 2008 Technical Meeting of the Central States Section of The Combustion Institute