Bylines

Enabling Multiphysics Solutions: Coupled Approaches and Beyond
by Subham Sett, SIMULIA

Multiphysics refers to the simultaneous inclusion of multiple physics fields in order to accurately predict device behavior or system response. Instances of multiphysics in industry such as airplane wing flutter, underwater shock effects on naval structures, tire hydroplaning, and so on, are not only widespread but also diverse.

Developments in software technology and hardware capacity are now enabling the simulation of a comprehensive set of real-world problems taking into account the effects of a range of physics involved. We will now look at how the realistic simulation of fluid-structure interaction of FSI, a major multiphysics focus area, is being made possible by such advances in technology.

Fluid structure interaction (FSI) represents the class of multiphysics where we consider the effects of fluid flow on compliant structures and their subsequent interactions. The primary fields interacting across the fluid and the structure domain are pressures and displacements, respectively. For problems where thermal effects are significant, temperature is an additional field in both the domains. In addition to these primary fields, there are secondary fields such as piezo-electric effects in the structural domain or cavitation effects in the fluid domain that indirectly contribute to the fluid-structure interaction. 

In Abaqus, a co-simulation technique is used to solve complex FSI analyses by coupling Abaqus to external CFD based flow solvers. In this solution scheme, Abaqus and the external CFD solver run concurrently, solving structural and fluid equations independently while exchanging converged solution quantities from the interacting fields at the interface. This approach to FSI, often called the partitioned approach, offers the most general-purpose tool for solving real-world FSI problems.

Communication between Abaqus and a CFD solver is critical
It is likely that your Abaqus and CFD solutions are running on different hardware configurations and even on remote systems.  Hence, essential to the coupled approach to FSI is seamless communication between Abaqus and the CFD solver. This communication is facilitated by using either an independent coupling, or a direct coupling interface, both of which are supported in the recently introduced SIMULIA Multiphysics Platform.

Independent or direct coupling?
The independent coupling offering (Figure 1) provides the flexibility of coupling Abaqus to a CFD code the user may already have.  This approach is useful when a company wants to encourage collaboration between their existing FEA and CFD engineering groups.  This coupling is through MpCCI, a code-coupling interface from the Fraunhofer SCAI; MpCCI allows Abaqus to couple to Star-CD from CD-adapco, and FLUENT from ANSYS.

Figure 1.  Independent coupling.

The Direct Coupling Interface or DCI (Figure 2), on the other hand, provides users the advantage of tighter integration between Abaqus and select third-party CFD solvers, without requiring additional software components. This approach is useful when an existing FEA engineering group needs to solve certain kinds of FSI problems on its own.  The first version of this direct interface, made available with the launch of Abaqus Version 6.7, provides support for AcuSolve from ACUSIM Software.

Figure 2.  Direct coupling.

While the current focus is on FSI, it is worth noting that either approach is scalable to included multiphysics with other external physics solvers as well.

Enabling the FSI analysis
An easy-to-use FSI module for Abaqus/CAE has been created to provide a unique interactive approach to manage the entire FSI workflow in the Abaqus/CAE environment. This workflow is standardized regardless of the choice of CFD code for the coupled analysis.  It includes defining the CFD model, defining the interactions and properties, and finally running the analysis and postprocessing the solution.  The FSI module, developed using Abaqus/CAE customization tools, is available as an additional plug-in.

We will now demonstrate an FSI solution with a peristaltic pump (Figure 3) as an example. Peristaltic pumps are used in a wide array of industries requiring the pumping of clean/sterile fluids (biomedical devices) or very corrosive fluids (chemical process industries).

Figure 3: A peristaltic pump works on the principle of peristalsis, where a rotor compresses a flexible hose that contains the fluid to be pumped.  The compression and restitution (or relaxation) of the flexible hose causes the fluid to move.

Visualizing the problem
Here, we start with a single Abaqus/CAE database with two models: one for the flexible hose and the relevant pump structural parts, and one for the corresponding fluid region (Figure 4). The structural model is fully defined and setup to run as an Abaqus/Explicit analysis. For the fluid model, the mesh has been created. Basic CFD settings for this model will be provided in the FSI module.

Figure 4: Abaqus/CAE FSI Module main window, displaying peristaltic pump assembly with internal fluid representation.

Defining the FSI model
The FSI model is defined in what is called the Study. This step allows the user to select the CFD code, identify the coupling step in the structural model and define the basic CFD model. As part of the CFD model definition the user specifies the boundary conditions and material type as shown in Figure 5. The fluid model data is then written in a format supported by the selected CFD code.  The applicable CFD preprocessor will then be launched and the user can complete the rest of the CFD model definitions.

Figure 5. FSI Study settings for peristaltic pump analysis using direct coupling with AcuSolve (LEFT) as the CFD code, and independent coupling (RIGHT) with STAR-CD via MpCCI.

Defining interactions and properties
The next step is to associate the structural model and the fluid model that are part of the FSI Study. The “Create Interaction” dialogue box (Figure 6) allows the user to choose from the list of available regions in the structural model (the hose exterior, interior, etc.), and in the fluid model (blood flow in, flow out, wall boundary etc.) that can be coupled.   
With the FSI interface selected as shown, you can now use the “Create Interaction” Property to select the solution quantities. With the interactions defined, we can proceed to solving the FSI problem.

Figure 6. Interaction settings for the peristaltic pump. The fluid structure interface is defined by selecting the Abaqus surface named HOSE-INTand the AcuSolve wall boundary named WALL. The solution quantities being exchanged are the Concentrated forces (Import) and Nodal displacements (Export).

Solving the FSI problem
This step involves defining coupling specifics such as time steps (Figure 7), providing Abaqus specific settings and CFD specific settings. For the independent coupling using MpCCI, additional settings for MpCCI can be specified.  The analysis can now be submitted for execution.

Figure 7. Job settings: Here, the direct coupling interface is used to couple Abaqus with AcuSolve using a fixed coupling time step of 0.01s.

Finally, the solution results can be post-processed in the Visualization module of Abaqus/CAE (Figure 8).  The fluid data can be extracted in the Abaqus output database format and be viewed in conjunction with the structural results. The coupled results illustrate the pumping action of the pump.

Figure 8.  Postprocessing the FSI results with the Visualization module. Here, the acuOdb tool provided by ACUSIM is used to extract the AcuSolve data. 

Beyond today’s coupled solutions
The partitioned approach enabled by the SIMULIA Multiphysics Platform, as exemplified by the FSI solution above, is an excellent tool for solving a wide range of multiphysics problems spanning many industries.  However, this coupled approach has its limitations: numerical challenges can occur, especially when handling interface conditions in the structural domain that include extreme contact, severe deformation, and damage or failure.      

An example of such a problem would be the sloshing of liquid inside a tanker that is hit by an outside force (Figure 9).  Tankers carrying liquids need to withstand impact loads during transportation.  The liquid itself is responsible for a significant portion of the container loading and any severe deformation of the tank can lead to rupture and potential spillage.  The event is highly dynamic in both the fluid and the container and there is a need to study the progressive damage and failure of the interface material as well.

A Coupled Eulerian-Lagrangian (CEL) method is being developed in Abaqus/Explicit to handle such FSI problems. The capability uses multi-material finite element formulation to handle the structure and simple fluids behavior in a single framework, thus alleviating requirements on continuity in fluid mesh topology which is a requirement for the coupled approach.  The CEL approach will be suitable for solving many interesting FSI problems in industry including tire hydroplaning, automotive air bag inflation, and liquid product dispensing.

The tools for solving FSI multiphysics problems are increasing in sophistication, flexibility and usability, clearly enhancing the dialogue between structural and fluid engineers.  Deciding which tool to apply to a specific problem still depends on understanding how far an analysis needs to be taken to achieve the desired results – in other words, traditional engineering insight.