The FOTOCAOS project has been divided into three target blocks:
The first block, MICROSCALE, comprising objectives 1 to 3, deals with the development of applications, based on metaheuristics and neural networks, for obtaining intrinsic kinetic parameters. Applications will be included in a graphical user interface, and a repository of kinetic parameters and reaction mechanisms will be developed.
- Resolution model of the matter conservation equation.
- Application of resolution of intrinsic kinetic parameters by means of metaheuristics.
- Application of resolution of intrinsic kinetic parameters by neural networks.
The second block, MESOSCALE, targets 4 to 7, brings together optimization and parallelization techniques of simulations. The standard parallelization in the Finite Volumes method consists of dividing the geometry and resolving each part independently, sending the values of the interfaces between processes. The optimization technologies of this block are not proposed as alternatives, but as complementary, between them and with the original.
- Multithreaded parallelization using shared memory in cross-field operations.
- Migration of equation resolution to GPU (Graphic Processing Unit).
- Optimization and horizontal parallelization in the radiation model.
The third block, MACROSCALE, objective 8 to 10, focuses on extending the usefulness of the simulations and exploiting the results of the previous block. The block ranges from higher computational cost simulations, such as the feasibility study of photocatalytic paintings on 3D city maps, to CFD simulations embedded in industrial process simulation platforms, to parametric simulation to automate design optimization.
- Parametric simulation application.
- Process-oriented simulation application
- High computational cost simulations.