Time-resolved fast-neutron radiography of air-water two-phase flows in a rectangular channel by an improved detection system

TL;DR

This study enhances fast-neutron radiography for air-water flows in rectangular channels by improving detector design and image processing, achieving higher quality images at lower exposure times, enabling better flow parameter analysis.

Contribution

The paper introduces a tailored detector and optimized post-processing techniques that significantly improve image quality and reduce exposure times in fast-neutron radiography of two-phase flows.

Findings

Achieved image quality improvements with a new detector design.

Reduced exposure times to as low as 3.33 ms.

Enabled more accurate prediction of flow parameters.

Abstract

In a previous work we have demonstrated the feasibility of high-frame-rate, fast-neutron radiography of generic air-water two-phase flows in a 1.5 cm thick, rectangular flow channel. The experiments have been carried out at the high-intensity, white-beam facility of the Physikalisch-Technische Bundesanstalt, Germany, using an multi-frame, time-resolved detector developed for fast neutron resonance radiography. The results were however not fully optimal and therefore we have decided to modify the detector and optimize it for the given application, which is described in the present work. Furthermore, we managed to improve the image post-processing methodology and the noise suppression. Using the tailored detector and the improved post-processing significant increase in the image quality and an order of magnitude lower exposure times, down to 3.33 ms, have been achieved with minimized motion artifacts. Similar to the previous study, different two-phase flow regimes such as bubbly slug and churn flows have been examined. The enhanced imaging quality enables an improved prediction of two-phase flow parameters like the instantaneous volumetric gas fraction, bubble size and bubble velocities. Instantaneous velocity fields around the gas enclosures can also be more robustly predicted using optical flow methods as previously.