RheoCube is a predictive simulations (or digital experimentation) tool, offering the R&D industry an easier way to understand what is happening at a molecular level in chemical formulations.
Below we list the latest additions and improvements to the RheoCube platform since the release of version 1.0:
Data Analysis and Visualization (DaVis) additions and improvements:
High Performance Computing
Access Data Analysis and Visualization (DaVis) sessions from within RC project
We’ve improved the user experience (UX) allowing for faster/direct access to Data Analysis and Visualization (DaVis). This improves the user’s workflow as the user no longer needs to go via the homepage to open a DaVis session to look at a project’s simulation results .
Video generation at fixed snapshot - Slicing
Previously it was only possible to pick one setting for video generation and generate it for an x amount of snapshots. Now users are able to pick a snapshot and then either slice every frame or rotate the view of the box for every frame.
The benefit of this new feature means that users will be able to create even more insightful videos. Some structures are very hard to visualize from a fixed camera angle due to the 3D nature of those structures. This improvement/extension to the video generation functionality gives users more insight into the structures
Davis: Derived attribute for difference between actual velocity and expected velocity (Non-newtonian velocity - meso only)
The non-Newtonian velocity is a blob attribute and is defined as the difference between the velocity of a blob in an experiment and the velocity that this blob is expected to have when it was only subject to shear. Thus, in a simple shear flow experiment, this attribute will be negligible. However, in experiments where the shear flow is interrupted, e.g. by particles, the non-Newtonian velocity shows how and where the flow is distorted.
Molecule analysis widget - Proper Radial Distribution Function for molecular simulations
The new Radial Distribution Function (RDF) can show the distance between parts of molecules in a MiMo box. By implementing this RDF feature, users can now quantify the ‘spread’ of molecules throughout the box. This can be useful when comparing structures between different experiments.
Better performance for molecular simulations
Two factors influence the improvement in performance for molecular simulations in this release. First, a more powerful compute resource is chosen, which costs approximately 4x more per hour. Molecular simulations can scale nearly linearly on the larger resource, meaning that the total cost per simulated time will stay roughly the same, but will run 4x faster. Second, we have removed the option to compute the heat conductivity, as this is very computationally expensive.
As a result, the "production" phase of the simulation is nearly 4x faster. We plan to reintroduce the heat conductivity as an independent simulation parameter that can be activated on a case-by-case basis. The net result is that with default simulation parameters (simulation split equally between equilibration and production) the full simulation can now be completed approximately 8x faster, at 2x the cost reduction.
Molecular SDK forcefield: a few bead changes
The aromatic beads BER, PEP, and XYR now have an extra carbon included in their SMARTS string definitions. This leads to better coarse graining outcomes.
Customers can now have a clear picture of what can be done with the models in RheoCube and how accurate the outcome of their simulations will be in comparison to what would be expected in the lab.
Validated SDK ionics in MiMo
Value : The majority of surfactants are ionic in nature, and have some explicit molecular detail. Capturing basic molecular structure interactions and ionic interactions in CGMD will enable users to examine such systems at the molecular scale.
Most of our formulation clients need basic molecular interactions in order to quantitatively capture microscale physics. This feature will have value for such clients.
High Performance Computing allows RheoCube to run bigger, and more complex, simulations for longer within reasonable result times. It is not automatically available, but rather on a case-per-case basis.
High Performance Computing (HPC): Parallel cluster ( Pcluster) with Automated choice of resources
This means that for 1.0 projects and larger systems the BE will automatically try to run 2-4-8 node jobs on AWS-pcluster.