Lightweight Design worldwide

, Volume 11, Issue 4, pp 54–59 | Cite as

Flexible Production of Thermoset FRP Components

  • Christian Hopmann
  • Max Ophüls
  • Markus Hildebrandt
  • Kai Fischer
Production Impregnation Process

The Institute of Plastics Processing at RWTH Aachen University has developed a process for the economical production of continuous fiber-reinforced, thermoset components. For this purpose, the researchers have expanded the double diaphragm forming process for the processing of thermosets. The process offers a high degree of flexibility even for small and medium series.

Automation for Small Series Production

Endless Fiber-Reinforced Plastics (FRP) are used in a wide variety of applications due to their outstanding potential for lightweight design. The production scope ranges from large series applications to manual, individualized parts with lot size one. The production of FRP parts in general is influenced by the increasing diversification and individualization of products. This desire for individualized production at low prices presents manufacturers with challenges in the areas of product development, process flexibility and cost efficiency.

Looking at the current FRP market, it becomes clear that despite the high growth rates in the field of thermoplastic reinforced plastics (TPC), thermoset matrix systems still account for the majority of applications. Highly automated production processes for processing long fiber-reinforced molding compounds — the annual production quantity of Sheet Molding Compound (SMC) and Bulk Molding Compound (BMC) is 274 kt [1] — and the manufacture of components using Resin Transfer Molding (RTM) with a production quantity of 141 kt [1] are large areas of application. However, due to high tooling costs, these processes are currently not suitable for the production of small and varied series as well as individualized products. For the large number of applications in small quantities, manual open processing methods with a production volume of 237 kt still account for a significant proportion of the total production volume [1, 2] of Glass Fiber-Reinforced Plastics (GFRP) of 1096 kt. These processes are characterized by a low degree of automation, long cycle times and human-dependent reproducibility, especially in the areas of manual lamination and prepreg processing.

Thermoset matrix systems still account for the majority of applications.

The aim of current work at the IKV is to provide a flexible process chain that makes it possible to react quickly and flexibly to new market and customer requirements, to improve reproducibility and automation compared to manual processes and at the same time to minimize the costs for plant and tool technology in the sense of individualized production. The Reactive Double Diaphragm Forming process (R-DDF) currently developed at RWTH Aachen University offers precisely this process flexibility combined with a high degree of automation.

Double Diaphragm Forming

The reactive double diaphragm forming process represents a further development of the diaphragm forming process. The impregnation bag, into which the dry semi-finished fiber product is placed, plays a key role here. The dry semi-finished products are stacked according to the respective component requirements and placed in the impregnation bag. The desired layer structure can either be inserted as a pre-impregnated prepreg in the impregnation bag, or it can be individually constructed directly from individual layers and welded in the impregnation bag. The impregnation bag enables simple and quick handling during processing on the one hand and prevents the system technology from getting dirty during the impregnation of the semi-finished fiber products with a low- viscosity reactive resin system on the other. The second point results in particular from the fact that in the R-DDF process, unlike in the classical RTM and pressing processes, no closed, two-sided tool technology is used, but the resin is instead injected directly into the impregnation bag. The impregnation bag is shown in Figure 1. The impregnation bag was made of two welded polyamide 6 (PA 6) films from Airtech Europe. The PA 6 film has a thickness of 0.05 mm. An opening is provided in the middle of the bag for the injection of the resin system, the lateral openings serve for evacuation.
Figure 1

Impregnation bag for the R-DDF process (© IKV)

The process chain of the R-DDF process is shown in Figure 2. The impregnation bag is placed on a silicone membrane in the DDF system, which is clamped into a transport frame. After the resin system is applied, a second silicone membrane is automatically applied and the area between the membranes is evacuated. The transport frame (shuttle) is then moved into a heating station. Here the resin system is heated under pressure, achieving a good impregnation of the fibers. The final geometry of the component is then formed in a forming station. The forming is carried out by means of air pressure into a heated tool on one side. Tools made of aluminum or steel can be used here. In order to enable individualized production, however, the use of geometrically variable or adaptive tools or alternative materials such as thermally conductive molding resins is also possible. An example of the successful use of an adaptive tool concept is presented in [3]. This step must be completed before the gel point of the resin system is exceeded. The component can be removed after the curing process.
Figure 2

R-DDF process schematic (© IKV)

The spatial separation of the process steps “Impregnation” and “Forming and Curing” as well as the automated process technology via the shuttle system make it possible to process resin systems that, with a gelation time after approximately 3 min, react much faster than open processing methods.

Process-technological Investigations

The following section presents selected results on the influence of the impregnation time on the accuracy of the forming process.


The forming accuracy and the formability depending on the process parameters are evaluated by investigations on a trunctuated pyramid geometry. This geometry is particularly suitable for investigations because of the pronounced corner areas and edges, as small differences between fiber structures and the different impregnation times can thus also be made visible. The analysis of the forming accuracy is carried out by means of strip light projection with a Comet L3D measuring device from Steinbichler Optotechnik. The projection unit projects a coded stripe pattern onto the target object. This coded projection is recorded with two cameras and converted into a 3-D point cloud by means of triangulation calculations. The point cloud is transferred to an element mesh and can then be evaluated using the Inspect evaluation software from GOM. The actual geometry (determined mesh) is compared with the target geometry (CAD model). The comparison is made at the four corner points, as the strongest deviations from the nominal geometry are expected here. The position of the measuring points and an exemplary target-performance comparison are shown in Figure 4.
Figure 4

Measurement of impression accuracy at the example of a pyramidal structure (© IKV)

Materials Applied

Due to their excellent mechanical and thermal properties, epoxy resin systems are used for numerous structural applications in the aerospace and automotive sectors. Due to the high economic relevance of this material, an epoxy system is selected for the investigations. An epoxy resin system of type Epikote 05475 from Hexion with the reaction agent Epikure 05443 from the same manufacturer is used for the tests.

Resin systems that react much faster can be processed.

The processing window of the resin system was determined in preliminary tests on a Rheostress RS 75 rotary plate rheometer from Haake. Plates with a radius r of 10 mm, a plate spacing d of 0.5 mm and a rotation frequency ω of 10 s-1 were selected as test settings. A plate start temperature of 120 °C was selected. According to [5], a resin system is formable until its dynamic viscosity rises above 10 Pas and thus gels. This assumption is adopted for the resin system used to determine the processing time. The time between addition of the hardener component and exceeding the gel point is measured to determine the maximum processing window. The determined curve of the dynamic viscosity during processing time is shown in Figure 5.
Figure 5

Viscosity versus processing time behavior of Epikote 05475 (© IKV)

In the R-DDF process, the processing time includes the process preparation times, i.e. mixing, impregnation and forming times. A cycle time of 5 min, composed of impregnation duration and reaction time, is aimed for, which is why the processing time should be between 180 and 300 s, which is given for the present system.

The influence of the semi-finished fiber products used is investigated using various semi-finished products. A Fabric with 2/2 twill weave (GK 2-2) from P-D Interglas Technologies and a 0°/90° biaxial non-crimped fabric (Draoptex) from the company Gustav Gerster is used. The twill-woven fabric product has a base weight of 390 g/m2 and the non-crimped fabric has a base weight of 1029 g/m2. In order to achieve a comparable component thickness, five layers of the twill-woven fabric are stacked, or two layers of the non-crimped fabric, with the structure 90°/0°/0°/90° selected here for symmetry reasons.

Impregnation Time and Forming Accuracy

In order to analyze the influence of the impregnation time in the heating station on the forming accuracy, the impregnation time is successively increased from 30 s by 20 s to 110 s in each case. The temperatures in the heating press and in the forming tool are both set to 120 °C. The impregnation pressure in the heating press is a constant 3 bar across all test points, the forming pressure in the forming station is 6.5 bar. The forming time is also kept constant at 250 s. The determined deviations of the component geometry on the outside from the nominal geometry are shown in Figure 6.
Figure 6

Influence of impregnation time on forming accuracy on the outside of the part (© IKV)

For the short heating times of 30 and 50 s, the deviation from the nominal geometry on the outside is very small at approximately 0.25 mm. With these test settings, the tool contour is shaped almost precisely. If the impregnation time is further increased, the impression quality of the tool contour deteriorates. With an impregnation time of 110 s, the outside deviates 10 mm from the nominal contour of the tool. This deviation cannot be tolerated in this form, which is why we must speak here of an incomplete forming behavior. However, it should be noted that large pure resin areas are formed in the corner areas. This means that the tool contour is reproduced exactly on the outside of the component, but the drape of the fibers is insufficient. By extending the impregnation time, the amount of heat supplied is increased. This initially reduces the viscosity of the resin system and the semi-finished fiber can be better draped, which leads to improved forming accuracy. With a further increase in the impregnation time, the crosslinking reaction of the resin begins, while the viscosity of the entire system increases. The semi-finished product is more difficult to drape and no longer forms the corner areas of the truncated pyramid. This effect is further intensified when increasing impregnation time, as the crosslinking progresses further. However, the increase in impregnation time and the associated increase in viscosity during forming also mean that the resin is not pressed out of the fiber bundles during forming and thus forms large areas of pure resin. Figure 7 shows an exemplary microscopic image of a corner area with a pure resin area (left) and dependence of the size of the pure resin area on the impregnation time (right).
Figure 7

Neat resin area depending on the impregnation time (© IKV)

The high potential results in particular from the low tooling costs.

It can be seen that for an impregnation time of 70 s almost no resin is pressed out of the fibre bundles and the pure resin area has thus disappeared. With regard to the resulting pure resin areas, the best test results are achieved, which is why the impregnation time of 70 s is kept constant for further investigations. However, it should be noted that the quality of the forming accuracy also deviates from industrial quality at this process point by 5 mm. However, it should be noted that very strong flank angles of a modular tool were used.


The R-DDF process promises great potential for the cost-effective, flexible production of fiber-reinforced components with thermoset matrix. The high potential results in particular from the low tooling costs compared to conventional liquid impregnation processes such as the RTM process with simultaneous high automation capability and short cycle times if less than 300 s. Cost-effective starting materials can be used, which is a clear advantage in terms of product variability, especially in comparison to pre-impregnated semi-finished products. An impregnation bag minimizes the cleaning effort between two production cycles, as the liquid resin system is never in direct contact with machine components. Furthermore, the impregnation bag can remain on the component as a protective film during further transport.

The examination of forming accuracy reveals the current shortcomings in the draping of the semi-finished products. However, it should be noted that the modular truncated pyramid tool represents a chicane geometry with steep flank angles. It can be assumed that flatter geometries with lower degrees of deformation can still be formed with the method. In order to nevertheless increase the forming accuracy of three-dimensional geometries, further investigations are carried out using a matched die or another tool concept [4] with an adaptive cavity surface. Furthermore it is necessary to transfer the extension of the R-DDF process to the processing of carbon fibers, which have different draping and impregnation properties than glass fibers.

Main Influencing Factors During the Forming Process

The main factors influencing the component quality of three-dimensional components in the R-DDF process are the inherent reaction kinetics of the resin system, the parameters for impregnating the dry semi-finished products (temperature, pressure and impregnation time) and the forming parameters (forming pressure, curing temperature). Here, heating temperature and pressure are limited by material-side or machine-side boundary conditions. Due to the characteristic viscosity curve of a thermosetting resin system during processing, the impregnation time is a particularly important parameter.

Mold Concept with Adaptive Cavity Surface

By combining modular and adaptive tooling technology, new degrees of complexity and flexibility can be achieved in the manufacture of thermoset components.

The required cavity geometry is represented by adaptively adjustable positioning pins. The mold has two mold halves with 3600 pins each for imaging the cavity surfaces. Modular metallic standard geometries as well as additively manufactured components can be used to image filigree structures in order to increase the possible geometrical complexity. In addition, indirect temperature control of the cavity via the positioning pins and direct temperature control of the semi-finished products are to be made possible as well. Figure 3 schematically shows the tool concept and essential properties.
Figure 3

Schematic representation of the lower tool half of the adaptive tool (© IKV)

A central challenge in production with a geometrically adaptive tool is the adjustment of the cavity surfaces for each new component cycle. Within the framework of the mold concept to be developed for the repair strategy, passive approaches to the adjustment of the cavity surface are pursued [3, 4].


  1. [1]
    Witten, E.: Composites-Marktbericht 2016 Der GFK-Markt Europa AVK (ed.): November 2016Google Scholar
  2. [2]
    Acmite Market Intelligence: World Carbon Fiber Composite Market. Marekt report, 2010Google Scholar
  3. [3]
    Hopmann, C.; Wagner, P. N.; Fischer, K.; Böttcher, A.: Industry 4.0 Repairs of Automotive CFRP Parts by Individual Repair Patch Production with Geometrically Adaptive Manufacturing Molds. Proceedings SAMPE USA 2017, Seattle (USA), 2017, pp. 816–829Google Scholar
  4. [4]
    Hopmann, C.; Kuhn, C.; Wagner, P.; Ophüls, M.; Fischer, K.: Endlosfaserverstärkte Strukturbauteile individualisiert produzieren. Umdruck zum 29. Internationalen Kolloquium Kunststofftechnik, Aachen, 2018Google Scholar
  5. [5]
    Neitzel, M.; Mitschang, P.; Breuer, U.: Handbuch Verbundwerkstoffe. Werkstoffe, Verarbeitung, Anwendung. München/Wien: Carl Hanser Verlag, 2014Google Scholar

Copyright information

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2018

Authors and Affiliations

  • Christian Hopmann
    • 1
  • Max Ophüls
    • 1
  • Markus Hildebrandt
    • 1
  • Kai Fischer
    • 1
  1. 1.Institute for Plastics Processing (IKV)RWTH Aachen UniversityAachenGermany

Personalised recommendations