Development and Validation of an MRI-Based Method for Particle Concentration Measurement

Abstract

Dilute, dispersed multiphase flows are of critical importance in engineered, environmental, and biological applications. Traditional laser-based techniques are limited to data acquisition in discrete 2D planes with optical access required. Magnetic Resonance Imaging (MRI) is finding increasing use as a diagnostic for 3D measurements in complex turbulent flows where no optical access is possible. A new diagnostic called Magnetic Resonance Particles (MRP) has been developed whereby the volume fraction of dispersed microparticles in water can be measured quantitatively. Data obtained for dilute loadings of bismuth, graphite, and titanium powder suspended in agar gel showed excellent agreement with theoretical predictions from the MRI literature. The spatial extent of signal disturbance was investigated using an individual stainless steel particle suspended in gel. Finally, a full-scale experiment was performed in which a streak of titanium particles was injected into a complex, turbulent channel flow. Calibration values were obtained using an integral flux method and the resulting particle concentration distribution was analyzed. The calibration values showed some disagreement with theory; this issue will be the focus of upcoming calibration experiments in a channel flow with homogeneous particle distribution. INTRODUCTION Dilute, dispersed multiphase flows are of interest in a wide range of applications including dust ingestion in aircraft engines, medicinal inhalants in the human airways, and sedimentary transport in coastal ecosystems. In commercial aircraft, the introduction of atmospheric dust can shorten engine service life significantly (Dunn et al., 1987), while volcanic ash can cause near-instantaneous engine failure (Dunn, 2012). Improved knowledge of, and prediction capabilities for particle transport characteristics in a turbine engine’s various subsystems would lead to improved efficiency in a major global industry. When inhaled by humans, particulate matter can be either beneficial (e.g., medicines) or harmful (e.g., airborne pollutants), but in either case it is useful to know where the particles are likely to flow and accumulate. A comprehensive review of particle transport in the human airways was provided by Kleinstreuer ∗Contact Author & Zhang (2010). Despite the wealth of applications in which detailed particle flow quantification would be beneficial, there are significant limitations on the collection of particle concentration data using currently available techniques. Such techniques are almost entirely based on optical methods in which data are acquired in distinct 2D planes through the use of a laser sheet and camera. This 2D acquisition limits the rate at which data can be obtained. Moreover, allowing undistorted optical access for the laser-camera pair imposes severe restrictions on the complexity of the geometry that can be studied. While 3D data sets can be amassed as a stack of 2D planes, the time, cost, and complexity of compiling such data sets is quite high. Given the scarcity of fully 3D data sets in the literature, it may be concluded that these concerns are rather prohibitive. Magnetic Resonance Imaging (MRI) is becoming an increasingly widespread tool for fundamental study of fluid mechanics and model validation in complex turbulent flows. Two diagnostics making use of a standard medical-grade MRI scanner are currently available. The first, known as Magnetic Resonance Velocimetry (MRV) allows 3D measurement of the 3-component velocity field (Elkins & Alley, 2007), while the second, known as MRC, is used to measure the 3D scalar concentration field in two-stream mixing studies (Benson et al., 2010). Both techniques are becoming increasingly popular because they are much more efficient than other 3D measurements—data sets comprising several million points can be acquired over a roughly 12 hour period—and because they permit the use of geometrically complex 3D-printed models. Recent studies highlighting these capabilities include flow in a coral colony (Chang et al., 2009), flow through a series of porous fins (Coletti et al., 2014b), and flow through a patient-specific human lung model (Banko et al., 2015). An early proof-of-concept study which included qualitative MRI measurements of a streak of glass particles in a turbulent pipe flow was performed by Coletti et al. (2014a). The objective of this paper is to present development and validation details for a new MRI-based method known as Magnetic Resonance Particles (MRP) that will enable researchers to quickly and quantitatively measure the 3D, time-averaged particle concentration distribution. The combination of 3D velocity data with the 3D particle concentration field obtained using MRP will provide a means to advance our funda-