Process simulation chain combined with a structural simulation to capture the actual process-induced effects and ‘as-built’ characteristics
Topic(s) : Material and Structural Behavior - Simulation & Testing
Co-authors :
Maximilian POLLAK (AUSTRIA), Franz MAIER (AUSTRIA), Roland HINTERHÖLZL (AUSTRIA)Abstract :
Introduction
To fully harness the potential of the CFRP material, it is crucial to find optimised, yet reliable designs. However, designing with CFRP poses challenges due to the inherent inhomogeneity, anisotropy, and process-induced effects and defects.
Iterative product and process optimization, with loops of prototyping and testing, results in expensive and time-consuming part development. Accounting for unknown inhomogeneities with generous safety factors further diminishes the lightweight potential.
Process simulations are common tools to overcome these challenges. By linking process simulations with structural simulations, we provide an 'as-built' simulation, accounting for process-induced effects and defects.
Material and Methods
An "as-built" simulation of a 3-point bending test for a C-section part, made from UD carbon fiber reinforced epoxy prepregs (HexPly® M79/34%/UD300/CHS), was set up in Abaqus 2019. To compare and validate the simulation results, parts were manufactured in a diaphragm forming process.
The geometric distortion, fiber orientations and residual stresses (discrete field quantities) were determined in a process simulation and mapped onto the part for the simulated bending test. The process simulation is composed of a macroscopic draping simulation, using quadrilateral shell elements with an empirical material model [1] and a curing simulation, using CHILE model from Compro plug-in with brick elements. The simulations were linked by mapping the results, allowing to consider the multi-physical processes with adequate material models and element types.
A mapping tool for the Abaqus programming interface (API) was developed in python and fortran for linking or transferring field quantities. Field quantities can be read from a simulation result and mapped to an FE-mesh with shell or solid elements. Field quantities are assigned by element number and material point numbering. This allows for a simple transfer, without interpolation of field quantities in the initial step by the Abaqus user subroutine ‘SIGINI’ and ‘ORIENT’.
Results
The "as-built" simulation was compared to an ‘as-designed’ version with ideal, fold free geometry, no residual stresses and perfect fiber placement.
The location of failure initiation in the presence of folds can be predicted with the ‘as-built’ simulation (Fig. 1). Force-displacement data from a bending test were used for validation of the simulation results (Fig. 2). The failure criteria from the simulation shows good agreement with the test data. The onset of the criteria corresponds with drops in the force response of the bending test. These drops could also be represented in the simulation with the linear damage model from Abaqus.
By mapping the field variables across the simulation steps, a strength loss of up to 50% was observed, which clearly shows the influence of process-induced residual stresses, distortion and fiber angle deviations, on the mechanical integrity.
To fully harness the potential of the CFRP material, it is crucial to find optimised, yet reliable designs. However, designing with CFRP poses challenges due to the inherent inhomogeneity, anisotropy, and process-induced effects and defects.
Iterative product and process optimization, with loops of prototyping and testing, results in expensive and time-consuming part development. Accounting for unknown inhomogeneities with generous safety factors further diminishes the lightweight potential.
Process simulations are common tools to overcome these challenges. By linking process simulations with structural simulations, we provide an 'as-built' simulation, accounting for process-induced effects and defects.
Material and Methods
An "as-built" simulation of a 3-point bending test for a C-section part, made from UD carbon fiber reinforced epoxy prepregs (HexPly® M79/34%/UD300/CHS), was set up in Abaqus 2019. To compare and validate the simulation results, parts were manufactured in a diaphragm forming process.
The geometric distortion, fiber orientations and residual stresses (discrete field quantities) were determined in a process simulation and mapped onto the part for the simulated bending test. The process simulation is composed of a macroscopic draping simulation, using quadrilateral shell elements with an empirical material model [1] and a curing simulation, using CHILE model from Compro plug-in with brick elements. The simulations were linked by mapping the results, allowing to consider the multi-physical processes with adequate material models and element types.
A mapping tool for the Abaqus programming interface (API) was developed in python and fortran for linking or transferring field quantities. Field quantities can be read from a simulation result and mapped to an FE-mesh with shell or solid elements. Field quantities are assigned by element number and material point numbering. This allows for a simple transfer, without interpolation of field quantities in the initial step by the Abaqus user subroutine ‘SIGINI’ and ‘ORIENT’.
Results
The "as-built" simulation was compared to an ‘as-designed’ version with ideal, fold free geometry, no residual stresses and perfect fiber placement.
The location of failure initiation in the presence of folds can be predicted with the ‘as-built’ simulation (Fig. 1). Force-displacement data from a bending test were used for validation of the simulation results (Fig. 2). The failure criteria from the simulation shows good agreement with the test data. The onset of the criteria corresponds with drops in the force response of the bending test. These drops could also be represented in the simulation with the linear damage model from Abaqus.
By mapping the field variables across the simulation steps, a strength loss of up to 50% was observed, which clearly shows the influence of process-induced residual stresses, distortion and fiber angle deviations, on the mechanical integrity.