Search

Pulling force modeling and analysis in conventional pultrusion and Resin Injection Pultrusion (RIP) of vinylester resin were carried out in this study. Based on the friction coefficient value of the liquid resin, an analytical model was developed to predict the resistance force in the injection die. A thermal expansion– polymerization shrinkage model and a friction coefficient model, combined with the temperature and resin conversion profiles along the die, were used to predict the resistance stress in the heating die, and to determine the composite separation point from the die wall. The simulation results were verified by the experimental data in the conventional pultrusion and RIP processes. The comparison shows that the models and the simulation tool developed in this study are capable of predicting the pulling force in the entire die with good agreement.

Pulling force is the summation of resistance forces along a pultrusion die. Its mechanism is complicated because the impregnated resin inside the compressed fiber reinforcement changes from a liquid to a gel, and finally to a solid in the die. Two methods, a ‘mat tracer’method and a ‘short die length’ method, were used to determine the pulling force distribution. Results show that the downstream part of the die contributes little to the pulling force. In order to predict the resistance force in both the injection die and the heating die, two models were developed in this study. A friction measurement device was designed to measure the friction coefficient between the composite and the mold surface. The effect of process variables, such as temperature, resin conversion, normal force, and line speed, on the friction coefficient was investigated and a friction coefficient model was proposed based on the experimental results. A volume change model of vinylester resin was also developed to predict the thermal expansion–polymerization shrinkage during curing. The parameters of the model were determined by dilatometry, thermomechanical analysis (TMA) and differential scanning calorimetry (DSC).

Blister formation, a major concern in pultrusion, has not been studied in detail. The objective of this study is to investigate the mechanism of blister formation and the effect of process variables such as die temperature, pulling speed, die length, and composite thickness on blister formation. Dies with different length and a stop-start method were used to investigate blister formation. The results show that for a given die and resin system, low temperature and a higher pulling speed lead to blister formation. Longer dies can prevent blister formation. Based on the experimental results, process windows for 4-mm-thick and 6-mm-thick composites are established for a vinyl ester resin. This study suggests that heater power input should be optimized in high-speed pultrusion.

In this study, a computer simulation code is developed to predict the dynamics of heat transfer in the pultrusion process. The die block and heater arrangement are included in the heat transfer analysis so that the simulation can provide the temperature profile at the interface between the die and the composite. The measured interface temperature profiles are then used to validate the simulation code. Energy management, i.e. heater power control along the heating die, is also considered in the simulation code. The code is capable of carrying out transient thermal analysis for both start-up and steady-state operation of the pultrusion process. From the experimental observations on the part quality in terms of blister formation, a processing window was obtained by showing the relation of the die length and the line speed to the part quality. The processing window is then generated numerically using the computer code based on the definition of a critical die length proposed in this work, and the result shows good agreement with the experimental data.

Gas diffusion is important in composite manufacturing, while moisture diffusion is important for composite performance. In this study, analytical models are developed to predict the mass diffusivity of both square and hexagonal arrays of cylindrical cylinders. These models are able to bridge the gap between continuum calculations via finite differencing of finite element analysis and simple/trivial models. By breaking the liquid region into zones and making some general assumptions about the diffusion in those zones, simplified models have been developed, which rely solely on the geometry and fiber volume fraction. The models are fairly accurate for a broad porosity range, but specially at low porosity values. A single weight factor is proposed to compensate for the error introduced by the model assumptions. When the weight factor is incorporated, the modified models agree extremely well with the experimental data for the whole porosity range.

This work presents the results of modeling, numerical simulation, and experimental study of resin flow and heat transfer in the resin injection pultrusion (RIP) process. A control volume/finite element method (CV/FEM) was used to solve the flow governing equations, together with heat transfer and chemical reaction models. Resin viscosity, degree of cure, and fiber stack compressibility and permeability were measured in order to understand their influences on the process. An analytical flow model has also been developed based on the one-dimensional flow approximation for the resin flow in the injection die. A high-pressure small-taper injection die was tested with different line speeds. Experimental data were used to verify the simulation results and the analytical solutions.

This work investigates the preforming process of composite ducts made of biaxial braided fabrics. Two models based on the geometric modeling were developed to predict the fiber orientation and surface density. It is found that fabrics tend to slip during preforming in the case of curved ducts. A slip coefficient was introduced in the model to take this into account. A series of experiments were conducted to validate the models. The results show good agreement between model prediction and experiments.

Compared with SCRIMP based on the high-permeable medium, SCRIMP based on grooves has the advantage of a much higher mold filling rate. This paper analyzes the influences of various molding conditions on mold filling and presents models that can be used to predict the filling time and flow pattern in SCRIMP based on grooves. Mold filling experiments were carried out to investigate the effect of various factors such as the size of the groove, groove spacing, number of fiber layers and resin viscosity on mold filling. A leakage flow model was developed to simplify the simulation of the mold filling process in SCRIMP based on grooves. An “equivalent permeability” was introduced to represent the average flow capacity in the grooves. Compared with the Control Volume/Finite Element Method (CV/FEM) model, the leakage flow model greatly reduced computation time and yet provided simulation results that were in good agreement with experimental observations.

Mold filling in SCRIMP based on a high-permeable medium is complicated because of the considerable difference in the permeabilities of the fiber reinforcement, the peel ply and the high permeable medium. The objectives of this paper are to understand the flow mechanism through flow visualization experiments and to present models that can be used to predict the filling time and flow pattern. Permeabilities of a stitched fiber mat, a high-permeable medium and a peel ply were measured. Flow visualization of SCRIMP mold filling was carried out under various molding conditions. It was found that although the resin flowed faster in the high-permeable medium than in the fiber reinforcement, the flow front lead-lag was not very large and it remained nearly constant through the entire mold filling process. A three-dimensional Control Volume/Finite Element Method (CV/FEM) was adopted to solve the flow governing equations, i.e. the Darcy's law, and the influences of the flow properties of the high-permeable medium, the fiber reinforcement and the peel ply on filling time were investigated. Based on experimental observations and CV/FEM simulation, a simplified leakage flow model is also presented. The comparison of experimental and simulation results show good agreement.

This paper investigates the manufacturing method and mechanical properties of composite tubes made from braided fiber preforms. The tubes molded by Vacuum-Assisted Resin Transfer Molding (VARTM) process with different resins, fiber reinforcements and braiding architecture were tested for weeping and burst strength. Both biaxial and triaxial braided fiber preforms were used. The failure modes of tubes made from these two braiding architectures were different. For biaxially reinforced tubes, the initial cracks are aligned along the direction of the fiber tows, while for triaxially reinforced tubes, the initial cracks along the longitudinal direction appear first. Tubes with and without flanges were also fabricated. The strength tests indicated that the tube with flanges is much stronger than the one without flanges. Two theoretical models based upon mechanics of material were employed to predict the mechanical properties. The pressure vs. deformation curves from the weep/burst tests are consistent with the theoretical predictions in the elastic region.