2019 

Alessandro Corbetta; Vlado Menkovski; Roberto Benzi; Federico Toschi Deep learning velocity signals allows to quantify turbulence intensity Journal Article arXiv, 2019. Abstract  BibTeX  Tags: condmat.statmech, cs.AI, cs.LG, physics.fludyn @article{046d24a1bab542e983a477781595c64f, title = {Deep learning velocity signals allows to quantify turbulence intensity}, author = {Alessandro Corbetta and Vlado Menkovski and Roberto Benzi and Federico Toschi}, year = {2019}, date = {20190101}, journal = {arXiv}, publisher = {Cornell University Library}, abstract = {Turbulence, the ubiquitous and chaotic state of fluid motions, is characterized by strong and statistically nontrivial fluctuations of the velocity field, over a wide range of length and timescales, and it can be quantitatively described only in terms of statistical averages. Strong nonstationarities hinder the possibility to achieve statistical convergence, making it impossible to define the turbulence intensity and, in particular, its basic dimensionless estimator, the Reynolds number. Here we show that by employing Deep Neural Networks (DNN) we can accurately estimate the Reynolds number within $15%$ accuracy, from a statistical sample as small as two largescale eddyturnover times. In contrast, physicsbased statistical estimators are limited by the rate of convergence of the central limit theorem, and provide, for the same statistical sample, an error at least $100$ times larger. Our findings open up new perspectives in the possibility to quantitatively define and, therefore, study highly nonstationary turbulent flows as ordinarily found in nature as well as in industrial processes.}, keywords = {condmat.statmech, cs.AI, cs.LG, physics.fludyn}, pubstate = {published}, tppubtype = {article} } Turbulence, the ubiquitous and chaotic state of fluid motions, is characterized by strong and statistically nontrivial fluctuations of the velocity field, over a wide range of length and timescales, and it can be quantitatively described only in terms of statistical averages. Strong nonstationarities hinder the possibility to achieve statistical convergence, making it impossible to define the turbulence intensity and, in particular, its basic dimensionless estimator, the Reynolds number. Here we show that by employing Deep Neural Networks (DNN) we can accurately estimate the Reynolds number within $15%$ accuracy, from a statistical sample as small as two largescale eddyturnover times. In contrast, physicsbased statistical estimators are limited by the rate of convergence of the central limit theorem, and provide, for the same statistical sample, an error at least $100$ times larger. Our findings open up new perspectives in the possibility to quantitatively define and, therefore, study highly nonstationary turbulent flows as ordinarily found in nature as well as in industrial processes.  
Xiao Xue; Luca Biferale; Mauro Sbragaglia; Federico Toschi Particle settling in a fluctuating multicomponent fluid under confinement Journal Article arXiv, 2019, (12pages, 7 figures). Abstract  BibTeX  Tags: condmat.soft, physics.compph, physics.fludyn @article{d03238a0027f4132ad45a1d7327e35d7, title = {Particle settling in a fluctuating multicomponent fluid under confinement}, author = {Xiao Xue and Luca Biferale and Mauro Sbragaglia and Federico Toschi}, year = {2019}, date = {20190101}, journal = {arXiv}, publisher = {Cornell University Library}, abstract = {We study the motion of a spherical particle driven by a constant volume force in a confined channel with a fixed square crosssection. The channel is filled with a mixture of two liquids under the effect of thermal fluctuations. We use the lattice Boltzmann method to simulate a fluctuating multicomponent fluid in the mixedphase, and particlefluid interactions are tuned to reproduce different wetting properties at the particle surface. The numerical setup is first validated in the absence of thermal fluctuations; to this aim, we quantitatively compute the drift velocity at changing the particle radius and compare it with previous experimental and numerical data. In the presence of thermal fluctuations, we study the fluctuations in the particle's velocity at changing thermal energy, applied force, particle size, and particle wettability. The importance of fluctuations with respect to the mean drift velocity is quantitatively assessed, especially in comparison to unconfined situations. Results show that confinement strongly enhances the importance of velocity fluctuations, which can be one order of magnitude larger than what expected in unconfined domains. The observed findings underscore the versatility of the lattice Boltzmann simulations in concrete applications involving the motion of colloidal particles in a highly confined environment in the presence of thermal fluctuations.}, note = {12pages, 7 figures}, keywords = {condmat.soft, physics.compph, physics.fludyn}, pubstate = {published}, tppubtype = {article} } We study the motion of a spherical particle driven by a constant volume force in a confined channel with a fixed square crosssection. The channel is filled with a mixture of two liquids under the effect of thermal fluctuations. We use the lattice Boltzmann method to simulate a fluctuating multicomponent fluid in the mixedphase, and particlefluid interactions are tuned to reproduce different wetting properties at the particle surface. The numerical setup is first validated in the absence of thermal fluctuations; to this aim, we quantitatively compute the drift velocity at changing the particle radius and compare it with previous experimental and numerical data. In the presence of thermal fluctuations, we study the fluctuations in the particle's velocity at changing thermal energy, applied force, particle size, and particle wettability. The importance of fluctuations with respect to the mean drift velocity is quantitatively assessed, especially in comparison to unconfined situations. Results show that confinement strongly enhances the importance of velocity fluctuations, which can be one order of magnitude larger than what expected in unconfined domains. The observed findings underscore the versatility of the lattice Boltzmann simulations in concrete applications involving the motion of colloidal particles in a highly confined environment in the presence of thermal fluctuations.  
Roberto Benzi; Thibaut Divoux; Catherine Barentin; S é; Mauro Sbragaglia; Federico Toschi Unified theoretical and experimental view on transient shear banding Journal Article arXiv, 2019, (5 pages, 4 figures  supplemental 5 pages, 4 figures). Abstract  BibTeX  Tags: condmat.soft, condmat.statmech, physics.fludyn @article{67228bed5d894c6688f967ac3899f499, title = {Unified theoretical and experimental view on transient shear banding}, author = {Roberto Benzi and Thibaut Divoux and Catherine Barentin and S é and Mauro Sbragaglia and Federico Toschi}, year = {2019}, date = {20190101}, journal = {arXiv}, publisher = {Cornell University Library}, abstract = {Dense emulsions, colloidal gels, microgels, and foams all display a solidlike behavior at rest characterized by a yield stress, above which the material flows like a liquid. Such a fluidization transition often consists of longlasting transient flows that involve shearbanded velocity profiles. The characteristic time for full fluidization, $tau_textf$, has been reported to decay as a powerlaw of the shear rate $dot gamma$ and of the shear stress $sigma$ with respective exponents $alpha$ and $beta$. Strikingly, the ratio of these exponents was empirically observed to coincide with the exponent of the HerschelBulkley law that describes the steadystate flow behavior of these complex fluids. Here we introduce a continuum model based on the minimization of an outofequilibrium free energy that captures quantitatively all the salient features associated with such textittransient shearbanding. More generally, our results provide a unified theoretical framework for describing the yielding transition and the steadystate flow properties of yield stress fluids.}, note = {5 pages, 4 figures  supplemental 5 pages, 4 figures}, keywords = {condmat.soft, condmat.statmech, physics.fludyn}, pubstate = {published}, tppubtype = {article} } Dense emulsions, colloidal gels, microgels, and foams all display a solidlike behavior at rest characterized by a yield stress, above which the material flows like a liquid. Such a fluidization transition often consists of longlasting transient flows that involve shearbanded velocity profiles. The characteristic time for full fluidization, $tau_textf$, has been reported to decay as a powerlaw of the shear rate $dot gamma$ and of the shear stress $sigma$ with respective exponents $alpha$ and $beta$. Strikingly, the ratio of these exponents was empirically observed to coincide with the exponent of the HerschelBulkley law that describes the steadystate flow behavior of these complex fluids. Here we introduce a continuum model based on the minimization of an outofequilibrium free energy that captures quantitatively all the salient features associated with such textittransient shearbanding. More generally, our results provide a unified theoretical framework for describing the yielding transition and the steadystate flow properties of yield stress fluids.  
2016 

S Kramel; G A Voth; S Tympel; F Toschi Preferential rotation of chiral dipoles in isotropic turbulence Journal Article arXiv.org, ePrint Archive, Physics, 2016. Abstract  BibTeX  Tags: physics.fludyn @article{05679cb42abe4f178b839c71697fa9f8, title = {Preferential rotation of chiral dipoles in isotropic turbulence}, author = {S Kramel and G A Voth and S Tympel and F Toschi}, year = {2016}, date = {20160101}, journal = {arXiv.org, ePrint Archive, Physics}, abstract = {Particles in the shape of chiral dipoles show a preferential rotation in three dimensional homogeneous isotropic turbulence. A chiral dipole consists of a rod with two helices of opposite handedness, one at each end. We can use 3d printing to fabricate these particles with length in the inertial range and track their rotations in a turbulent flow between oscillating grids. High aspect ratio chiral dipoles will align with the extensional eigenvectors of the strain rate tensor and the helical ends will respond to the strain field by spinning around its long axis. The mean of the measured spinning rate is nonzero and reflects the average stretching the particles experience. We use Stokesian dynamics simulations of chiral dipoles in pure strain flow to quantify the dependence of spinning on particle shape. Based on the known response to pure strain, we build a model that gives the spinning rate of small chiral dipoles using Lagrangian velocity gradients from high resolution direct numerical simulations. The statistics of chiral dipole spinning determined with this model show surprisingly good agreement with the measured spinning of much larger chiral dipoles in the experiments.}, keywords = {physics.fludyn}, pubstate = {published}, tppubtype = {article} } Particles in the shape of chiral dipoles show a preferential rotation in three dimensional homogeneous isotropic turbulence. A chiral dipole consists of a rod with two helices of opposite handedness, one at each end. We can use 3d printing to fabricate these particles with length in the inertial range and track their rotations in a turbulent flow between oscillating grids. High aspect ratio chiral dipoles will align with the extensional eigenvectors of the strain rate tensor and the helical ends will respond to the strain field by spinning around its long axis. The mean of the measured spinning rate is nonzero and reflects the average stretching the particles experience. We use Stokesian dynamics simulations of chiral dipoles in pure strain flow to quantify the dependence of spinning on particle shape. Based on the known response to pure strain, we build a model that gives the spinning rate of small chiral dipoles using Lagrangian velocity gradients from high resolution direct numerical simulations. The statistics of chiral dipole spinning determined with this model show surprisingly good agreement with the measured spinning of much larger chiral dipoles in the experiments. 
publications
2019 

Deep learning velocity signals allows to quantify turbulence intensity Journal Article arXiv, 2019.  
Particle settling in a fluctuating multicomponent fluid under confinement Journal Article arXiv, 2019, (12pages, 7 figures).  
Unified theoretical and experimental view on transient shear banding Journal Article arXiv, 2019, (5 pages, 4 figures  supplemental 5 pages, 4 figures).  
2016 

Preferential rotation of chiral dipoles in isotropic turbulence Journal Article arXiv.org, ePrint Archive, Physics, 2016. 