Hybrid Optimization Software Suite (HOSS)

A different type of fluid-solid interaction solver is needed to simulate the effects of an explosively-generated shock wave propagating through air and impinging on building structures. For this purpose the FSIS (Fluid-Solid Interaction Solver) was developed and implemented into HOSS. Some of the main design parameters for the FSIS were: a) to be explicit in terms of time integration and maintain a time step similar to the solid phase time step; b) to allow for several fluid grids that could move with the solid; and c) to allow for the resolution of multi-phase flow problems.

Munjiza, A., Rougier, E., Lei, Z. & Knight, E.E. 2020. FSIS – A novel Fluid-Solid Interaction Solver for Fracturing and Fragmenting Solids. Comp. Part. Mech. doi.org/10.1007/s40571-020-00314-9

Cohesive Zone Models (CZMs) usually are used to simulate the fracture and fragmentation of solids. Existing CZMs have different disadvantages, such as artificial compliance and time-discontinuity. In order to avoid these drawbacks, the Unified Cohesive Zone Model (UCZM) was implemented into HOSS. In UCZM, damage surfaces are dynamically inserted into the model according to the local stress state, much like the extrinsic CZM approach. However, in order to avoid sudden jumps in the simulations, the state variables are smoothly transitioned from continua to discontinua through an algorithm that properly balances the nodal forces during the process, while maintaining highly efficient simulations.

Coupled Thermo-Hydro-Mechanical (THM) simulators are necessary to explore the physical processes involved in complex subsurface operations such as unconventional fossil energy production and underground nuclear test detection. In HOSS pre-existing discrete fractures in the rock media can be explicitly modeled and their influences are sensed by both the solid and fluid domains. In terms of the linking of the two codes, HOSS provides information such as porosity and permeability to the fluid solver, while the fluid solver passes fluid state variables such as pressure and temperature back to HOSS.

Jansen, G., Valley, B., Miller, S.A. THERMAID-A matlab package for thermo-hydraulic modeling and fracture stability analysis in fractured reservoirs. arXiv: 1806.10942.

Knight, E.E., Rougier, E., Lei, z., Euser, B., Chau, V., Boyce, S., Gao, K., Okubo, K., Froment, M. 2020. HOSS: An Implementation of the Combined Finite-Discrete Element Method. Comp. Part. Mech. doi.org/10.1007/s40571-020-00349-y

When using lower order 2D or 3D finite element formulations, such as constant strain triangles (CSTRI) and tetrahedrons (CSTET), an issue that often appears is that these types of elements will numerically lock. This is due to their full integration scheme and is often recognized by a checkerboard pattern in the stress distributions. To alleviate this deficiency composite triangular (COMPTRI) and tetrahedral (COMPTET) finite elements were developed. Both COMPTRI and COMPTET formulations use selective integration of stresses in order to avoid this artificial stiff response or locking. The developed finite elements utilize a unified hypo-/hyper-elastic approach that allows linkage to user-defined (isotropic or anisotropic) material models. A demonstration of HOSS’ generalized anisotropic deformation kinematics is shown below.

a) Model of a representative geologic structure (all dimensions in meters). Four different types of materials are included in the model. b-e) Comparison of the wave propagation through the geologic medium: anisotropic material (b, c) and isotropic material (d, e). Units for speed are meters per second.

Lei, Z., Rougier, E., Knight, E.E., Munjiza, A., & Viswanathan, H. 2016. A generalized anisotropic deformation formulation for geomaterials. Comp. Part. Mech 3, 215-228. doi.org/10.1007/s40571-015-0079-y

Lei, Z., Rougier, E., Knight, E.E., Frash, L., Carey, J.W., & Viswanathan, H. 2016. A nonlocking composite tetrahedron element for the combined finite discrete element method. Engineering Computations. 33(7), 1929-1956. doi.org/10.1108/EC-09-2015-0268

The problem with existing distributed potential contact force algorithms is that the potential field introduces artificial numerical non-smoothness in the contact force, i.e. the contact forces calculated experience a jump (in amplitude and/or direction) when the contact points move from one finite element to another finite element. To overcome this issue, a new solution, which is named the Smooth Contact Algorithm (SCA) was developed in HOSS.

In the SCA, a smooth potential field is introduced according to the global geometry information of each discrete element. In particularly, the SCA calculates contact potential at nodes of the finite element mesh by taking into account nodal connectivity and existing discrete element boundaries. The contact force is then calculated as a function of the gradient of the potential field. Thus, a smooth contact evolution for a smooth surface is recovered.

Lei, Z., Rougier, E., Bryan, E., Munjiza, A. A smooth contact algorithm for the combined finite discrete element method. Computational Particle Mechanics, 2020. doi.org/10.1007/s40571-020-00329-2

In 2013 LANL unveiled its Integrated Solid-Fluid (ISF) Interaction Solver (US Patent #US10275551B2) to accurately resolve the interaction between fluid and solid domains under an enhanced fluid pressure boundary condition. The HOSS-ISF accounts for fluid flow through fracturing porous solids, fluid flow through crack manifolds, pressure wave propagation through fluid and fluid-solid interaction and is applicable to hydraulic-fracture problems, enhanced geothermal systems, or carbon capture related systems. One of the main advantages of the HOSS-ISF is that the fluid phase is described using the same grid as the solid phase via a modified Eulerian formulation. This eliminates the need of continuously mapping variables between the fluid and solid domains. Most importantly, the HOSS ISF features an explicit time integration solver with an aperture-independent critical time step size.

Lei, Z., Rougier, E., Munjiza, A., Viswanathan, H., & Knight, E.E. Simulation of discrete cracks driven by nearly incompressible fluid via 2D combined finite‐discrete element method. International Journal for Numerical and Analytical Methods in Geomechanics, 43:1724-1743, 2019. doi.org/10.1002/nag.2929

Rougier, E., Knight, E.E., & Munjiza, A. Integrated solver for fluid driven fracture and fragmentation. US Patent US10275551B2, granted 30 April 2019.

Utilizing a generalized massively parallel solution based on a virtual parallel machine for FDEM, it has been demonstrated that large increases to the number of processors only results in marginal increases in the specific time. This solution is problem-specific (as opposed to computer architecture specific) thus, a parallel FDEM code can be adapted to different hardware platforms ranging from desktops to 10,000+ node high-performance-computing (HPC) clusters.

Lei, Z., Rougier, E., Knight, E.E., & Munjiza, A. 2014. A Framework for Grand Scale Parallelization of the Combined Finite Discrete Element Method in 2D. Comp. Part. Mech. 1, 307-319. doi.org/10.1007/s40571-014-0026-3