Lightweight Design worldwide

, Volume 10, Issue 5, pp 42–47 | Cite as

Effector for automated direct textile placement in rotor blade production

  • Marvin Richrath
  • Jan Franke
  • Jan-Hendrik Ohlendorf
  • Klaus-Dieter Thoben
Production Handling Technical Textiles

The process chain for the production of rotor blades for wind turbines is still mainly characterised by manual process steps. It is planned to reduce the long production times and fluctuations in quality that this causes through automated laying up of dry technical textiles. Scientists from the University of Bremen illustrate the possible development of a suitable effector for the automated direct textile placement.

Automation in Rotor Blade Production

Wind turbine rotor blades are manufactured primarily from fibre reinforced plastics (FRPs), owing to their excellent mechanical properties. The main production method used is the vacuum infusion process, where glass-fibre non-crimp fabrics are inserted manually into the mould. The rotor blade mould with the dry, near-net-shape layer structure, the so-called preform, is then sealed air-tight with a vacuum foil. The space between the mould and the vacuum foil is subsequently evacuated, and the preform impregnated with a resin/hardener mixture and allowed to cure.

A significant quality criterion for the structural properties of the subsequent component is the load-conforming alignment of the non-crimp fabric in the mould. The predominantly manual process step of laying up the fabric requires time and results in reduced quality as a result of the positioning of the fabric, which is often difficult to replicate. Since rotor blade production accounts for a large proportion of the manufacturing costs of a wind turbine — around 21 % — [1], it is intended to meet the high quality demands and the desire for shorter cycle times through automated production processes.

The Institute for Integrated Product Development (BIK) at the University of

Bremen has for several years been researching and developing various technologies and processes for handling technical textiles for the automated manufacture of FRPs. Investigations are performed into handling technologies for various rotor blade components and lay-up strategies along the production chain of a rotor blade. Preforming the root ring and the transition section, Figure 1, of the rotor blade was implemented through an even lay-up of the technical textiles in the research projects “preblade” [2] and “mapretec” [3]. Likewise, the automated spar cap lay-up along the slightly rotated and curved surfaces of the rotor blade was developed in the “LAR” [4] joint project.
Figure 1

Layout of a modern rotor blade, divided into relevant rotor blade sections (following [5]) (© BIK | Institut für integrierte Produktentwicklung)

As part of the “BladeMaker” joint project [6], the BIK is pursuing the alternative strategy of continuous lay-up in a multi curved mould in the sub-project Direct Textile Placement (DTP). It involves handling very large cuttings that generally have to be picked up by an effector, stored, transported, draped and positioned.

The Challenge of Laying Up in Multi Curved Moulds

Defects often occur when laying up technical textiles into a multi curved rotor blade mould as a result of transposing two-dimensional into three-dimensional geometry (draping). For example, folds in the textile may arise where curvature progression is extreme, leading to unwanted fibre displacement during evacuation in the vacuum infusion process. The resulting shear loads can also cause the fibre layers to shift in relation to each other, meaning that so-called looping occurs during draping, Figure 2. Such defects must be avoided, as the flow of resin in the subsequent infusion process is impaired at these points, for example by inclusions of air, and this lowers the quality of the component. Furthermore, the deflection of the fibres (undulation) curtails the benefit of high-tensile strength.
Figure 2

Wrinkles produced after faulty direct placement of technical textiles (left) and loops formed during a draping test with a standardised sample size (right) (© BIK | Institut für integrierte Produktentwicklung)

Automated Direct Laying Up in the DTP Process

The DTP process is being developed for the continuous direct laying up of technical textiles in multi curved rotor blade moulds, Figure 3. A special effector is operated on a gantry robot, allowing all process steps to be performed in a fully automated manner. In an upstream process step, a pre-assembled cutting is produced in the machining area of the gantry robot that is then passed to the effector, which winds it up and stores it. The continuous direct textile placement is then effected from the starting position at the tip to the root ring of the rotor blade. During this process, the previously wound cutting with a maximum width of 1270 mm is unwound and precisely draped, positioned and laid up in the mould. After this process step finishes, a further cutting can be wound up and then laid up. Laying up longitudinally rather than transversally brings a time-saving benefit, as fewer cuttings need to be wound up, and thus fewer cuttings produced. In addition, the problem of longitudinally separated rovings is avoided.
Figure 3

Diagramme of the DTP process for continuous direct textile placement (© BIK | Institut für integrierte Produktentwicklung)

The layout of the DTP effector, Figure 4, consists of the support structure that accommodates the sub-assemblies of material storage and draping unit as well as additional components for monitoring and controlling actuators and sensors. Owing to the very long cuttings used, the material storage is implemented as a cylindrical winding core and forms the central element on which the cuttings can be wound up and stored. The draping unit, consisting of seven draping modules, is critical for laying up dry technical textiles wrinkle-free and exactly in position. An integrated textile gripper is used to feed the cutting from a defined edge of the guide plate located on the winding core towards the surface of the rotor blade mould, where it is specially adapted to the curved geometry (draped). The precise laying up and draping of the material is performed by the draping modules.
Figure 4

Structure of the DTP effector (representation in simplified rotor blade mould) (© BIK | Institut für integrierte Produktentwicklung)

Material Storage

Owing to the relatively large dimensions of the cuttings being handled, it is not possible to use the so-called pick-and-place method for picking and positioning the large pieces of textile. Instead, it is necessary to have the effector stores the cutting, which can be as long as the maximum length of the rotor blade. In this situation, it makes sense to make use of a storage mechanism to wind up the cuttings; this can be implemented as a cylindrical reel.

Geometric restrictions resulting from the rotor blade mould used in the sub-project, which has a radius of 950 mm at the root ring, led to an initial draft design of the material storage with an internal drive, Figure 5 (above). This has a beneficial effect on the width of the effector, allowing it to be minimised. However, the diameters of the cylindrical reel and other components increase.
Figure 5

Draping module concepts shown with its possible kinematics using two CFRP springs (concept 1, left) as well as one CFRP spring and integrated gripping technology (concept 2, right) (© BIK | Institut für integrierte Produktentwicklung)

A second draft design shows two safety chucks that can accommodate a winding shaft from above, allowing standard cardboard cores to slide over the winding shaft and be clamped pneumatically, Figure 5 (below). The installation space of the material storage decreases, owing to the smaller diameter of the winding shaft, while the length of the cutting remains the same. This results in the centre of mass shifting towards the robot’s tool centre point. This is also helped by the absence of the internal drive. Theoretically, the drive can be located anywhere within the load-bearing structure, meaning that the resulting bending moment on the structure of the effector can be reduced by locating the motor in a favourable position.

Draping Unit

The multi curved rotor blade mould and the defined textile width mean that it is necessary to configure the draping unit very flexibly. For this reason, the draping unit has been split into seven structurally identical draping modules that have one rotational and one translational degree of freedom.

In an initial draft design, a double-acting pneumatic cylinder surrounded on both sides with CFRP springs connects the draping module pick-up with the draping tip, Figure 6 (left). The rotatory positioning of the draping modules is controlled by stepping motors. The use of ball joints between the pneumatic gripper cylinder and draping tip allows the draping tip to be adjusted adaptively to match the surface geometry of the rotor blade mould. During laying up and draping, the pneumatic cylinder can spring back, with the CFRP springs providing additional counterforce. This means that the draping force can be varied by means of the stepping motors and the material thickness of the CFRP strips. The disadvantage of this draft design is the small draping surface that results during draping, owing to the CFRP springs located on each side.
Figure 6

Material storage concepts with internal motor (concept 1, above) and external motor (concept 2, below) (© BIK | Institut für integrierte Produktentwicklung)

In a second draft design, the draping module is modified such that it only has one CRFP spring, thus increasing flexibility of the draping tip, Figure 6 (right). The gripping technology is integrated into the draping tip, and the draping surface enlarged. Tests show that the second draft design allows greater pressure to be generated at the rotor blade mould.

Both draft designs call for lightweight materials, such as aluminium and carbon-fibre reinforced plastics.

The Support Structure

Since the initial focus is on developing the material storage and the draping unit as well as performing tests to prove the concepts, the support structure for the first working model is designed completely from aluminium system profiles, Figure 8. The working model allows the process steps for direct textile placement to be successfully developed and tested.
Figure 8

Working model of the handling effector during direct textile placement in a segment of the rotor blade mould in the BIK Handling Technology Competence Centre (© BIK | Institut für integrierte Produktentwicklung)

Finite Element (FE) topology optimisation is used in an initial draft design to develop an attached ideal truss structure in order to increase stiffness. Here, a distinction is made between static and dynamic load cases, working in three spatial directions. The strain in the static load cases results from weight that remains unchanged over time. Additional inertial forces occur in the dynamic load cases as a result of acceleration and braking, which reach their maximum value in the event of an emergency stop.

The acting forces must be defined at the beginning of topology optimisation and the installation space selected for the design of the attached structure. The components used and their possible ranges of movement may not be ignored when defining the installation space available. This applies, for example, to the swapping cycle of the material storage or the rotatory positioning of the draping modules. The aim of topology optimisation is the creation of a structure that is as stiff as possible given the specified volume restriction (percentage of the defined installation space). With regard to the elaborate calculation, the topology is determined for the first load case where only the weight is considered. The topology serves as the basis for developing a bar structure in which high specific stiffness is to be achieved together with very low weight through the use of CFRP-UD tubes. In order to use the optimum stiffness properties of the CFRP tubes, the bar structure is then transformed into a truss structure so that the CFRP tubes are only subjected to tensile and compressive forces. This involves relocating nodes of bars and adding new bars. The calculation of the maximum values for static and dynamic deflection, of the equivalent tensile stress and of the natural frequencies is then performed for all load cases.

At around 21 %, rotor blade production accounts for a large proportion of the manufacturing costs of a wind turbine.

Both draft designs call for lightweight materials such as aluminium and carbon-fibre reinforced plastics.

Converting the calculated truss concepts into a producible design is achieved by using CFRP tubes, aluminium profile connectors and aluminium gussets, Figure 7. The profile connectors are executed as swivel parts with a cross hole, glued on one side to the CFRP tubes, and on the other side, attached to the gusset by an articulated joint. The potential of lightweight design in the support structure is not fully exploited in the case of the attached truss structure. Increasing stiffness and lower maximum deflection comes at the cost of additional mass of the attached structure. For this reason, topology optimisation is recalculated taking into account the planned effector components, which, viewed independently of the aluminium system components, make full use of the potential of lightweight design. This concept combines elements of the truss and bar structure. Tensile-compressive tubes on articulated joints are chiefly used; however, the longitudinal and transverse CFRP tubes of the lower layer are executed as a bar structure and are able to absorb torsional moment.
Figure 7

Concepts of attached truss structure (above) as well as a combined truss and bar structure (below) (© BIK | Institut für integrierte Produktentwicklung)

Use of the DTP Effector in the Mould Segment

The process steps for laying up dry technical textiles can be demonstrated with the help of the working model of the handling effector in the transition section of a rotor blade mould. The concepts of the material storage with internal motor, the draping modules with one CFRP spring and integrated gripping technology as well as a support structure made from aluminium system profiles are implemented, Figure 8.

In principle, all of the shown concepts can be regarded as modules that can be combined and used for the automated direct textile placement. Owing to the higher masses involved with the first material storage concept and the resulting deflection of the support structure, preference is given to the second material storage concept with external motor. Both material storages can be separated from the effector and loaded independently in order to shorten effector downtimes by using several material storages. The attached truss structure achieves further stiffness in the support structure and is also easy to install. The second draping module concept facilitates a larger draping area and a more precise draping adjustment and is deployed for these reasons. The effector components described are currently being put together, Figure 9.
Figure 9

Fully developed handling effector during direct textile placement in a segment of the rotor blade mould in the BIK Handling Technology Competence Centre (© BIK | Institut für integrierte Produktentwicklung)

Both material storages can be separated from the effector and loaded discretely.

Conclusion and Outlook

The handling effector developed as part of the DTP sub-project allows a further step to be taken towards the automated laying up of technical textiles for the production of wind turbine rotor blades. Through the use of various gripping technologies, it is possible for the effector to handle different technical textiles, including those that are unsuitable for conventional manual process steps. This supports material development with regard to the draping properties of technical textiles and therefore with regard to optimising the geometry of the mould. Besides use in the construction of rotor blades, it is conceivable that the DTP handling effector might be used for the automated production of components with similarly large dimensions, such as aircraft wings.

From an automation engineering perspective, many of the project’s goals have already been met. Assembly of the various components and a verification of the process steps will be made by the end of the project. In a final step, the process for continuous laying up will be shown in the “BladeMaker” demonstration centre in Bremerhaven, and all the project partners will demonstrate the entire process chain for the automated production of a rotor blade.



We would like to thank the Federal Ministry for Economic Affairs and Energy, which is funding the “BladeMaker” joint research project based on a resolution of the Bundestag, as well as project lead partner Jülich, for coordinating the scheme.


  1. [1]
    Hau, E.: Windkraftanlagen: Grundlagen, Technik, Einsatz, Wirtschaftlichkeit. 5., neu bearb. Auflage. Berlin: Springer Vieweg, 2014Google Scholar
  2. [2]
    Müller, D. H.; Weigel, L.: Gemeinsamer Technischer Abschlussbericht für das Verbundprojekt preblade / Bremer Institut für Konstruktionstechnik, Abeking & Rasmussen Rotec GmbH Co. KG. 2008. Research reportGoogle Scholar
  3. [3]
    Ohlendorf, J.-H.; Rolbiecki, M.; Schmohl, T.; Franke, J.; Ischtschuk, L.: mapretec — ein Verfahren zur preform-Herstellung durch ebene Ablage für ein räumliches Bauteil als Basis einer automatisierten Prozesskette zur Rotorblattfertigung (0329926C/F) / Institut für integrierte Produktentwicklung. 2015. Research reportGoogle Scholar
  4. [4]
    Rolbiecki, M.; Worthmann, F.; Maas, R.; Müller, D. H.; Thoben, K.-D.: Large Area Robot — Flexibel automatisierte Produktion großflächiger Faserverbundstrukturen durch mobile Fertigungseinrichtung (02PK2052) / Institut für integrierte Produktentwicklung. 2009. Research reportGoogle Scholar
  5. [5]
    Ohlendorf, J.-H.: Untersuchung der mehrlagigen Umformung von Fasergelegen zur Herstellung von Faserverbundstrukturen, University of Bremen, 2013Google Scholar
  6. [6]
    IWES: BladeMaker: Projektseite. Fraunhofer IWES, Bremerhaven (2017). Online:, access: 29 June 2017

Copyright information

© Springer Fachmedien Wiesbaden 2017

Authors and Affiliations

  • Marvin Richrath
    • 1
  • Jan Franke
    • 1
  • Jan-Hendrik Ohlendorf
    • 1
  • Klaus-Dieter Thoben
    • 1
  1. 1.University of BremenGermany

Personalised recommendations