Dr. Christopher Elkins

Senior Research Engineer

Mechanical Engineering

Stanford University



Three-Dimensional Velocity Measurements in Complex Physiologic and Turbulent Flows Using Magnetic Resonance Imaging


Most technological and natural flows are three-dimensional, transitional or turbulent, and highly complex making them nearly impossible to predict with computations or measure with the most advanced conventional laser-based techniques. Magnetic resonance velocimetry (MRV) presents a promising experimental alternative for use in these flows. MRV uses magnetic resonance imaging (MRI) techniques in conventional medical scanners, and it is a highly versatile non-invasive technique for measuring the full three-component mean velocity field for complex flows in and around opaque objects. While MRV was developed for measuring blood flow in patients, we are using it to measure velocities in a diverse collection of flows including the flow around coral colonies, the flow in the internal cooling passages of gas turbine blades, and the physiologic flows in diseased aortas, stent-grafts, and bypass grafts.


In this talk, the basics of the three-dimensional MRV technique will be presented, and its strengths and shortcomings will be illustrated with experimental results. Comparisons between MRV and particle image velocimetry (PIV) measurements will be presented for the physiologic flow in a rigid model of a human aorta and for the turbulent flow in a channel downstream of a grid and a backward facing step. Results for these flows demonstrate the ability of the MRV to accurately capture with high resolution the periodic flow phenomena in physiologic flows and the complex secondary flow structures in turbulent flows.


Future developments of MRV will be discussed in terms of its practical applications and its possible use for measuring turbulent velocity fluctuations. Because MRV delivers well resolved velocities in three-dimensional domains, it can be used for validating CFD results. Moreover, by combining MRV and rapid prototype manufacturing which is capable of quickly making complex flow models, it is possible to make measurements in a very short time (on the order of one week). This enables an iterative design process for highly complex flow passages using experimental data in much the same way one might use the results of CFD. Examples will be presented related to the internal cooling passages of gas turbine blades. While MRV in its original form provided mean velocities, we are adapting other MRI techniques to measure turbulent velocity fluctuations in three-dimensional flows. Promising preliminary results of this technique will be presented.