Abstract
High strength-to-weight ratio, excellent corrosion resistance, flexibility, superior fatigue performance, and cost competitiveness have made thermoplastic fiber reinforced polymer composites (TP-FRPCs) a material of choice for the manufacture of pipe products for use in the oil and gas industry. The TP matrix not only protects the composite structure from brittle cracking caused by dynamic loads, it also provides improved flexibility for bending of pipes to enable easier field installation and reduces the requirement for pre-fabricated bent connections. Despite the attractive mechanical performance, the design, development and qualification evaluation of TP-FRPC components for a large portion relies on experimental testing. The time and expense of manufacturing new composite prototypes and performing full-scale testing emphasizes the value of a predictive modeling. But, modeling TP-FRPC structures is not a trivial task due to their anisotropic and time-dependent properties. In this study, a new technique based on the finite element method is proposed to model anisotropic time-dependent behavior of TP-FRPCs. In the proposed technique the composite mechanical properties are captured by superimposing the properties of two fictitious materials. To that end, two overlapping three-dimensional elements with similar nodes were assigned different material properties. One of the elements is assigned to have time-dependent properties to capture the viscoelastic behavior of the matrix while the other element is given linear anisotropic properties to account for the anisotropy induced by the fiber reinforcement. The model was calibrated using data from uniaxial tensile creep tests on coupons made from pure matrix resin and uniaxial tension tests on TP-FRPC tape parallel to the fiber direction. Combined time hardening creep formulation, ANSYS 19.2 implicit analysis, and ANSYS Composite PrepPost were employed to formulate the three-dimensional finite element model. The model was validated by comparison of model predictions with experimental creep strain obtained from TP FRPC tubes with ±45° fiber layups subjected to uniaxial intermediate and high stress for 8 hours. The results obtained showed that for the tubes subjected to intermediate stress, the model predicted the creep rate in the secondary region with less than 5% error. However, for tubes subjected to high stress, the model overestimated the creep rate with over 30% error. This behavior was due to large deformation at this high level of stress and inability of the model to capture fiber realignment towards the pipe longitudinal direction and, therefore, capture an increase in stiffness. Overall, comparison of the simulation results with experimental data indicated that the technique proposed can be used as a reliable model to account for deformations caused by secondary creep in the design of TP-FRPC structures as far as deformations are relatively small and limited to a certain strain threshold. Acceptable predictions of the model, its simplicity in calibration, and limitations on available models that can simultaneously account for time-dependency and anisotropic properties, further emphasize the value of the developed model.
EXAMINATION OF THERMAL STRESSES OCCURRING IN CIRCULAR DISCS BY FINITE ELEMENT METHOD