Nowadays, demand in composite material such as Carbon Fiber Reinforced Plastic (CFRP) and Glass Fiber Reinforced Plastic (GFRP) keeps on increasing. This creates a substantial rise in the need for fabric cutting especially from the aerospace industry. Therefore, the production cadence has to be significantly increased while keeping a high level of precision. To face these tremendous challenges, automation of this process is unavoidable. This paper presents a discussion on the difficulties in automating textile cutting and how they can be overcome using a robotic system. More precisely, this automation project relies on the infrastructure developed by the 3D Textiles Laboratory of the CTT Group. This laboratory is equipped with a KUKA KR-100 installed on a 6 meter linear rail. The end-effector of the robot takes advantage of a pneumatic tool changer for the robot to use, among others, an ultra-sonic cutting blade as shown in Figure 1. Also, a vacuum table (on the bottom left of Figure 1) ensures the fabric is secured during the cutting operations. This table can hold fabric with a maximal dimension of 1.5 meter by 2.5 meter. To operate the table, the fabric is first unrolled on the table and then, covered with a plastic sheet to ensure sealing of the vacuum.
The common problems of the robotic ultra-sonic cutting process are divided into six categories. First, 3D textiles possess a wide range of structure. Therefore, some type of structure are harder to cut in specific directions, for example, layers in a non-crimp fabric (NCF) are not linked together allowing them to slip during the contact with the knife (Figure 2). Then, the precision of an industrial robot such as a KUKA KR-100 is not constant in it workspace. Therefore, some irregularities can occur during the cutting process, ruining sometimes the raw material as shown in Figure 3. In addition, if a joint velocity limit is exceeded during a cut, the robot engages its brakes and stops as quickly as possible. However, as the blade is in contact with the vacuum table, this can cause serious issues like damage to the table, the tool and the fabric. Also, the typical method to program an industrial robot by manually positioning the robot to generate the desired path is impractical here. Indeed, for complex patterns this task is very time consuming. Additionally, a common requirement when cutting for CFRP is a ±45° angle cut. However, angle cuts presents a challenge because fibers tend to slip on the blade and lift the plastic sheet causing vacuum leaks (Figure 4). Finally, the raw material (rolled on a carton spool) often needs to be replaced to ensure that the fibers are aligned to the table. This task also requires a lot of time and can cause errors in the fiber orientation of the final CFRP part.
With this scope, the optimization of the robotic process takes six steps. First, as every textile pattern presents different challenges, optimal cutting parameters are obtained with an identification program which allows the user to compare the resulting cut with different parameters. In addition, an optimal cutting precision is obtained using a technique computing the Jacobian matrix of the robot. Indeed, as the robotic system has seven degrees of freedom (DOF), it is possible to use the additional DOF to minimize the conditioning of the Jacobian matrix. Then, since the process requires high velocities, another technique is proposed to ensure that the joint velocities stay within acceptable limits. This technique use a lever effect of the two first joints of the robot to transfer fast Cartesian motion to slow joint motion. Subsequently, the implementation of the workcell in a commercial robotic simulator is presented. This simulator allows the operator to plan and to visualize the cutting operations with the scope of minimizing useless robot motion as well as material wastes. Also, a novel cutting technique particularly efficient when using the robot is presented to perform ±45° cuts through a textile fabric without damaging the fibers. This technic also allows the use of only one plastic cover sheet for every cutting operation. Finally, an assistance device was designed and implemented to help the operator to install and orient the fabric on the table. Indeed, the orientation of the fibers of the fabric is a primary concern to ensure the proper mechanical properties of the CFRP. This device allows the stacking of up to four fabric layers.
To conclude, this paper presents the results of the optimization of the fabric cutting process for CFRP fabrication of the CTT Group. In more details, it was possible to decrease approximately 10 times the total duration of a typical cutting operation by using this approach combining a simulation program, layer stacking cuts and single plastic sheet use. Also, the assistance device ensures that the fibers are properly aligned with the table reducing error in the fiber orientation.
Topics: Composites manufacturing, joining & repair , Topics: Process automation/robotization
