Lightweight Design worldwide

, Volume 10, Issue 2, pp 36–41 | Cite as

Industry 4.0 Enabler for CFRP Repairs in Vehicles

  • Christian Hopmann
  • Lutz Eckstein
  • Robert Schmitt
  • Uwe Reisgen
Construction CFRP Repair


Point Cloud Data Management System Damage Detection Damage Assessment Tool Concept 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Increasing production figures of FRP-based cars and the manufacturers’ variety of individual components require a state of the art automated repair processes. Together with 60 industry partners, the RWTH Aachen University is currently working on a project for an integrated repair strategy. It is based on standardised damage detection, a database-oriented damage assessment as well as the individualised production of repair patches conforming to Industry 4.0 guidelines.

Current situation

Carbon-fibre reinforced plastics (CFRP) offer a significant contribution to an increase in energy efficiency of cars. It follows, that the everyday use of CFRP-part containing cars results in a larger amount of damaged CFRP parts which can’t be repaired by using standard methods common in car body construction. In addition, it is difficult to detect damages in CFRP structures in comparison to metal parts. Damages can negatively affect mechanical properties of CFRP parts without showing any indication on the surface of the parts.

The application of different materials and implementation of CFRP components in modern vehicles results in complex repair procedures. Currently, repairs are done by the cars’ manufacturer and require a high amount of model- and brand-specific competence. They include the replacement of entire components or predefined repair sections, without taking the actual extent of damage into account. Repairs via individual repair patches which fit the precise damage geometry and replace the affected area provide a material-efficient and lower-cost alternative. Therefore, at RWTH Aachen University the Institute of Plastics Processing (IKV) and the six research institutes ika, ISF, IBF, WZL, MMI, FIR as well as 60 industry partners develop new technologies for the detection and assessment of damages, the individualised production of repair patches and the installation of individual repair patches with an overarching quality control in place.


The repair strategy employs Industry 4.0 principles and methods to establish a collective data management and information platform for all repair steps. Decentralised multisensory-based damage detection at a local workshop is followed by a central damage assessment and the specification of local repair measures. Accordingly, individual repair patches are to be manufactured centrally and shipped to the decentralised workshops. Repairs are then completed through a sensor-based quality assurance of the repaired area, conforming to common standards. All complex singular aspects and correlations that are developed in regards to the repair strategy also hold great relevance in other industrial fields and can be equally employed there. This article outlines the repair strategy in depicted in the entire process chain of the repair procedure, Figure 1. The basis is a CFRP demonstrator part (cap profile).
Figure 1

Process chain of the repair strategy (© RWTH Aachen University)

Data Management

All data obtained within the individual repair steps is saved in a centralised data storage and management system within which a digital twin model of the damaged component is created. The data storage system includes interfaces for a partly-automated data exchange with any technical system used during the individual steps of the repair strategy. Therefore, the data management system needs to be able to import and export various kinds of data formats, including but not limited to sensor images, point clouds, material charts and different geometry exchange formats. Additionally, the data storage system features a damage database which makes an easy and quick damage assessment of already known repair cases possible. In the future, the data management system includes the possibility of adding further functions such as assistance systems for the workers involved in the repair process or virtual test beds for training and simulation-based repair optimisations.

Damaging a Component

An analysis of accident frequency, cause and type of damage, showed that the repair of a side sill is a particularly relevant case. For this purpose, a cap profile serves hereinafter as a component resembling the geometry of a side sill. Damage caused by ground contact or smaller collisions with the lower part of the vehicle are initated through controlled impacts from a drop tower. Here an impactor with a spherical head hits the cap profile with an energy input of 9 J, Figure 2. The point of impact is positioned slightly out of centre in order to cause the damage to spread over the edge of the cap profile. The high-speed recordings reveal a noticeable dynamic deformation in the component which completely returns to the original shape afterwards. This illustrates the insufficiency of a singular visual damage assessment for an accurate analysis of the damage and its effect.
Figure 2

Drop tower experiments for pre-damaging of the parts and the recovery of dynamic deformations (© RWTH Aachen University)

Geometry and Damage Detection

The vehicle repair procedure and repair patch production are based on the previously described data management system. It enables the data fusion of all repair steps, digital mapping of the damage to the part as well as the establishment of a collective damage register for a quick and efficient damage assessment. In order to use local repair patches, CAD data of the part area that is supposed to be replaced needs to be acquired through a manual laser scannig system, Figure 3. A hand-operated scanner projects a laser line onto the damaged part and surrounding area. During the movement over relevant parts a tracker registers the line offset and scanner position. During the digitisation process, a point cloud of the measured object is continuously generated, transmitted and converted into a CAD model by reverse engineering the surface. As an intermediate step, a polygon model is created from the point cloud. For the construction of a CAD model, all polygons are approximated by so-called NURBS surfaces (non-uniform rational B-splines). Data noise as well as possible measuring errors are compensated through specialised filter techniques.
Figure 3

Geometry detection with hand-operated scanner (© RWTH Aachen University)

As depicted in the schematic process in Figure 4, the multisensory-based damage detection combines various non-destructive testing methods to categorise and assess the damage. After the part geometry has been established, the damage position and region of interest (ROI) are determined with a large-area testing, active lock-in thermography. The active lock-in thermography is based on a thermal stimulation of the surface of the part, during which the heat is transferred inside the part. Any damages inside the structure will alter its thermal conductivity compared to unaffected parts of the CFRP laminate which results in visible heat peaks when comparing multiple infrared (IR) images during a subsequent cooling process of the part. Within the ROI the detection of the damage geometry is conducted with ultrasonic testing. The unit is traced along the surface of the part while emitting ultrasonic impulse. Through changes in the impulses’ length, transit time or intensity the geometry of the damaged part can be visualised three-dimensionally. The resulting geometry model acts as the basis for an assessment of the damage severity and extent of a possible repair.
Figure 4

Schematic sequence of the damage detection as the basis for the damage assessment (© RWTH Aachen University)

A singular visual damage assessment is insufficient for an accurate analysis of the damage and its effect.

Damage Assessment and Repair Method

Figure 5 schematically depicts the damage assessment process. First, all data from the multisensory-based damage detection and geometry data is merged in a 3-D model. In case of an unknown case of damage, this model is implemented in an FE model of the component that includes the extent of damage. This model of the damaged component is simulated with regard to its mechanical properties under different load cases and compared to those of an intact component. Any load cases used here are deduced from load cases used in the design of complete vehicles. If there is a significant decline in its mechanical properties the component will have to be repaired — or if need be replaced entirely. In case of a needed repair, a method of repair gets selected and configured on an FE basis. The size of the area that is to be removed, as well as the geometry of structural part preparation (for example scarfing, doublers, mechanical joints) are given. Additionally, the repair patch is designed, with regard to size, laminate build-up and preparative measures for adhesive joining. The repair data is sent to a central patch manufacturer for production. It is also passed on to the local repair work shop for the treatment of the damaged structure. As a last step, the new established case of damage is registered in the collective database. Should a similar case occur in the future all essential information can be taken from the database and a renewed FE based assessment of the damage is rendered unnecessary.
Figure 5

Schematic sequence of the damage assessment and the selection of a repair method (© RWTH Aachen University)

Tool Design and Manufacturing of FRP Repair Patches

The multitude of possible component geometries corresponds to a large diversity of repair patches necessary to recreate the area of damage as fast and flexible as possible. Such flexibility in geometry cannot be covered economically by conventional tool concepts. This is why a new tool concept for the repair strategy is developed. For the development economic efficiency and speed in the production of different patch geometries are key elements and the tools cavity should form based on the geometry detection and patch design of the damage evaluation. It follows, that the setting of the tools geometry is done by the service provider in charge of patch production who is in possession of the geometry data due to the collective data management system.

During the research project, the development of a flexible tool concept is based on different trials with several trial tools. On its cavity side, the mould surface was formed by use of incremental sheet forming (IBU) and embedded in a mix of resin and graphite for stabilisation and heating. Whereby short cycle times even for repair patches with thermoset resins can be achieved.

IBU is already being used in the manufacture of prototypes and small-batch productions of sheet components. A sheet of deep-drawing steel is placed inside a frame and fixated tightly. A punch traces the contours of the desired geometry line after line, Figure 6, upper right. During the process the sheet gets locally plasticised and shaped into form. After a review and optimisation, the component can be manufactured in just a few minutes, depending on its size and complexity. It is ready for further use in the CFR repair process directly after production and cleaning.
Bild 6

Production of a trial tool and machine technology of the DDF process for the analysis of final tool specifications and demonstration of the process chain (© RWTH Aachen University)

For the manufacturing of FRP repair patches with a thermoset matrix the dual-diaphragm forming process (DDF) is used. The DDF machine consists of the assembling, heating and forming station. At the assembling station, an FRP semipreg is placed between two diaphragms and fixated by vacuum pressure. A complete consolidation and preliminary cross-linking inside the heating station follow. The vacuum ensures the removal of any air trapped between laminate and diaphragms, which could otherwise lead to a reduction in part quality. After the heating phase is completed the frame moves to the forming station. A circumferential sealing inside the pressure plate creates a pressure-tight space between mould and frame. The laminate is shaped by the negative mould through overpressure from above and fully cross-linked into a finished repair patch. The patch is then milled to match the structural preparations done to the damaged component and sent to the work shop for the subsequent joining process, Figure 6.

The process data gained from the trials in the trial tools is used for the conceptual design of a flexible two-sided tool concept. It is being developed for the production of repair patches by a diaphragm manufacturing process in combination with a pin support structure. Figure 7 shows the tool concept and geometry adjustment. The support pin concept’s high flexibility also enables the combination with other modular tool elements, for example additive manufactured inserts or panels from the IBU. This combination of innovative tool concept and manufacturing process makes the production of thermoset and thermoplastic repair patches possible.
Figure 7

Flexible tool concept for the individual repair patch production at a central manufacturer (© RWTH Aachen University)

Preparation and Joining Procedure

The choice of a repair method and necessary preparations of the damaged part are based on the previous damage assessment. Since the project considers various joining techniques, for example adhesive bonding via scarfing, mechanical joining, employment of doppler for the repair patches, the individual preparations for the damaged part are stored in the data management system. With regard to the economical capabilities of independent workshops - the project’s main target group — research focuses on hand-operated tools.

In this particular case the preparation was done via continuous scarfing. The large angle in the scarfing area offers a larger bonding surface, Bild 8, for better force transmission in compliance with material requirements. Furthermore, this kind of repair is visually and aerodynamically inconspicuous. For the scarfing process grinding is preffered over milling, due to the lower material removal per time which favours the manual handling and guiding of the tool. Industrial vacuum cleaners protect workers from respirable CFRP particles. Other health and safety measures include protective clothing and the enclosure of working areas. Shortly before applying the adhesive to the surface it needs to be pre-treated with an abrasive fleece, which can easily be done in the workshop.

Following the structure preparations, the geometry detection can be repeated on the altered part of the component. The renewed data supports the manufacturing company of the repair patch in the production of a perfectly fitted patch with all necessary joining surfaces so as to avoid any post processing at the work shop.

Adhesive systems used in the repair process need to meet the following criteria for an application in the vehicle:
  • ▸ high mechanical strength for structural adhesive bonding of FRP

  • ▸ viscosity suitable for vertical joints as well as bridging gaps due to geometrical variances

  • ▸ a reproducible curing process under workshop conditions

  • ▸ embedded glass beads to set a specific bonding gap

  • ▸ suitable for sanding/grinding post process.

The adhesive is manually applied with the help of a pneumatically driven glue gun equipped with a static mixer. After the repair patch has been fit into the part and the adhesive is fully cured, post-processing might be necessary (for example removing excess material).
Strictly manual handling of the structure treatment leaves room for improvement concerning reproducibility and functionality, as can be seen in Figure 8 (mid, right). Despite a renewed geometry detection, preparing the damaged part for an exact fitting while taking the bonding gap into account proves to be especially complex. Especially, since the final result should display a homogenous surface
Figure 8

Part preparation and joining process (© RWTH Aachen University)

Summary and Conclusion

In this article interim results of a research project focusing on a repair strategy under Industry 4.0 guidelines for CFRP based vehicle structures are presented. The repair process is outlined through all individual steps with a demonstrative part in style of a side sill.

It is demonstrated that the innovative approach of a decentralised damage detection and repair in workshops, as well as a centralised, database based damage assessment and automated patch production, is expedient in respect of an efficient repair of CFRP components. Any necessary data exchange between the respective repair steps is ensured by a collective data management system in accordance with Industry 4.0 principles.

The individual aspects of the multisensory-based damage detection create a basis for the standardised and quality-assured repair. The subsequent preparation of the vehicle structure and geometry detection of the prepared joining area and consideration for the patch production make the implementation of repair steps possible within the existing tool environment.



The authors express thanks to the co-authors Philipp Nicolas Wagner, Bernd Marx, Katharina Bethlehem-Eichler, Sarah Ekanayake, Markus Gottschalk, Daniel Losch und Kai Fischer.

The research projects 18757N, 18758N and 26 LN in context of the initiative “Leittechnologien für KMU” of the Forschungsvereinigung Kunststoffverarbeitung is sponsored as part of the “Industrielle Gemeinschaftsforschung und —entwicklung (IGF)” by the German Bundesministerium für Wirtschaft und Energie (BMWi) due to an enactment of the German Bundestag through the AiF. The Authors would like to extend their thanks to all organisations mentioned.

Copyright information

© Springer Fachmedien Wiesbaden 2017

Authors and Affiliations

  • Christian Hopmann
    • 1
  • Lutz Eckstein
    • 2
  • Robert Schmitt
    • 3
  • Uwe Reisgen
    • 4
  1. 1.Institute of Plastics Processing (IKV)RWTH Aachen UniversityAachenGermany
  2. 2.Institute for Automotive Engineering (ika)RWTH Aachen UniversityAachenGermany
  3. 3.Laboratory for Machine Tools and Production Engineering (WZL)RWTH Aachen UniversityAachenGermany
  4. 4.Welding and Joining Institute (ISF)RWTH Aachen UniversityAachenGermany

Personalised recommendations