JEC COMPOSITES MAGAZINE - Issue #57 - May 2010 - 51

The potential of such a control device was explored by studying the mechanical behaviour of various placement patterns. The force cell fitted on the placement head can monitor the force variation response during the placement process to verify the local mechanical properties of the preform. For this purpose, a compression test is carried out on-line, providing information on the changing compliance of the preform and detecting process deviations from the ideal scenario. The advantage of this approach is that it considers the entire placement history of the whole preform, whereas visual control only gives information on the last-placed layer (provided that the contrast is sufficient). The control methodology was calibrated by characterizing different fibre-placement patterns with a compression test prior to implementation in the robot programme. The preform compaction response is expected to change, which would indicate a different pattern or a placement problem. On-line visual control can be used as a complementary defect detection tool on an industrial device, or to clearly define the type of defect mechanically detected. The control actions required will be defined later, as they depend on the impact of the defect on the macroscopic properties of the preform, which implies another step in the current research programme.

Mechanical identification of preform lay-up patterns
The preforms produced were mechanically tested with successive compressive tests simulating the application of new plies. The preforms were tested between two flat platens mounted on a Zwick compression test unit. In a first step, the maximum fibre volume fraction that could be reached for each preform type was determined as a function of the load applied by the head and the roller radius. The load applied on each preform type and the fibre volume fraction derived from the preform thickness were monitored during the loading phase, as shown in Figure 3.

Reference lay-up patterns
The first step in this development consists in obtaining preforms with different pattern types in terms of fibre-tow spacing. Three patterns are proposed, representing configurations that can be found in industrial preforms, such as local displacement of a fibre tow, either overlapping the neighbouring tow or creating a gap. The maximum displacement observed is about 2 mm. The first pattern, with regularly placed fibre tows, is the nominal pattern. The second one is the overlapping pattern, where one out of five fibre tows overlaps the preceding one by 2 mm. The third pattern is the open pattern, where one out of five fibre tows is at a 2-mm distance from the adjacent fibre tow. The preforms are made of 8 layers oriented in different directions according to the defined stacking sequence [90,45,0,-45]s. The preforms are produced automatically by the robot. The stacking sequence and spacing between the fibre tows are implemented in the robot software. The three reference preforms produced are presented in Figure 2.

Fig. 3: Fibre volume fraction response as a function of the applied force

Fig. 2: Preform patterns designed and produced for reference

The preforms were loaded at a compression speed of 0.5 mm/min up to 90 kN. At a given load, the fibre volume fraction is lower for the samples cut from an open-pattern preform and higher for the overlap-pattern samples. The overlap pattern facilitates microstructural fibre arrangement when pressure is applied during fibre placement, resulting in a higher fibre volume fraction. Conversely, the open pattern retains gaps that are not filled during the compression process, reducing the fibre volume fraction. Compressive cycle tests were conducted in a second step. Eight successive compressive loading and unloading cycles were performed on samples of the different preforms, representing the application of eight extra layers. The compression load was defined according to the contact surface with the fibre preform and the load applied during manufacturing. The parameter of interest is the evolution of the preform’s plastic deformation that would represent a change in the preform thickness after one compression cycle. Figure 4 shows the evolution of plastic deformation after each compression cycle for all three preform types. For all three tested preforms, most of the plastic deformation was generated during the first compressive cycle. In the subsequent cycles, the additional plasticity per cycle remained small. The open preforms exhibited higher

No57 May 2010 / jec composites magazine

JEC COMPOSITES MAGAZINE - Issue #57 - May 2010

Table of Contents for the Digital Edition of JEC COMPOSITES MAGAZINE - Issue #57 - May 2010