Determination and verification of mechanical properties of GFRP filament wound pipes

Michał Krzysztoporski, Karolina Paczkowska, Zuzanna Pacholec, Wojciech Błażejewski

Pre-print pages 1-14

DOI:

keywords: filament winding, glass fiber reinforced polymers (GFRP), finite element (FEM) simulation, progressive failure analysis, materials testing, split disk test

abstract Filament winding is an efficient and versatile manufacturing technique utilised to create lightweight high-strength composite structures. Glass fiber reinforced polymers (GFRP) are widely used in filament winding and can be characterised by high tensile strength, corrosion resistance, and favourable stiffness-to-weight ratios. These properties make GFRP composites suitable for various industries such as aerospace, automotive, marine, and civil engineering. Despite their widespread use, accurately identifying and verifying the mechanical properties of GFRP filament wound structures presents significant challenges. This study addresses these challenges by presenting methods to ascertain and verify the mechanical properties of GFRP filament wound pipes. Commercial pipes from Plaston-P composed of an inner PVC layer and an outer shell of glass fiber roving and mat impregnated with polyester resin were examined. Various mechanical tests were conducted, including tensile, compression, and shear tests, following ASTM standards. This paper describes the steps taken to prepare the specimens required for those tests with a strong focus on reproducing the most representative structure, highlighting potential inaccuracies in parameter identification. Finite element (FE) simulations were performed to verify the obtained parameters, using a nonlinear orthotropic material model with a progressive failure approach. The results showed that the simulated value of the apparent tensile strength of the specimen is 75.94 MPa. The fracture of the element was initiated by failure of the roving-resin layers, which was sudden and brittle. The simulation results were compared with the experimental data obtained from split disk tests according to ASTM D2290. The average apparent tensile stress from the experiment was 80.65 MPa and the specimens failed in a brittle manner. The comparison showed a satisfactory correlation between the simulation and the experiment with a value difference of approximately 6%. The failure mechanism was also identical. It proves that the adopted method of identification allows the mechanical properties to be characterised correctly. Future research will focus on improving the correlation between the simulation and experiment by incorporating parameters to account for delamination and continuous damage of the composite.