Adaptive Control of Filling Pattern in Resin Transfer Molding Process

Abstract

During the Resin Transfer Molding (RTM) process, the filling pattern is strongly affected by race-tracking phenomenon and this usually leads to dry spot formation at the end of the filling stage and results in defective parts. This situation can be alleviated by placing air vents at the locations of the Last Point to Fill (LPF), where dry spots usually form. However, due to the variation of the race-tracking effects, the LPF location does not remain in the same position during RTM batch cycle production. Dry spots can still form and the part is still scraped. In order to ensure that the LPF location always coincides with the predetermined vent during RTM batch cycle production, an adaptive control strategy based on a model reference adaptation system that can regulate the LPF location during the RTM filling stage is proposed, to adjust the inlet pressures on-line. Multiport flow is simplified as many individual one-dimensional ‘spine’ flows and each ‘spine’ extends from the gate to the vent along the short path. The spine flow is controlled independently by adjusting the pressure of the related inlet. The control input is the distance between the inlet and the flow front along the spines, and the reference model of the flow is that the resin is driven by a uniform velocity from the gate to the vent along the spine. The proposed adaptive control has the capability to adapt to high levels of system disturbances. The proposed adaptive control is tested numerically by incorporating the control algorithm into a RTM time-implicit filling simulation. The resulting simulations show that the proposed control strategy can eliminate the dry spot formation under a wide range of situations resulting from the variations in the race-tracking effects.

Advanced Polymer Composite Molding Through Intelligent Process Analysis and Control

“…Direct adaptive control was tested in [19] in a simulated simple 1-dimensional RTM process. In this work, an algorithm controlling the injection pressure calculates the filling error and uses it as control feedback to make the flow front follow a predefined filling pattern.…”

“…At each time step, the control system retrieves the distance between inlet and flow front and compares it with the predefined filling pattern to calculate the error and correct the pressure for the simulated 1-dimensional RTM process. In [20], Minaie and Chen used flow front positions traced along three spinal sensors as adaptive control feedback parameters in a time-implicit 2-D simulation of flow in porous media. In this study, the filling pattern uncertainty resulting from the racetracking effect was evaluated by changing the permeability along different mold edges.…”

“…In the following case study, instead of real RTM processes, the numerical RTM simulations are used. Hence, this control interface is connected to an implicit 2-dimensional Control Volume Finite Element Method (CVFEM) RTM simulation program [20] that is based on Darcy's Law and the following assumptions: (1) density and viscosity are constant; (2) inertial effects and surface tension are negligible; (3) the preform is rigid and never deflects; and (4) as the resin flows through the cavity, it fully saturates the fiber preform. The listed assumptions induce some simulation limitations on the RTM simulations.…”

“…Since the gap size is affected by many parameters such as preform cutting, stacking and placement, the effect is usually nonrepeatable and greatly reduces the reliability of RTM. To compensate for the effect of these disturbances, flow control strategies and appropriate sensing devices have been employed to minimize dry spot formation in composite parts [9][10][11][12][13][14][15][16][17][18][19][20][21][22]42].…”

“…The basic concept of this approach is to monitor the position of resin flow front during RTM process and use adaptive control algorithms to adjust the injection flow rate or pressure at the different inlet gates to obtain a desirable filling pattern. To effectively track the resin flow front movement, several linear sensors with spinal arrangement, i.e., spinal sensors, are placed from inlet gates to the vent [2,19]. As the flow front moves inside the mold cavity, its profile intercepts each one of these linear sensors [15,17].…”