Optimisation of four-sensor probes for measuring bubble velocity components in bubbly air–water and oil–water flows

Abstract

In a previous paper Lucas and Mishra (2005) [3] a local four-sensor conductance probe was introduced to measure the velocity vectors of dispersed bubbles in bubbly two-phase flow in which the continuous phase is water. There are a very limited number of alternative methods available for bubble velocity vector measurement with which results from, for example, computational fluid dynamic models can be compared and so the four-sensor probe technique is of interest to the multiphase flow community. In the previous paper [3] a mathematical model was presented to calculate the velocity vector of each gas bubble from seven time intervals which were measured using the output signals from each of four ‘needle’ conductance sensors located within the probe. In the present paper, a new technique for making the local four-sensor probe is introduced to minimise interference with the measured bubbles. A new signal processing method is presented using criteria to ensure that (i) the group of sensor signals from which the bubble velocity vector is to be determined are all produced by the same bubble and (ii) bubbles which contact the local four-sensor probe in an ambiguous manner are ignored. The accuracy with which the locations of each of the rear sensors in the probe relative to the lead sensor can be measured influences the accuracy with which the bubble velocity vector can be measured. However, the degree to which the accuracy of the measured velocity vector is affected by errors in the measured probe dimensions is dependent upon the geometrical arrangement of the four sensors within the probe. Experimental results and an error analysis are presented which show that the susceptibility of the velocity vector measurement technique to errors in the measured probe dimensions is reduced if the geometrical arrangement of the four sensors is optimised. As a result of this initial work, an optimised probe, known as the P30 probe, was designed and built and results obtained from the P30 probe in swirling oil-in-water bubbly flow are presented. A probe calibration factor is defined in this paper which can be interpreted as a measure of the interference of a probe with the motion of the bubbles with which it interacts. For the probes described in this paper the calibration factor was found to be much closer to unity than for previous four-sensor probes described in the literature (e.g. [3]) suggesting that these new probes have a much smaller effect on the bubbles’ motion than previous probes.