This report concerns the ongoing collaboration on the numerical modeling of the shear rheology for capillary suspensions via the Fast Stokesian Dynamics algorithm. Up to and including December 2024, we have completed major work on the base simulation software: Python-JAX Fast Stokesian Dynamics (JSFD). This includes enabling shear, both oscillatory and steady, and the implementation of several validation checks for the physics under consideration – a publication on the software is under preparation and this will be submitted in the coming month. JSFD is already publicly available and being tested by collaborators external to the project (Prof. G. Petekidis, FORTH, Greece). It is presently capable of simulating the effect of low-Reynolds-number hydrodynamic interactions between monodisperse spheres that are subjected to external forces and various interparticle potentials. The latter can be readily modified to purpose.
In order to compare experimental rheology with our numerical modeling, we have to map experimentally obtained – using confocal microscopy – particle coordinates into our simulation. We have devised such a mapping and have had success in creating simulationappropriate initial conditions based on experimental input. We remove any overlaps that come from replacing polydisperse particle (10%, experiment) by monodisperse spheres using gradient descent. The quality of this mapping has been characterized and we are investigating whether we can improve upon it, for example, by adding information on the position of the capillary bridges in the sample as obtained using confocal.
Our first rheological measurements were shown during the IFPRI 2024 Annual General Meeting (AGM), although these were based on very preliminary analysis. We have since realized that there are some fundamental issues that need to be overcome. We are presently studying the literature to find out the most accurate way to compute the loss and storage modulus within the Stokesian Dynamics framework. There appears to be a lack of consensus on the topic, which needs to be resolved, as we will detail in our report. Our present approach is to reproduce the rheology of a dilute dumbbell fluid, where the connection between the beads made using a Hookean spring. Lastly, we are in the process of implementing changes to the interaction potential that models the capillary bridges, as per the suggestions received during the AGM. We aim to complete these activities by May 2025, which closes the collaboration project, also including publication of our findings.