Prediction of Process induced deformations in curved CFRP-metal hybrids
Topic(s) : Manufacturing
Co-authors :
Michael PAULITSCH (AUSTRIA), Liane GÜSSER (AUSTRALIA), Matthias THON (AUSTRIA), Franz MAIER (AUSTRIA), Roland HINTERHÖLZL (AUSTRIA)Abstract :
Process simulation of fiber reinforced plastics (FRP) and FRP-metal-hybrids is a valuable tool to predict process induced warpage and stresses in the manufactured part. Successfully validated, it enables a reduction of time-consuming and cost-intensive experiments.
Two different lay-ups of 2mm thick pre-consolidated laminates of unidirectional (UD) carbon fiber reinforced PA6 (SIGRAPREG®TP C U157-0/NF-T340/46%; CF-PA6) were formed into L-shapes or spherical caps– with or without an additional 2 mm sheet of either 1.4301 stainless steel or EN AW-1050A aluminum, to form a CFRP-metal hybrid. For the hybrid components, the adhesion promoter Vestamelt Hylink (Evonik) was deposited onto the sandblasted metal surfaces.
Manufactured L-shapes varied in inner radius (10mm, 17mm), CF-PA6 lay-up (010, 02/902/02/902/02) and composition (CF-PA6 or hybrid). 2mm thick CF-PA6 and 4mm thick hybrid spherical caps were produced with an inner radius of 123mm. The CF-PA6 laminates were heated to 280°C for approximately 90s in the IR-station of a thermoforming press, subsequently pressed into shape with 100kN, held for 30s on a 150°C warm steel tool and finally demolded.
Metal sheets for the hybrid components were preformed and heated for 5-7 minutes prior to the CF-PA6 forming step to activate the adhesion promoter (see Fig. 1).
The manufactured samples were immediately scanned with a laser scanner (RS6, Hexagon). To investigate the effect of moisture absorption, L-shapes were dried for 72 hours at 80°C and scanned a second time.
A cool-down and warpage simulation was set up in Abaqus, using the COMPRO plug-in (see Fig. 2a). The simulation utilizes recorded temperature data from the experiment. Material data was determined according to [1]. A schematic result can be seen in Fig. 2b.
Mesh convergence studies were carried out to determine the maximum element size for the overall specimen and as well as in the radius. The deformed geometry in the cooled-down state after processing was exported as an orphan mesh.
Geometries of the experiments and simulations were compared with two methods. Firstly, established metrics, i. e. spring-in angle (only L-shape) and radius (fitted cylinder or sphere) were determined [2-6]. Secondly, a surface comparison of the entire part, thus the 3D deviations of every point obtained in the scanning process, was performed.
The model can be successfully applied to all investigated configurations. The maximum surface deviation between simulation and experiment is at most in the range of about ± 0.6 mm. Higher local deviations were typically found at the corners, since some of the manufactured specimens exhibited torsion.
The samples showed no change in radius or angle of hybrid L-shapes before and after drying. The angle of the CF-PA6 L-shapes showed an increase in spring-in between 0.20° and 0.28°, which indicates little moisture absorption.
Overall, the established simulation showed good agreement with the experiments.
Two different lay-ups of 2mm thick pre-consolidated laminates of unidirectional (UD) carbon fiber reinforced PA6 (SIGRAPREG®TP C U157-0/NF-T340/46%; CF-PA6) were formed into L-shapes or spherical caps– with or without an additional 2 mm sheet of either 1.4301 stainless steel or EN AW-1050A aluminum, to form a CFRP-metal hybrid. For the hybrid components, the adhesion promoter Vestamelt Hylink (Evonik) was deposited onto the sandblasted metal surfaces.
Manufactured L-shapes varied in inner radius (10mm, 17mm), CF-PA6 lay-up (010, 02/902/02/902/02) and composition (CF-PA6 or hybrid). 2mm thick CF-PA6 and 4mm thick hybrid spherical caps were produced with an inner radius of 123mm. The CF-PA6 laminates were heated to 280°C for approximately 90s in the IR-station of a thermoforming press, subsequently pressed into shape with 100kN, held for 30s on a 150°C warm steel tool and finally demolded.
Metal sheets for the hybrid components were preformed and heated for 5-7 minutes prior to the CF-PA6 forming step to activate the adhesion promoter (see Fig. 1).
The manufactured samples were immediately scanned with a laser scanner (RS6, Hexagon). To investigate the effect of moisture absorption, L-shapes were dried for 72 hours at 80°C and scanned a second time.
A cool-down and warpage simulation was set up in Abaqus, using the COMPRO plug-in (see Fig. 2a). The simulation utilizes recorded temperature data from the experiment. Material data was determined according to [1]. A schematic result can be seen in Fig. 2b.
Mesh convergence studies were carried out to determine the maximum element size for the overall specimen and as well as in the radius. The deformed geometry in the cooled-down state after processing was exported as an orphan mesh.
Geometries of the experiments and simulations were compared with two methods. Firstly, established metrics, i. e. spring-in angle (only L-shape) and radius (fitted cylinder or sphere) were determined [2-6]. Secondly, a surface comparison of the entire part, thus the 3D deviations of every point obtained in the scanning process, was performed.
The model can be successfully applied to all investigated configurations. The maximum surface deviation between simulation and experiment is at most in the range of about ± 0.6 mm. Higher local deviations were typically found at the corners, since some of the manufactured specimens exhibited torsion.
The samples showed no change in radius or angle of hybrid L-shapes before and after drying. The angle of the CF-PA6 L-shapes showed an increase in spring-in between 0.20° and 0.28°, which indicates little moisture absorption.
Overall, the established simulation showed good agreement with the experiments.