Sample menu:

Research activites

My research activity at Montpellier is now mainly focused on computational cardiovascular biomechanics; the developments made within the group I am leading are gathered in the YALES2BIO solver for the description of blood flows at both micro and macro scales. I am also still working in close connection with CERFACS , addressing issues related to combustion noise, thermo-acoustic instabilities and wall modeling. Some examples of recent realizations I was involved in are given below. This is not an exhaustive list; refer to my publications and/or PhD students to obtain a more complete view of my research activities.

Subgrid scale modeling for LES

The WALE model (Nicoud and Ducros, Flow Turb. and Comb., 62(3), 1999) is now recognized as a simple yet efficient way to represent the effect of subgrid scales in wall bounded turbulent flows in complex geometries. However, a careful examination of the underlying spatial operator demonstrates that this model overestimates the eddy-viscosity and thus subgrid scale dissipation in some academical flows (e.g.: 2D vortex). This motivated the design of the so-called Sigma model (Nicoud et al., Phys. Fluids, 23(8), 2011) which has the unique property to vanish as soon as the resolved field is either two-dimensional or two-component, including the pure shear and solid rotation cases. As the WALE model, it also has the appropriate cubic behavior in the vicinity of solid boundaries without requiring any ad-hoc treatment. This model proved more accurate that the dynamic Smagorinsky model in several cases, including the pulsatile hot jet experiment developed at IFPEN during Hubert Baya Toda's thesis.

Animation of an iso-surface of temperature to visualize the generation of the jet and associated vortex ring as well as the impingement of the hot jet with the wall.
(Baya Toda et al., Phys. Fluids, 26, 2014)

Entropy-acoustic coupling in a realistic aeroengine

Thermoacoustic instabilities are frequently encountered during the development of low pollutant emissions combustors. While Large-Eddy Simulation is now accepted as a very powerful tool to numerically reproduce these instabilities, it also proves useful to use low order models to better understand the nature of these instabilities. The AVSP solver (Nicoud et al., AIAA J., 45(2), 2007) solves the Helmholtz equation for the acoustic pressure over a non-isothermal flow in complex 3D domains bounded by arbitrary boundary impedances. Fed with appropriate data for the flame response to acoustic excitations, this approach proved efficient to reproduce the stability map of several academic (e.g.: Silva et al., Comb. Flame, 160(9), 2013) and industrial (e.g.: Selle et al., Comb. and Flame, 145, 2006) combustors, including annular chambers with multiple burners (Sensiau et al. Int. J. Aeroacoustics, 8(1), 2009) without or with multiperforated liners (Gullaud and Nicoud, AIAA J., 50(12), 2012). The Helmholtz equation relies on the so-called zero Mach number assumption stating that the mean velocity is very small compared to the speed of sound. A recent study suggests that the domain of validity of the zero mean flow assumption might be rather small (Nicoud and Wieczorek, Int. J. Spray Comb. Dyn., 1(1), 2009). One reason for this is that this equation does not support entropy waves, preventing the acoustic generation due to the entropy spots being accelerated in the dowstream nozzle to be accounted for. A Delayed Entropy Coupled Boundary Condition was recently introduced to correct this drawback of the Helmholtz approach (Motheau et al., J. Sound and Vibration, 2013) and successfully applied to an actual SAFRAN aeroengine (Motheau et al., J. Fluid Mechanics, 749, 2014).

Animation of an iso-surface of heat release to visualize the turbulent flame. Top: the flow is turbulent but stable. Bottom: The flow is pulsating due to a low frequency instability. The hot fluid packets shed from the primary zone generate a low pressure wave which in turn pulsates the flame. (from E. Motheau, CERFACS)

Cardiac turbulence

The hemodynamics within a human left heart was computed thanks to the YALES2BIO solver used together with the Medical Imaging-CFD framework developed within the OCFIA research program and successfully applied to large vessels (Midulla et al., Eur. Radiol, 22(10), 2012). Ten ECG gated CT images were acquired along the cardiac cycle and used to generate a 3D finite volume mesh evolving consistently with the observed geometry changes. The Arbitrary Lagrangian Eulerian methodology was then used to solve the Navier-Stokes equations within the computational domain which contains the four pulmonary veins, the left atrium and ventricle and the ascending aorta. The mitral and aortic valves were accounted for thanks to a simplified Immersed Boundary Method. A multi-cycle Large-Eddy Simulation allowed to clearly establish the turbulent nature of the intra cardiac flow during certain phases of the cycle (Chnafa, Mendez and Nicoud, Comp. Fluids, 94, 2014). A more recent application of the same methodology can be found in Chnafa, Mendez and Nicoud, ABME, 2016.

Animation of the vorticity magnitude in a realistic human left heart.
(from C. Chnafa, IMAG)

Blood flows

The YALES2BIO solver is also used in order to describe the dynamics of red blood cells in microfluidics. An immersed boundary method is used to model the fluid-structure interaction between the membrane and the inner/outer fluids (Mendez, Gibaud and Nicoud, JCP, 2014). The methodology was validated in several cases including the damped oscillations of a capsule (Martins Afonso, Mendez and Nicoud, JFM, 2014), the stretching of a capsule in a linear shear flow or of a red blood cell in an optical tweezers experiment (Siguenza et al., JCP, 2016). The YALES2BIO simulations allow to retrieve the red blood cell dynamics observed experimentally by Lanotte et al., PNAS, 2016, including the trilobed shape at large shear rate. The solver is now being used within the RheoBlood project to study the complex rheological behavior of blood.

Deformation of a red blood cell stretched as in the optical experiment of Mills et al., 2004.
(from J. Siguenza, IMAG)