TL;DR
This paper introduces a novel Finite Volume Method (FVM) based simulation approach for reinforced reactive injection molding (RRIM) of thermosets, addressing a gap where existing software often focuses on thermoplastics. The FVM solver, implemented in OpenFOAM, accurately captures complex multiphase flow, including partial wall contact and air behavior, and is validated against experimental data and commercial Finite Element Method (FEM) software.
What Problem Does This Paper Address?
The production of high-performance composite parts relies heavily on reinforced reactive injection molding (RRIM). Accurately simulating this process is crucial for predicting part properties and optimizing manufacturing. However, current commercial simulation software often simplifies the complex flow behavior of thermosets, which differ significantly from thermoplastics. Specifically, these tools tend to neglect the presence of air as a distinct phase and struggle to model phenomena like partial wall contact, which are critical for thermoset RRIM. This leads to a gap in reliable virtual process design for this important manufacturing technique.
Methods
The core of this work is the development of a new solver within the open-source Computational Fluid Dynamics (CFD) toolbox, OpenFOAM. This solver utilizes the Finite Volume Method (FVM) to model compressible, non-isothermal multiphase flows, treating air and polymer as distinct phases. Key methodological contributions include:
- Multiphase Flow Modeling: The solver employs the Volume-of-Fluid (VoF) method to track the interface between air and polymer.
- Phase-Dependent Boundary Conditions: A novel boundary condition for the outlet is implemented, allowing air to exit the mold while retaining the polymer, crucial for accurate mold filling simulation.
- Eulerian Approach: Unlike FEM, FVM calculates flow at cell surfaces using an Eulerian approach, which is argued to provide a more physically significant representation of fluxes and better handle partial wall contact.
- Integrated Material Models: The solver incorporates established models for:
- Curing Kinetics: The Kamal-Malkin model is used to describe the cross-linking process.
- Viscosity: The Castro-Macosko viscosity model captures the temperature, shear rate, and cure-dependent viscosity of thermosets.
- Fiber Orientation: The Folgar-Tucker model is implemented to predict the orientation of reinforcing fibers.
Datasets
The paper uses experimental data collected from injection molding trials of a glass fiber reinforced phenolic resin. This data includes pressure measurements at specific sensor locations within the mold. The simulations are also compared against results from a commercial FEM software (Autodesk Moldflow Insight 2018.1).
Key Insights and Ideas
- FVM's Advantage in Multiphase Flow: The FVM approach, by calculating fluxes at cell faces, offers a more realistic representation of multiphase flow compared to FEM, particularly in capturing phenomena like partial wall contact.
- Realistic Air Handling: Explicitly modeling air as a separate phase with phase-dependent boundary conditions is shown to be critical for accurate mold filling simulation, a simplification often made in commercial FEM tools.
- Partial Wall Contact is Key: The FVM solver successfully predicts regions of partial wall contact at the flow front, a characteristic of thermoset RRIM that is not captured by FEM simulations, which tend to predict sharp interfaces.
- Integrated Modeling Power: The successful integration of curing kinetics, viscosity, and fiber orientation models within the FVM framework demonstrates its potential for comprehensive RRIM simulation.
Key Results
- The FVM-based solver accurately predicts pressure distribution during mold filling, showing good agreement with experimental data across various filling speeds and mold temperatures.
- The simulation results for pressure are generally closer to experimental measurements than those from the commercial FEM software, especially at sensor position 1.
- The FVM solver successfully captures the phenomenon of partial wall contact, which is absent in the FEM simulations.
- The implemented curing kinetics model shows good agreement between FVM and FEM simulations, validating the correct implementation of this aspect.
- Fiber orientation predictions from FVM show reasonable agreement with FEM, with some differences noted in edge regions and alignment in the z-direction, attributed to the different flow modeling approaches.
Limitations and Future Work
- Material Characterization: While the FVM simulation results are generally good, the authors note that further material characterization beyond manufacturer-provided data could potentially improve accuracy.
- Air Simulation: Although air is modeled as a separate phase, the paper doesn't delve deeply into the specific challenges or advanced modeling of air behavior beyond its role in filling.
- Future Work Suggestions:
- Diffusion and Weld Lines: Incorporating diffusion and transport models to analyze weld line formation.
- Curing Heat Energy Balance: Extending the energy balance to include the heat generated by curing to improve temperature and viscosity modeling for thicker parts.
- Anisotropic Viscosity and Viscoelasticity: Implementing anisotropic viscosity models and viscoelastic behavior to better model extensional flows and improve the simulation of partial wall contact.
- Advanced Fiber Orientation: Further refinement of fiber orientation models.
Summary updated