Comparison Of Signal Processing Algorithms For Multi-Sensor Intrusive Phase-Detection Probes

Abstract

Measurements of two-phase flow properties in highly turbulent, self-aerated free-surface flows remain a challenging task. The combination of dual-tip intrusive phase-detection probes with the recently developed adaptive window cross-correlation (AWCC) signal processing allows for measurements of local pseudoinstantaneous 1-D velocities. However, the applied window averaging reduces the data rate and thereby limits the retrieval of time-resolving turbulence statistics. Nevertheless, this represents the best practice in the field of hydraulic engineering. Meanwhile, intrusive phase-detection probes with more than two sensors are wellestablished in nuclear and chemical engineering. The use of four or more probe tips in combination with suitable signal processing algorithms enables the measurement of local instantaneous 3-D particle velocity vectors and particle diameters. The applied signal processing algorithms are typically based on eventdetection and thus yield an increased data rate in comparison to the AWCC approach. In this work, we compare two signal processing algorithms for multi-sensor intrusive phase-detection probes and evaluate their applicability for measurements of two-phase flow properties in highly turbulent free-surface flows. A dual-tip intrusive phase-detection probe in combination with the AWCC technique serves as a benchmark. To this end, we use synthetically generated particle velocity time series, which are obtained by modeling bubble trajectories based on synthetic stochastic velocity fields. This approach is advantageous over experimentally obtained signals in that the true instantaneous velocities and particle shapes are known, allowing for a direct comparison of the signal processing algorithms. Furthermore, we discuss the limitations of the considered signal processing algorithms and highlight future research opportunities.