MANTA: An industrial strength open source high performance explicit and implicit multi-physics solver
Please login to view abstract download link
MANTA (Mechanical Analysis Numerical Toolbox for Advanced applications) is an open-source effort from the CEA and EDF to develop a multiphysics solver for quasi-static and fast-transient simulations in fluids and solids. MANTA aims to replace the 40 years-old Cast3M and Europlexus solvers and provide larger physical modeling abilities using up to date technologies. The project has been designed to meet several objectives. First the quality assurance compatible with safety-critical studies in the nuclear industry, then compatibility with high performance computing, and finally ease of use for mechanical engineers or researchers. Indeed, MANTA targets two main kind of users: 1. The mechanical engineers or researchers which exploit the output of numerical simulations to design or analyse physical systems of interest. which would interact with a frontend offering some API (C++ and python) where most numerical details are hidden by default. 2. The researcher in the field of numerical methods which would like to implement and test various algorithms. MANTA's components provide a flexible ways to implement new unstructured-mesh-based numerical methods. Considering the goal of inheriting the large scope of both Cast3M and Europlexus, MANTA aims at enabling the numerical modeling of various physical phenomena such as the explicit and implicit treatement of solid mechanics and heat transfers, in the discretization context of the finite elements method. These bricks enable the treatement of elasto-visco-plastic solid behaviors as well as damage mechanics, or contact and friction. Usual heat transfer mechanisms as well as their coupling with solid mechanics are also available. MANTA also targets a large set of features in the field of compressible fluids dynamics with the finite volume method, and their interactions with deformable structures. Some functionalities for the simulation of multicomponent flows (e.g. steam/water) or pipe flows are already available. Some short term perspectives encompass AMR, parallel contact detection modeling, usual numerical methods in the context linear elastic fracture mechanics, and robust non-linear solvers.
