Jeffrey M. Lawrence

An approach to couple mold design and on-line control to manufacture complex composite parts by resin transfer molding

During impregnation of the resin into a closed mold containing the preform in the resin transfer molding (RTM) process, increased yield of successful parts can be achieved if one could account for the inherent disturbances such as race-tracking and preform variability. One way to address this is to use sensors and actuators to control the resin flow dynamics during the impregnation process to counteract the disturbances. In this paper, we use a mold filling simulation tool to develop a design and control methodology that, with the help of sensors and actuators, identifies the flow disturbance and redirects the resin flow to successfully complete the mold filling process without any voids. The methodology is implemented and experimentally validated for a mold geometry that contains complex features such as tapered regions, rib structures, and thick regions. The flow modeling for features such as ribs and tapered sections are validated independently before integrating them into the mold geometry. The approach encompasses creation of software tools that find the position of the sensors in the mold to identify anticipated disturbances and suggest flow control actions for additional actuators at auxiliary locations to redirect the flow. Laboratory hardware is selected and integrated to automate the filling process. The effectiveness of the methodology is demonstrated by conducting experiments that, with feedback from the sensors, can automate and actively control the flow of the resin to consistently impregnate all the fibers completely despite disturbances in the process.

Design and Testing of a New Injection Approach for Liquid Composite Molding

Resin Transfer Molding (RTM) is used to manufacture advanced fiber-reinforced composite parts. In this process, a fiber preform is placed into the mold cavity and sealed shut. Then, a liquid thermoset resin is injected into the mold through the injection gates and the gates are not closed until the resin arrives at the vent locations. One expects the resin to arrive last at the vents after having saturated the complete preform to obtain a successful part. However, the resin path to impregnate the preform is usually not repeatable due to the inherent nature of not being able to place the fiber preform exactly in the same location in the mold. A micron-size gap between the preform and the mold edge will introduce flow deviations and alter the flow pattern causing the resin to arrive prematurely at the vent locations. Thus, it is desirable to have multiple gate and vent locations, to aid in distributing the resin throughout the mold effectively and efficiently. In this paper, we present the design and validations of an automated RTM intelligent workstation. The workstation will have distributed injection gates and vents that will be triggered by the control methodology selected based on the on-line flow pattern of the resin during the impregnation process. Point sensors placed in the mold walls that will monitor the arrival of the resin will detect the on-line flow pattern. To automate and control the fluid flow during the RTM process it is also necessary to individually open or close various injection locations. A new method, which can deliver the resin to various locations within the mold, while still allowing any individual gate to be opened or closed at any time during the filling, was designed and tested. Development of such versatile injection scheme provides the potential to have better control on the distribution of the resin during impregnation and will lead to successful mold filling without any dry regions despite the presence of disturbances created by the quality and the placement of the perform.

Dependence Map-Based Flow Control to Reduce Void Content in Liquid Composite Molding

ABSTRACT Prohibitive costs are preventing liquid composite molding processes, such as resin transfer molding, from gaining greater popularity. The high costs are mainly due to lack of automation and repeatability. Variations inherent in the preform cutting and placement process lead to unpredictable resin infusion patterns that sometimes form regions devoid of resin. This compromises the structural integrity of the composite and the part is rejected as scrap. Control during the filling stage of liquid composite molding processes have been shown to redirect the flow towards regions difficult to impregnate, providing a greater percentage of acceptable manufactured composite parts. However, many of the current techniques have various limitations. Many off-line control approaches depend on anticipation of problems, and on-line approaches are geometrically limited and can be computationally intensive if executed during manufacturing. The present work combines off-line and on-line approaches to infuse and fill the mold containing fabrics in an attempt to eliminate their shortcomings and reduce the limitations. First, off-line computationally intensive control algorithms based on the specific part geometry and locations for the sensors and injection gates are created. Next, on-line control is initiated with the off-line parameter guidelines. The approach will be presented and illustrated with several case studies to demonstrate geometrical independence in a simulation environment. Comparison will be made between part success rate using no control, other control techniques, and the proposed control approach. Validation in a laboratory environment is also conducted.

Experimental Validation of Dependence Map based Control in Liquid Composite Molding

Liquid composite molding (LCM) processes offer composite makers the environmental benefits realized by the closed molding process, and can provide better dimensional tolerances as well as part complexity. One drawback is the unpredictability in the flow behavior during the filling stage due to material and process variability that can result in resin void areas, or dry spots. Parts containing dry spots must be discarded, thus adding superfluous cost to the process. Flow control has demonstrated the ability to correct those flow anomalies to eliminate dry spots. Previous efforts developed a hybrid on-line–off-line flow controller for the LCM process and detailed the development and application of the methodology. In this study, the methodology is summarized, and the technique is applied in a manufacturing environment. A resin transfer molding workstation is developed and built to meet the needs of the control technique. A series of experiments with three different geometries with increasing complexity is conducted to validate the new flow control methodology. The results are compared with experiments without flow control to gauge the usefulness of the methodology proposed.

The Compaction Behavior of Fibrous Preform Materials during the VARTM Infusion

In Vacuum Assisted Resin Transfer Molding (VARTM), fiber preform is placed on a rigid tool and sealed with a flexible plastic bag. Vacuum is drawn at the vent to compact the preform under atmospheric pressure with the injection line closed. When the injection line is opened the resin flows into the preform. The pressure of the incoming resin relieves some of the pressure being borne by the preform which increases the thickness of the preform during the infusion process. As the resin pressure varies from the injection location to the vent, the thickness of the part will not be uniform due to this resin pressure variation and its coupling with the compression characteristics of the preform. Compaction studies of the preform have identified elastic, plastic, and even viscoelastic behavior and a change in behavior as one goes from a dry preform to a wet preform (lubricated). This paper will describe results from a model that couples the fiber compaction behavior with the resin infusion process. Experiments wi...