Sensitivity studies on the multi-sensor conductivity probe measurement technique for two-phase flows

The objective of this study is to advance the local multi-sensor conductivity probe measurement technique through systematic investigation into several practical aspects of a conductivity probe measurement system. Firstly, signal “ghosting” among probe sensors is found to cause artificially high bubble velocity measurements and low interfacial area concentration (a_i) measurements that depend on sampling frequency and sensor impedance. A revised electrical circuit is suggested to eliminate this artificial variability. Secondly, the sensitivity of the probe measurements to sampling frequency is investigated in 13 two-phase flow conditions with superficial liquid and gas velocities ranging from 1.00–5.00 m/s and 0.17–2.0 m/s, respectively. With increasing gas flow rate, higher sampling frequencies, greater than 100 kHz in some cases, are required to adequately capture the bubble number frequency and a_i measurements. This trend is due to the increase in gas velocity and the transition to the slug flow regime. Thirdly, the sensitivity of the probe measurements to the measurement duration as well as the sample number is investigated for the same flow conditions. Measurements of both group-I (spherical/distorted) and group-II (cap/slug/churn-turbulent) bubbles are found to be relatively insensitive to both the measurement duration and the number of bubbles, as long as the measurements are made for a duration long enough to capture a collection of samples characteristic to a given two-phase flow system (or a statistical ensemble). Fourthly, investigation into the orientation of a double-sensor probe in the pipe indicates that the sensors should be oriented parallel to the pipe wall to ensure symmetric bubble velocity measurements. Lastly, Monte Carlo simulations are performed to study the effects of the axial (s) and lateral (d) probe sensor separation distances. In addition to previous criteria on the ratio of s to the bubble diameter, it is found that s/d should be greater than four to minimize errors in the measured bubble velocity.