Community Research and Development Information Service - CORDIS

Final Report Summary - PUL-AERO (High quality curved aerospace composites using pultrusion manufacturing)

Executive Summary:
The first activities of the project focused on the definition of key specifications of the technology development and in particular the design of stringer, the selection of resin and fabric for the stringer, the modelling and simulation requirements, the design of the pultrusion line and the relevant processing windows and finally the requirements on the resin injection system and the monitoring and Quality Assessment (QA) systems.
In the modelling and simulation section, the pre-former (linear profile - Stage 1) modelling of the pultrusion process was developed. The model solution was carried in a commercial simulation platform where all relevant sub-models were coded. The energy balance at Stage 1 and Stage 2 (curved profile) of the pultrusion line has been performed. This model has been coupled with the pre-former model to account for the effects of advection and heat transfer. Material models were incorporated through subroutines. Execution of both models demonstrates the effectiveness of the strategy and verified their operation.
The modelling of the distortion in the pultrusion process was performed and has been used in the design of process for curved stringers. The steady state version of this problem, which is relevant for both stages in the pultrusion dies, required only weak coupling with the solution of the heat transfer/cure model.
The material characterisation campaign in the project started after the selection of the resin system for the pultrusion process. It included cure kinetics model, chemorheology model, thermal conductivity/CTE predictions, mechanical properties characterisation and permeability measurements.
During the pultrusion line implementation activities, the manufacturing of the various parts (die, manifold, post-former) for the construction of the line was made.
The resin injection system was constructed and installed. The injection system includes, resin and hardener tanks, process control box, pumps and mixing head, heated connection tubes and pressure sensor. It has been designed to control direct injection to the manifold at very low flow rate. In the injection system a viscosity monitoring system was built based on dielectric sensors positioned in the manifold. The resin viscosity level was checked for successful wetting of fibres.
The on-line QA system was constructed and installed. The system includes ultrasound sensors and corresponding motors for scanning the flange and the web of the stringer, water bath, electronic hardware for the measurement and the motion control, encoder for distance measurement and software for data acquisition, C-scan plotting, defect detection and evaluation of material quality.
The pultrusion line was assembled and prepared for operation. The line operation was performed for the production of a first set of profiles (where the line components were checked), the production of straight stage 1 stringers with variable conditions on the line (to select the optimal conditions) and the production of serial stage 1 profiles for feeding to stage 2 (at Exel) and stage 2’ (at IAI).
In stage 2 processing of stringers, deformed profiles were produced indicating that optimal tooling is required for the production of aerospace grade curved stringers. In stage 2’ processing, the stringers were co-bonded to panels through LRI or prepreg route. Mechanical testing of the semi-cured stringers and the co-bonded structures did not produced satisfactory results, however the relevant standards and methods were adapted to the industry requirements.
The on-line NDI system was calibrated by NDT standard reference profiles produced by Exel (pultrusion) and IAI (LRI). The improvements on the detection algorithm resulted in the satisfactory detection of defects similar to a standard NDI system. The production of serial stringers showed the robustness of the pultrusion line, however optimisation of the release agent involved in the process is necessary for the manufacturing of acceptable quality stringers.
Project Context and Objectives:
Pultrusion is a continuous manufacturing process for composite parts. Pultrusion is the process of pulling dry fibres (from roving bobbins) or/and fabrics (from rolls), impregnated with resin, through a series of heated dies. The resin solidifies in the dies and the final part is formed. The final part exits the dies and a cut off saw cuts the continuous profile to appropriate lengths for transportation and application. The whole process is driven by reciprocating pullers, which define the speed of the line.
Traditionally, pultrusion is used for production of linear, non-aerospace grade composite parts. Typical pultrusion products include glass, carbon or other fibre polyester and vinylester profiles. Typical industries and applications include: the construction industry, chemical resistant floors, construction of mobile towers, stiffeners for industrial equipment (such as tanks), cable trays and other. A wide range of shapes of profiles can be manufactured with pultrusion. The ease of the process and the flexibility offered by adding different chemicals and colour pigments in the resin bath make pultrusion an ideal process for low-cost straight profiles.
The main benefits of the pultrusion process are the possibility for continuous production, the repeatability of the process and the low production costs (calculated as cost per meter of composite produced).
The main disadvantages of the pultrusion process is the difficulty in producing curved parts and the difficulty in handling aerospace grade epoxy resins, mainly due to the relevant stringent requirements regarding exposure to elevated temperatures and their slow cure (which results in slow pultrusion line speed).
PUL-AERO aims to resolve the technical problems of pultrusion when it comes to production of curved aerospace composites parts and take advantage of the benefits of the process.
PUL-AERO will design and build a pultrusion line for the production of linear and curved composite stringers for the aerospace industry. The pultrusion line innovations will be the following:
• The separation of the process in two stages. In stage 1, linear heated dies will cure the resin partially, in order to obtain dimensional stability. Then, in stage 2, a post forming die will induce the final curvature of the profile and fully cure the resin. The benefit of such configuration is that profiles of different curvatures can be produced without having to re-set the whole of the pultrusion line. Only the post former die needs to be replaced. This is a great advantage for the aerospace industry where it is usually the case that a stringer with the same profile is needed at different sections of the aircraft with different curvature characteristics.
• The development of an alternative pultrusion process route involving production of a semi-solid intermediate product. In this approach the product of pultrusion is in the state corresponding to stage 1 of the integrated process. The material is then partially cured so that it forms a glass at ambient temperature allowing handling, transport and integration in processing assemblies. In a second stage applicable to a Liquid Composite Moulding (LCM) scenario (stage 2’), the semi-cured pultruded stringer will be integrated with the dry reinforcement of the skin. Upon the application of heat for the skin injection the material of the stringer would become a rubber with the capability for creep flow. This will lead to spontaneous keying of the stringer with the reinforcement through the formation of effective nails generated by the slowly flowing matrix of the stringer. In this way, the need for joining stringers with the skin through adhesives is rendered obsolete.
• The design, construction and operation of a resin injection system. The system will inject resin onto the fibres and fabrics in a manifold just before the entrance of the first heated die. Sensors in the injection manifold will monitor the resin temperature and viscosity. The resin excess can flow out of the manifold in a small quantity as the injection flow rate will match the resin consumption in the process.
• The integration of a process monitoring and QA system. The system will collect information from all sensors from the line and from the QA systems installed. Single software will provide the user with information about the pultrusion line operation. The output of the system will be a complete signature of the process, which can accompany the produced part to the customer. This is very crucial for the aerospace industry, as process certification and process standardisation issues are very important for the acceptance of any new manufacturing process.

The main goal of PUL-AERO is the development of a material state based controlled pultrusion process for the manufacturing of curved and partially cured stringers that comply with the stringent specifications of the aerospace industry and offer significant savings in production costs.
PUL-AERO will create production processes and technologies with a focus on improved cost efficiency while taking into account the environmental related aspects. The pultrusion line will be split in two stages, adding flexibility to the manufacturing of stringers with different curvatures and improving the cost effectiveness of the production. The whole process will be driven and monitored by a process monitoring and QA system that will provide real time information and control of the line.

The main objectives of the PUL-AERO project are:
• The accurate modelling of the thermoset resin properties in order to design the pultrusion process based on the values of material properties.
• The prescription of the temperature profiles in the resin injection system, stage 1 dies and stage 2 tools (die or LRI tool) based on modelling and simulation of the line. Calculation of the fibres undulation in the curved stringer and provide technical solutions for minimising the effect on the stringer properties.
• The correlation of the thermoset resin viscosity to dielectric measurements in order to monitor the resin condition in the resin injection system.
• The development of special dies and pulling mechanisms with curved contours in order to curve the pultruded profiles according to specifications.
• The development or adaptation of methods for in situ Quality Assessment in the pultrusion line.
• The design and implementation of a pultrusion process that allows resin injection with a benefit of approximately 60% reduction in resin wastage.
• The development and implementation of a pultrusion process that allows the continuous production of curved stringers of aerospace quality with a cost benefit of 20% in comparison to current routes for the incorporation of stiffeners.
• The quantification of the keying opportunities offered by the integration of semi-cured stringers in a liquid moulding assembly and the associated mechanical performance of the interface with a cost benefit of 30% associated with the elimination of the necessity for bonding.
Project Results:
1. Process Design
The first activities of the project focused on the definition of key specifications of the technology development and in particular the design of stringer, the selection of resin and fabric for the stringer, the modelling and simulation requirements, the design of the pultrusion line and the relevant processing windows and finally the requirements on the resin injection system and the monitoring and Quality Assessment (QA) systems.

More specifically:
The profile of the stringers has been defined and shown in Figure 1.
It was agreed that both constant curvature and joggled configurations for pultrusion lines will be evaluated, and therefore both configurations are presented.
To minimize capital investment on tooling for the PUL-AERO project it was agreed to utilize existing tooling for production in this project. The stiffener section that was chosen is identical to that manufactured for the G280 horizontal stabilizer by Exel, and will be manufactured with the existing pultrusion die.
The stiffeners longitudinal geometry and length was chosen to fit with existing tooling that was used for previous R&D projects. Constant radius geometry of 933mm is of similar size to a small business jet fuselage (for comparison, the Gulfstream-IAI G200 fuselage outer diameter is about 2280mm (~90in)), but the curved pultrusion technology developed could be used easily on larger aircraft as well. The joggled end is a typical design detail, and therefore the technology developed for this configuration should be versatile so it could be implemented for different lengths and thicknesses of the joggle.

The design of the new pultrusion process developed in the project requires the development and implementation of modelling and monitoring tools. These will enable prediction and verification of the evolution of state variables such as the temperature, degree of cure and resin flow speed to be made. The simulation capability will comprise predictive models of the pre-forming, curing and post-forming stages of the process. The pre-forming simulation requires solution of the flow through porous media problem. Curing simulation corresponds to solution of the heat conduction problem in combination with cure kinetics and simulation of post-forming requires solution of the stress problem. The different physics involved are coupled thought the dependence of material properties on state variables associated with two or more of the models used. This coupling results in the need for development of 12 constitutive models, involving the determination of about 65 parameters. In addition, a set of 14 boundary conditions need to be defined. The overall simulation chain will be able to predict the pressure and resin flow speed fields in the pre-former, the evolution of temperature and degree of cure along the whole process, the process stress and associated deformation developed in the post-forming stage and the distortion occurring upon release of the component form the die/tool. The monitoring system will comprise temperature and dielectric sensors accompanied by the corresponding acquisition systems. The sensors will be capable of operating with the temperature, pressure conditions of the process and the interrogation systems will be able to provide on-line information on the state variables (temperature, material state). Thermal monitoring will address all stages of the process (resin bath, per-former, curing dies, post former), whilst dielectric monitoring will focus on the rheological development of the material in the resin bath and the pre-forming zone.


The selection of the resin matrix system and the fibres is reported in this deliverable. A number of candidate resin system, suitable for infusion have been shortlisted. The selection was performed according to the specifications set by the End Users.
The candidate materials were ranked according to three properties, defined by the End Users:
• Glass transition temperature at wet conditions higher than 140°C
• Elongation as high as possible
• Fracture toughness (G1c) as high as possible
The XU3508/XB3473 has been selected.


The methodology for the design of the cure process for the pultrusion line is presented. The design is split into the three stages outlined in the project:
• Stage 1: cure the material up to a specific point that will allow mechanical and dimensional integrity. It will also allow safe transportation (for the case of stage 2’)
• Stage 2: Post cure the material after curvature is applied in the pultrusion line
• Stage 2’: Transport the partially cured composite to the end user site and co-cure with a composite skin.

The main design parameter for the determination of the cure cycle is the glass transition temperature. The cure cycle is split to 4 phases:
• Infusion: This phase is not present in the pultrusion line. The reinforcements will be fully saturated either by passing through a resin bath or by passing through an infusion manifold located before the first die.
• Curing (see Figure 2): Curing during stage 1 is arrested at a specific degree of cure that will ensure dimensional integrity for the material and safe transportation. Curing will be continued in stage 2. In the case of stage 2’ (co-curing) the full cure cycle of the material will be followed
• Post curing: Post curing will be effected in stage 2 with the use of post former die(s).
• Cooling: Controlled cooling in all stages will be needed in order to control the buildup of stresses due to shrinkage and abrupt temperature variation


The design of the pultrsion line for the production of curved aerospace grade stringers is reported in this section of the deliverable. All the components of the new line have been designed and presented. The key components of the line are:
• resin injection system (see Figure 3)
• Straight pultrusion dies
• Line pulling system
• Curved pultrusion dies
• On-line monitoring and quality control system


The design of the process monitoring system (sensors and thermocouples at the resin injection system) and the design of the quality assessment system (ultrasonic sensors and devices at the pultruded profiles) has been performed. The utility of these systems is very significant for the reliable operation of the pultrusion process for the production of curved aerospace stringers.


2. Modelling and simulation
In the modelling and simulation section, the pre-former (linear profile - Stage 1) modelling of the pultrusion process was developed. The model solution was carried in a commercial simulation platform where all relevant sub-models were coded. The energy balance at Stage 1 and Stage 2 (curved profile) of the pultrusion line has been performed. This model has been coupled with the pre-former model to account for the effects of advection and heat transfer. Material models were incorporated through subroutines. Execution of both models demonstrates the effectiveness of the strategy and verified their operation.
The modelling of the distortion in the pultrusion process was performed and has been used in the design of process for curved stringers. The steady state version of this problem, which is relevant for both stages in the pultrusion dies, required only weak coupling with the solution of the heat transfer/cure model.
The material characterisation campaign in the project started after the selection of the resin system for the pultrusion process. It included cure kinetics model, chemorheology model, thermal conductivity/CTE predictions, mechanical properties characterisation and permeability measurements.

More specifically:

The energy balance at the stages 1 and 2 of the pultrusion line, has been made using Fourier’s law coupled with a model for chemical kinetics and exothermic heat. This model was coupled with the pre-former model to account for the effects of advection on heat transfer. The solution was carried out using the commercial solver Marc®. The thermal boundary conditions were incorporated through a convection condition, with parameters depending on the current location of the nodal points. The pulling is represented by a fixed displacement boundary condition through a user defined subroutine. The interaction with the die was taken into account using a contact condition of the die, which was considered isothermal and fully fixed. Material models were incorporated through user subroutines. The boundary conditions of the thermal problem were defined. The geometry and lay up considered for testing the model at this stage correspond to the final component. Execution of the model (see Figure 5) demonstrated the effectiveness of the strategy and verified its operation. The results of this model will be further used in the pre-former model for the incorporation of material state and temperature effects on flow phenomena, in the simulation of the pultrusion line for the accurate prediction of material state at the end of stages 1 and 2 of the process, and in the coupling with the distortion model to allow the definition of the dependence of mechanical properties on temperature and material state.
This thermo-mechanical problem was addressed with a standard coupled solution in the finite element model. The steady state version of the problem, which is relevant to stage 1 and stage 2 in the pultrusion die, required only weak coupling with the solution of the heat transfer/cure model of the previous sub-task. A transient version of the problem was established for the stringer-skin co-curing process. The non-linear sub-models of composite moduli dependence on material state was implemented using look up tables. The envelope of conditions relevant to the co-curing problem was defined.
The viscous flow of the semi-cured resin when it is in contact with a dry textile under pressure was simulated using a viscoplastic material model implemented in a subroutine. The condition envelope relevant to this problem and the benchmark requirements related to resin penetration were provided.


A specific study was made to analyse the interfacial properties for a process in which partially cured stiffeners are co-cured with infused skins. This process option can deliver efficient integration of pultruded stiffeners in aerospace components. The investigation was carried for two different material systems: a unidirectional carbon reinforcement with a standard epoxy matrix and the material combination of the project involving a braid. Mode I delamination testing showed that the fracture toughness of the interface is a decreasing function of the degree of cure of the pre-cured sublaminates. There is a critical level of cure over which the reduction becomes steep; this level is at about 50% for the standard system (see Figure 7, left) and at about 70% for the system of the project. The reduction is far more significant for the system of the project (80%) compared to the standard system (35%). The critical point in pre-cure is linked with the point at which the fractures follows the interface formed by the surface of the pre-cure processing (see Figure 7, right). Therefore, when fracture across the primary surface occurs, the fracture toughness is sensitive to the type of peel ply used in precure. This is also in agreement with the stronger effect of the transition for the system of the project involving a braid, which results in a higher initial fracture toughness and a steeper loss when the fracture process moves to the pre-cured surface. Overall, the results showed that use of pre-formed stiffeners at procure levels lower than 70% for the system of the project allows subsequent integration in the skin infusion assembly and co-curing without significant loss of interfacial performance.


The material characterisation was performed for the selected materials. The cure kinetics characterisation relied in the execution of the plan for the DSC tests on the selected resin system to derive degree of cure and cure rate of the resin during the cure process. Viscosity measurements were executed for the plan for the rheology tests on the selected resin system to derive the chemorheological model of the resin system. The thermal conductivity measurements were made on the selected resin system to derive thermal conductivity parameters of the resin during the cure process. CTE measurements were made according to a plan for the TMA tests on the selected resin system to derive thermomechanical Tg and thermal expansion coefficients of the resin during the cure process (see Figure 8). Tg and cp measurements were made according to the plan for the DSC tests on the selected resin system. The Tg model is shown in Figure 6. The mechanical properties characterisation was based on a plan of mechanical tests on the selected resin system to derive modulus and Poisson’s ration of the resin during the cure process.


The model for the simulation of the pre-former (Stage 1 of the pultrusion process) was developed. The model solves by means of FEA solver PAMRTM® the coupled flow through porous media, heat transfer and chemical problem. Constitutive material models have been implemented through user defined tables. Moving pressure to simulate the pultrusion injection system and moving temperature boundary conditions have been applied. Initial degree of cure is set equal to 2%. The simulation successfully deals with the problem under study resulting in the prediction of degree of cure distribution at the exit of the die which will be the initial degree of cure of the simulation for Stage 2. The results simulations (see Figure 9) are compared with glass transition temperature measurements at the end of Stage I showing good agreement and validating the modelling strategy and material models utilised. The simulation shows that for the geometry considered in the project the through thickness degree of cure gradient at Stage 1 die exit is negligible. The final degree for a cure profile where the two heaters are kept at 160 ⁰C and pulling speed is 5 cm/min is 67%.

The simulation of the Stage 2 post former of the pultrusion process was made. The model includes solutions of the heat transfer and stress development process. Constitutive material models involving thermal, chemical, mechanical and thermos-mechanical properties have been implemented via user-subroutines. Boundary conditions have been applied also through user subroutines to represent the thermal profile undergone by the part and the pulling force. An initial temperature condition has been applied using a transient thermal load case where cure was not active and an initial degree of cure of 67% was applied as obtained from the simulation of Stage 1 of the process. Different simulation considering different thermal profile and pulling speed have been considered. The simulation developed constitutes a powerful tool at the design stage of the process able to predict outcome of the process in terms of stresses developed and final degree of cure. The simulation allows the estimation of process stress during curved pultrusion (see Figure 10), showing that the level of stress reached is not problematic for feasibility of the process. The simulation also shows that the process is feasible at relatively high pultrusion speeds, identifying an efficiency opportunity in process design.



3. Pultrusion line implementation
During the pultrusion line implementation activities, the manufacturing of the various parts (die, manifold, post-former) for the construction of the line was made.
The resin injection system was constructed and installed. The injection system includes, resin and hardener tanks, process control box, pumps and mixing head, heated connection tubes and pressure sensor. It has been designed to control direct injection to the manifold at very low flow rate. In the injection system a viscosity monitoring system was built based on dielectric sensors positioned in the manifold. The resin viscosity level was checked for successful wetting of fibres.
The on-line QA system was constructed and installed. The system includes ultrasound sensors and corresponding motors for scanning the flange and the web of the stringer, water bath, electronic hardware for the measurement and the motion control, encoder for distance measurement and software for data acquisition, C-scan plotting, defect detection and evaluation of material quality.


More specifically:
Key pultrusion line components were prepared and adapted at the pultrusion site. The manifold was designed and manufactured (see Figure 11). The pre-former, the main die, the pre-former and the line puller were adapted for the line, so that the line set-up can take place at the next step (see Figure 18).

The construction of a new resin injection system was performed. The resin injection system constructed meets the requirements set earlier in the project. The basic steps of the injection system operation were followed. The resin injection software (see Figure 14) has been made to be flexible enough to work with single-component resin systems as well as with two-component resin systems. Also it is compatible with epoxy, vinylester resins as well as polyurethanes. The integration of the injection system with monitoring sensors (see Figure 12) has been performed to allow the measurement of resin viscosity at all stages of the resin advancement to the die.
The injection system alongside the sensors for viscosity measurements has been installed on the pultrusion line (see Figure 13), and the first trials were performed. The installation run (see Figure 19) involved the selected resin system, constant temperature heating zones in the die, representative line speed and flow rate, pressure control as injection strategy and mild pre-heating of resin to reduce resin viscosity. The line operation was stable, all systems performed well and semi-cured profiles were produced.

The construction of a new viscosity monitoring and a novel on-line NDI system was performed. Both constructed system meet the requirements set earlier in the project. The basic steps of the systems operation were documented.
The viscosity monitoring system consists of dielectric sensors (see Figure 15) adapted to the manifold, electronic hardware (housed within the injection system, see Figure 12) and monitoring software (see Figure 20) presenting the viscosity data in real-time to the injection system.
The integration of the monitoring system to the injection system has been performed to allow the measurement of resin viscosity at all stages of the resin advancement to the die. The viscosity monitoring system was installed on the pultrusion line.
The injection system alongside the sensors for viscosity measurements were tested and found fit to be used in the production of controlled cure profiles.

The on-line NDT system relies on ultrasound sensors (see Figure 16) and has been made to be custom for the particular design of profile, although the measurement principle is generic for composite structures with mixed type of fabric (UD and triaxial). The adaptation of the on-line NDT system to a moving profile has been made to allow the evaluation of quality of parts in real time during production. Dedicated software has been developed for the operation of the system (see Figure 17). This QA system was installation on the pultrusion line. The installation trial run involved the standard T-section profile, standard fabric architecture, 2-zone die heating, representative line speed and two line pullers. The operation of the on-line NDT system was stable (see Figure 21) as every component of the system performed well and the quality of the produced profiles was monitored (see Figure 22).

The line was run by the factory’s operators using the newly installed equipment (manifold, resin injection system and on-line NDT system) for the production of first set of profiles (see Figure 23). During the line running, the conditions were kept constant in order to check the consistency of the production and the stability of the resin injection system (see Figure 24). 4 metres of profile length were produced and the thermal analysis testing showed the good level of consistency and the potential to produce undercured (stage 1) profiles for further processing towards the development of curved stringers. The quality of the first parts was visually acceptable, although there was potential for optimisation of the line in the next step of the project.


4. Process Monitoring
The pultrusion line was assembled and prepared for operation. The line operation was performed for the production of a first set of profiles (where the line components were checked), the production of straight stage 1 stringers with variable conditions on the line (to select the optimal conditions) and the production of serial stage 1 profiles for feeding to stage 2 (at Exel) and stage 2’ (at IAI).
In stage 2 processing of stringers, deformed profiles were produced indicating that optimal tooling is required for the production of aerospace grade curved stringers. In stage 2’ processing, the stringers were co-bonded to panels through LRI or prepreg route. Mechanical testing of the semi-cured stringers and the co-bonded structures did not produced satisfactory results, however the relevant standards and methods were adapted to the industry requirements.

More specifically:
During the forming trials of curved stiffeners (see Figure 25), it was found that formability is a critical issue for stage 1 profiles. The work suggested that formability is possible, but is limited by detail part geometry. The long beam sections formed under vacuum bag on a curved tool demonstrated a curved base but failed blade after unloading (see Figure 25). The trials using four point bending also produced blade failure, but no significant permanent base deformation. The suspicion is that the inherent geometrical stiffness of the blade is too resistant to in plane deformation to enable formability in this direction, despite the incomplete curing of the material and forming temperatures above the as-received Tg.

The resin injection system was operated at the pultrusion line and first runs were performed. For the appropriate operation of the injection system various tests were conducted on site and in the laboratory. The tests resulted in the calculation of the optimum volumetric ratio for the components of the thermoset matrix of the produced profiles and in the estimation of optimum flow rate and hydraulic pressure for the resin injection to the manifold (see Figure 26). Furthermore, the mechanical and thermal settings of the resin injection system were determined for continuous (uninterrupted) operation of the system in production and the resin temperature was tuned as it affected the resin viscosity during injection for proper wetting of fibers in the manifold.

The on-line NDI system has been one of the most significant developments of the current project. The final configuration of the sensing system was installed (see Figure 27). This QA system at the pultrusion line was operated during manufacturing of standard production profiles and new resin system profiles. The performance of the sensors was tuned (see Figure 28) and the detection sensitivity was verified through scanning NDT standard reference profiles (see Figure 29) made at EXEL (pultrusion) and IAI (LRI). The QA system’s sensitivity was calibrated to the defects of the standard reference profiles. A new analysis routine was developed for the accurate detection and dimensioning of the defects. This routine was integrated to the on-line version of the software and has operated successfully on the line.
The on-line NDT system has demonstrated similar sensitivity and performance to conventional ultrasonic NDT inspection methods (se Figure 30). This is a major achievement for the PUL-AERO programme which should be continued to the development of a practical commercial system. Off-line NDT is time consuming and expensive, and cannot help to correct process discrepancies in real time. The value of this development cannot be under-emphasized.
The stringers produced with the new resin system showed high attenuation (see Figure 32) which masks the actual quality of the profile. The industrial significance of the QA system has been adequately demonstrated.

As a long time user of pultrusion during serial aircraft component production, the primary function of the project end user was to direct the consortium towards the requirements of the aerospace industry. The development of curved and/or postformable pultrusion was identified as a method for overcoming some of the drawbacks of conventional, straight, pultrusion. The project developed technology to produce semi-cured stringers as stiffeners for integrated structures. During stage 2’ of the manufacturing process, QA procedures were applied. They included measurements of temperature and vacuum during LRI and specific mechanical testing of representative samples, whereas dedicated protocol for samples preparation and handling was developed.

Quality control procedures at the produced stringers were followed at both IAI and Exel. IAI produced a NDT standard profile with intentional defects which was used to calibrate the on-line NDI system. Microscopy (see Figure 31) was performed on semi-cured and post-cured samples. Also DSC and DMA tests were employed to measure the degree of cure and the glass transition temperature of the profiles and to tune the pultrusion line for the production of stage 1 stringers.

A number of line speeds and temperatures for the dies has been tried once the process was finalised. The assessment of the quality of the produced parts in each set of conditions was used to identify the optimum process parameters (injection temperature, line speed, die heating zones temperatures) to be used for series production of stringers in the next workpackage.


5. Pultrusion line operation and assessment
The on-line NDI system was calibrated by NDT standard reference profiles produced by Exel (pultrusion) and IAI (LRI). The improvements on the detection algorithm resulted in the satisfactory detection of defects similar to a standard NDI system. The production of serial stringers showed the robustness of the pultrusion line, however optimisation of the release agent involved in the process is necessary for the manufacturing of acceptable quality stringers.

More specifically:
The series production of Stage 1 stringers was made (see Figure 33). Half of profiles were shipped to the end user for Stage 2’ processing, while the other half was kept at the pultrusion factory for Stage 2 processing (bending). Initial testing of profiles as-received by the end user showed a very high attenuation, to the extent that no internal detail could be imaged. It was suspected that this was a result of the partially cured matrix. The partially cured epoxy could be expected to have a pseudo rubbery structure and produce high ultrasonic damping. Accordingly, post cured profiles were tested and also showed the same very high overall attenuation. The cause for this is unclear, but could be a result of the toughened XU 3508 epoxy system, interaction with the internal release, or both. Infused panels using the XU 3508/XB3473 resin combination did not exhibit this behaviour and could be inspected readily using conventional NDT. The suspicion has to be that the cause of the high attenuation is the internal release system or some other peculiarity of the pultrusion process which needs further evaluation.

Following the production of the Stage 1 material, Stage 2 at the pultrusion site was undertaken through bending trials of the material using a forming jig as proposed in Figure 34. The schematic of the bending configuration is shown in Figure 35.
Several trials have been completed using the jig in various manner to see how best the profile can conform to the shape. As shown in Figure 36 (left), weights were added to try and give some additional forming applied load. In all cases whilst the profile is ‘softening’ at the predicted temperatures, when it is placed in the oven (see Figure 36, right), however, it is kinking rather than bending to the required shape (Figure 37).

For the mechanical test programme described below, manufacturing of panels is required, by both prepreg autoclave technology, and LRI using the selected resin matrix system. In both cases, panels are required with and without film adhesive applied to the bondline. Co bonding uses an adhesive interface as standard. However, in this case, since the matrix of the pultrusion is only partly cured there is the possibility of some chemical attachment between the stiffener and base panel during the panel manufacturing process. In particular, for the LRI process, since the base panel and pultrusion use the same matrix resin the possibility of chemical attachment should be enhanced.
No problems were envisaged for the prepreg/autoclave manufacturing route. However, flow trials are usually required for infusion processing that incorporates cured details. A panel was manufactured initially to check infusion parameters, as shown in Figure 38. The panel was fully wetted at all points and was satisfactory according to NDT testing. The parameters developed for this trial panel were used for panel LRI manufacture for the mechanical test programme.

The long beam sections formed under vacuum bag on a curved tool demonstrated a curved base but failed blade after unloading. The trials using four point bending also produced blade failure, but no significant permanent base deformation. The suspicion is that the inherent geometrical stiffness of the blade is too resistant to in plane deformation to enable formability in this direction, despite the incomplete curing of the material and forming temperatures above the as-received Tg.
The on-line NDT system has demonstrated similar sensitivity and performance to conventional ultrasonic NDT inspection methods. This is a major achievement for the PUL-AERO programme which should be continued to the development of a practical commercial system.
Finally, calculation of the relative proportion of shear to bending displacement in tests (see Figure 39) suggests that for a three-point bend test a support span of greater than 700mm would be required to ensure the shear contribution to bending was less than 10% of the total. The values of G and E determined in these tests can be used to evaluate the total bending behaviour of the T-section in any loading configuration.

Much interesting progress has been made, but several open issues remain necessitating further work. The practicality of pultrusion with an amine curing epoxy has been demonstrated, but the mechanical properties measured indicate the need for process optimization and/or a study of the resin /hardener/internal release system compatibility. Although the pultrusion line has developed into a new platform of automation and quality, the combination of materials affects the surface and interface properties of the profile structure. Further study in this aspect is required. This study should also identify the reasons for the failure of conventional ultrasonic NDT systems to reveal internal detail in the pultrusions.

The production costs associated with the PUL-AERO pultrusion line have been analysed. The cost of a composite stringer produced by the PUL-AERO process has been compared to the cost of stringers produced by three alternative material/fabrication routes. Taking the additional capital costs of the PUL-AERO line over a five year period, the PUL-AERO product has been calculated to be 43% cheaper than a standard pultrusion. This provides the potential opportunity for composite manufacturers such as Exel to increase their profits to in excess of 26%. Additional benefits of the PUL-AERO process have also been highlighted.


6. Exploitation and Dissemination
Within the PUL-AERO project the relation between the material properties, the process conditions and the quality / mechanical performance of profiles was established. This results in knowledge based system for the studied geometries and configurations. Such a knowledge based must be made for every profile variation since it is not universally applicable. However, the possibility that PUL-AERO technology offers in tuning rapidly and safely the production line must be taken into account. Usually lengthy procedures and several failures are encountered before pultrusion line produces acceptable quality parts. Elements for this potential are:
• type of resin and fibre alongside available data (resin kinetics, fabric properties)
• variation in raw material condition
• variation in line speed and temperature
• influence of release agent
• other relevant elements for the specific case under study.

Changing process parameters are translated to mechanical properties, product quality (as evaluated by NDI methods) and, if relevant, long term properties after chemical aging (influence of kerosene, salt, et cetera). The results could best be summarized in test matrices for ease of reading and control of completeness.
Several routes can be followed to involve pultrusion for manufacturing of aviation composites in wide range. Aviation authorities are becoming more and more convinced of the efficient use of pultrusion for manufacturing of high performance composites. This view is supported by the ever increasing number of technical personnel at leading aerospace companies who becomes familiar with pultrusion and aware of the true potential of this manufacturing method. Therefore, PUL-AERO has opened a new avenue for this industry sector and shows the way for manufacturing high quality composites at the most time and cost efficient manner.

A requirement for 148,500km of aircraft stringers over the next 20 years has been identified, leading to a potential addressable world market for stringers of over €20.8 Billion. The stringers produced by the PUL-AERO pultrusion line are showing promise as a replacement to aluminium stringers and composite prepreg stringers in addressing this market, offering benefits of weight saving, and cost reduction.

The consortium has outlined the measures they are taking to target this market and hence benefit from the results of the PUL-AERO project. Exploitation routes for both pieces of equipment comprising the PUL-AERO pultrusion line and for the stringers produced by the line are identified. Specific actions to achieve the exploitation plan are outlined in the exploitation report. The dissemination activities have been mainly confined by the intention to keep details of the simulation and the on-line NDT system undisclosed for the potential of knowledge protection.
Potential Impact:
A. Strategic Impact
The scope of the PUL-AERO project is the development of an advanced pultrusion line with capability to produce curved aerospace composites while taking full advantage of the benefits of the process. The main benefits of the pultrusion process are the possibility for continuous production, the repeatability of the process and the low production costs (calculated as cost per meter of composite produced). The proposed developments will then lead to the optimal, cost-effective and reliable processing of curved CFRPs with aerospace quality standards. The project addresses all variable manufacturing parameters (materials, process conditions, equipment, simulation models, sensors, control strategies and quality/inspection issues) for pultrusion processing of composite materials before integrating the components to a functional and qualified production line. More specifically the project output consists of:
• A novel configuration of the pultrusion processing of composites through the separation of process in two stages: one with linear heated dies for curing and one with curved post forming tools for shaping.
• An alternative pultrusion process route involving production of a semi-cured intermediate product where the semi-cured stringers will be co-bonded with the panel of the structure.
• The resin injection system for pultrusion processing with full design, construction and operation guidelines including sensors and actuators for the supply of standard quality liquid resin towards full wetting of the fibres in a manifold before they enter in the heated die.
• A functional and integrated process monitoring and QA system based on sensors positioned in the resin bath and on-line inspection system for the detection of structural faults in the processed materials.
• A complete signature of the process accompanying the produced part from the supervisory control system comprising of temperature controllers output, sensor signals from the resin, on-line inspection data and pultrusion line operation data (line speed and pull force) giving rise to reliable and eco-efficient processing of composite structures.
• The manufacturing of curved aerospace grade components in a continuous process with improved cost efficiency and known quality.

Profiles like stringers, frames and beams are one of the most important structural elements of an airframe. Used for PAX and cargo floor grid, frames, stringers in fuselage, wing and stabiliser panels, profiles of different shape, size and complexity are estimated to represent about 1/3 of an airframe structural weight. The substitution of these, today mainly aluminium components, by CFRP parts offer in general a weight saving potential of more than 20%.
It can be definitely stated, that this weight reduction can only be afforded if suitable production methods for complex CFRP components become available, which can reduce the manufacturing costs by a significant factor compared to prepreg technology. Otherwise the desired increase of CFRP parts in an airframe will fail due to the high production costs.
Technologies on composite semi- or partly-curing (including textile performing) offer in combination with respective injection and curing technologies the highest potential for an optimised design following the structural mechanical requirements. Especially for complex shapes structures weight saving potential can even be higher than that of prepreg based structures. The reason is that a textile based design, combining if necessary several textile sub-preforms in one part, offers a higher potential for optimised fibre orientations (even curved) adapted to the loads, highly integrated net-shaping and even three-dimensional fibre reinforcements.
The PUL-AERO process has shown promise in achieving its goals of: (i) a weight saving of at least 20% compared to aluminium design, (ii) a weight saving of up to 5% compared to prepreg design for curved profiles and (iii) a cost saving of more than 35% compared to prepreg design for complex profiles. Detailed weight and cost analysis is supporting the above claims.

The improved cost efficiency will be achieved mainly by four factors:
• the integrated development and process loop
• the very high degree of automation on the process
• use of advanced manufacturing methods
• integrated quality control through novel QA procedures
A shorter development cycle, based on standardised methods for design and production has shown to contribute to the cost reduction by reducing the number of iteration loops (goal: first time right) and less costly design adaptations.
A consequent utilisation and application of the project results will lead to a very significant weight saving of the whole airframe and by this to an improvement of environmental compatibility without loss of performance. At the same time an also significant reduction of the aircraft acquisition costs can be expected. The main cost cutting issues compared to prepreg are: lower basic material costs,
The know-how of the partners, the partner composition, the work programme and the exploitation philosophy allows for the readiness of the technology in due time before the year 2020. This will allow enough time for more specific developments (design, calculations, tooling, ...) in order to transfer the basic know how into series applications by 2020.

B. Socio-economic benefits
B1. Growth and competitiveness in the European aeronautical industry
The aeronautics industry operates in a highly competitive and dynamic environment. New generations of aircraft are complex, costly and require a wide range of skills, expertise and facilities. As the strengthening of the competitiveness of the aeronautics manufacturing industry is an important constituent of the sustainable growth for Europe, it is therefore indispensable for the EU aeronautical industry to be able to respond to the need for economical aircraft and to be more competitive in the world market. The competition for the EU aeronautics manufacturing industry mainly comes from the strong position of the respective US industry, and especially Boeing, and this fact gives a continuous challenge to proceed in cost effectiveness. For the business aircraft sector, the same level of competition is met by the strong position of General Dynamics (Gulfstream).
Along these lines, the aeronautic industries are steadily increasing the percentage of composite materials in the structure of the aircraft in order to reduce weight and number of parts. However, the challenge in the composite materials’ production is to keep cost as low as possible while maintaining the known high part quality. The only feasible way to achieve this goal is to utilise optimised processing methods (in terms of cost and quality) in parallel to improving the component development procedures and utilising advanced technological solutions in monitoring and control systems.
To achieve the objective of increasing the competitiveness in aerospace manufacturing industry, the developments within the PUL-AERO project have developed a novel composite materials processing practice in order to produce lighter, lower cost and higher quality (thus more safe) structural parts. This improvement will help satisfy the potential of composite manufacturing methods, which is currently limited because of a lack of automation, the use of very high capital cost equipment and the requirement for highly skilled workers. As a consequence, the development of new and cost reducing processes and practices is of great importance for the increase of competitiveness in the market. Thus the activities of the PUL-AERO research project are very much in line with the EU's air transport policy and its objectives.

B2. Impact on the formation of a competitive supply chain
Time to market can be mainly reduced due to a higher flexibility in the process and supply chain. If required the whole process can be either, integrated in one facility or also separated into impregnation/pre-curing, curing and final shaping. The first concept aims on a reduction of part transport and storage costs; the latter offers the chance to leave sensitive tasks to specialists. The actual concept can be chosen as it is technically and economically suitable.
Today Japanese suppliers are dominating the market in the field of low complex pultruded CFRP profiles. The upper floor beam of Airbus A380 is a well known example. PUL-AERO has established a technology and product line for complex curved profiles in Europe with a realistic chance to be the first on the market and to claim corresponding intellectual property. To gain control or at least secure knowledge about the use of those technologies will improve the competitive situation of the European aerospace industry. Today the knowledge about the use of such technologies by the competitor (787 black fuselage) is only based on rumours.
It can be expected that, by developing and saving this know-how for the European industry, a considerable number of high tech jobs can be secured or even newly established.
To face the challenge of reaching all the stated goals in a time frame allowing a utilisation right before 2020, a joint effort of the most experienced partners with their specific know how and facilities across Europe is combined in the project.
A national approach would not allow to make all steps forward in parallel, which are needed to meet the goals of vision 2020 and to support significantly the European position in the global competition in aerospace. Nevertheless, the achievements of national research activities will be utilized in the project by the partners coming from the different EU countries.
Finally, it should be mentioned that also other European key industries like automotive, ship building, railway or architecture will benefit from the results of the project, because affordable and high performance profiles are needed in many structural applications.

B3. Impact on reduction of aircraft components production cost
The development within PUL-AERO of an intelligent process control system and the combination of material models, process optimisation and reliability tools with the sensing system and the process equipment will allow the European aircraft industry to reduce time-consuming trials and costly experimental work, contributing to cost savings in the development phase of up to 15%. The availability of modelling, process simulation and sensing tools, such as those developed in PUL-AERO, will help to define the Process Specification in fewer and more informative trials. To this end the utilisation of material behaviour knowledge will make smoother the shifting from one component design to another. In addition the planned industrialisation of the sensing devices and the ability to measure accurately and on-line the material state of the produced components will lead to more efficient design concepts due to higher confidence in meeting safety and certification requirements and thus will contribute further to reduce the development effort.
The proposed implementation of supervisory control on the pultrusion processing based on real-time measurement of material properties and process parameters (i.e. viscosity, Tg) linked to the process equipment will allow the aeronautical manufacturing industry to reduce the energy and infrastructure utilisation, contributing to cost savings in the actual manufacturing phase at 10-20% in average. The use of on-line sensing and quality control devices will allow the elimination of all safety margins and related overheads posed by the raw material manufacturers (i.e. resin manufacturers) in the recommended cure cycles of the matrices.
The reduction of scrap in manufacturing through the use of the combined sensing and QA system is substantiated with the capability of the sensors to provide in real-time measurements of the material properties and the capability of the on-line QA system to inspect the quality of the parts. This knowledge has a dual role: on one hand safeguards the accurate tracking of the process path according to the predefined schedule and on the other hand allows the timely correction of the process parameters (temperature at the die and post-former, line speed) in order to arrive at acceptable quality of the produced parts. Given that in approximately 15-25% of the currently produced aerospace quality profile lengths at EXEL the off line quality inspection detects structural faults, the use of the above mentioned technologies will reduce significantly the scrap in processing and will ensure quality in manufacturing. The reduction of scrap will save material and reduce waste of energy and human resources.

C. Societal benefits
Time to market can be mainly reduced due to a higher flexibility in the process and supply chain. If required the whole process can be either, integrated in one facility or also separated into impregnation/pre-curing, curing and final shaping. The first concept aims on a reduction of part transport and storage costs; the latter offers the chance to leave sensitive tasks to specialists. The actual concept can be chosen as it is technically and economically suitable.
Today Japanese suppliers are dominating the market in the field of low complex pultruded CFRP profiles. The upper floor beam of Airbus A380 is a well known example. PUL-AERO will establish a technology and product line for complex curved profiles in Europe with a realistic chance to be the first on the market and to claim corresponding intellectual property. To gain control or at least secure knowledge about the use of those technologies will improve the competitive situation of the European aerospace industry. Today the knowledge about the use of such technologies by the competitor (787 black fuselage) is only based on rumours.
It can be expected that, by developing and saving this know-how for the European industry, a considerable number of high tech jobs can be secured or even newly established.
To face the challenge of reaching all the stated goals in a time frame allowing a utilisation right before 2020, a joint effort of the most experienced partners with their specific know how and facilities across Europe is combined in the project.
A national approach would not allow to make all steps forward in parallel, which are needed to meet the goals of vision 2020 and to support significantly the European position in the global competition in aerospace. Nevertheless, the achievements of national research activities will be utilized in the project by the partners coming from the different EU countries.
Finally, it should be mentioned that also other European key industries like automotive, ship building, railway or architecture will benefit from the results of the project, because affordable and high performance profiles are needed in many structural applications.

D. Dissemination activities
D1. Public web site
A website (www.pul-aero.eu) was set up at the beginning of the project in order to disseminate the project concept and latest activities. A password protected area was incorporated for the storage of project documents to be accessed by the project partners. Updates were made to the website during the project.
The website will remain in place for three years post-project, until at least 28 February 2020.

D2. Publications/presentations
Due to the innovative nature (and potential patentability) of some of the project outcomes, a strategic decision was made not to publish too many details of the project until protection was in place.
Two publications, giving a general project overview were made during the project:
• “PUL-AERO”, Dr J Hartley, Exel Composites, Aerodays 2015, London, UK, 20-23 Oct 2015
• “Making curved composites the PUL-AERO way”, Results in Brief, EU Cordis Service, 20 March 2017 (http://cordis.europa.eu/result/rcn/191261_en.html)

D3. Industrial Seminar
An Industrial Seminar was scheduled for the end of the project. However, despite interest form some big players in the industry, insufficient interest was received to make the event viable, so regrettably the decision had to be made to cancel it. All companies that had showed an interest in the event have since been contacted with a presentation of the PUL-AERO technology and an invitation of a demonstration.

D4. Post-project dissemination
It has been decided that agreement continues to be sought from all partners prior to the dissemination of any results from the PUL-AERO project. Notification of the intent to publish will be sent to all partners. If no objection is received within 12 days, agreement will be assumed.
Among the dissemination activities currently planned, there are three presentations in trade shows and two journal papers.

E. Exploitation plan
E1. Exploitable results
To achieve the objective of increasing the competitiveness in aerospace manufacturing industry, the PUL-AERO project has developed a novel composite materials processing practice in order to produce lighter, lower cost and higher quality (thus more safe) structural parts. This improvement will help satisfy the potential of composite manufacturing methods, which is currently limited because of a lack of automation, the use of very high capital cost equipment and material and the requirement for highly skilled workers.
The PUL-AERO project has developed an advanced pultrusion line with the capability to produce curved aerospace composites while taking full advantage of the benefits of the process. The specific project outputs consist of:
• A novel configuration of the pultrusion processing of composites through the separation of process in two stages: one with linear heated dies for curing and one with curved post forming dies for shaping.
• An alternative pultrusion process route involving production of a semi-solid intermediate product where the semi-cured stringers will be integrated with the dry reinforcement of the skin.
• The resin injection system for pultrusion processing with full design, construction and operation guidelines including sensors and actuators for the supply of standard quality liquid resin towards full wetting of the fibres before they enter in the heated die.
• Simulation models of the materials and processing.
• A functional and integrated process monitoring and on-line NDT system based on sensors positioned in the resin bath and on-line inspection system for the detection of structural faults in the processed materials.
• A complete signature of the process accompanying the produced part from the supervisory control system comprising of temperature controllers output, sensor signals from liquid and polymerised resin, on-line inspection data and pultrusion line operation data (line speed and pull force) giving rise to reliable and ecoefficient processing of composite structures.
• The manufacturing of curved aerospace grade components in a continuous process with improved cost efficiency and known quality.

E2. Commercialisation strategy – equipment
A number of sub systems from the PUL-AERO production line have been identified as having individual commercialisation potential. The actions highlighted in order to achieve sales success with these innovations are shown below.

Actions:
(a) Marketing of resin injection system - Partner: Isojet - Timescale: 2017 onwards
(b) Explore patenting of in-line NDT system - Partners: Advise/Exel - Timescale: July 2017
(c) Consider software opportunities for process model - Partner: Cranfield - Timescale: 2017

E3. Commercialisation strategy – straight stringers
The PUL-AERO project has demonstrated a production run (24 hours) of straight stringers and the quality has been verified by NDT techniques. However, in order to be an acceptable manufacturing route for the aerospace industry, various qualification processes need to be met. The consortium is prepared to fund these activities in order to gain the exploitable advantage they perceive the process will bring. The schedule of necessary activities is outlined below. It is anticipated that serial production could begin as early as 2018, ahead of the year 2020 predicted in the Description of work.

Specific actions for the commercial production of straight stringers have been planned, each with assigned budget, including (a) optimisation/ further verification of in-line NDT system, (b) Certification/qualification of NDT discussions, (c) Certification of personnel and service, (d) Process Optimisation for quality, (e) Obtain qualification (demonstrate equivalence), (f), Market benefits of PUL-AERO stringers, (g) Aim to produce commercial produce.

E4. Whilst the PUL-AERO project has demonstrated production runs for straight stringers, it is acknowledged by the consortium that further development is required to fully demonstrate the production potential of curved stringers. Although further development is still required, Exel believes the advantage to be gained from the pultrusion of curved stringers warrants the development costs and is prepared to fund some of this work. As stated in the Description of Work, it is still expected that the technology will be ready for serial production in 2020.

Specific actions for the commercial production of curved stringers have been planned, each with assigned budget, including: (a) Further development of stage 2 processing, (b) Look for further funding e.g. follow-on project, (c) Optimisation of processing of curved stringers, (d) Incorporate curvature operation into production line, (e) Aim for serial production.







List of Websites:
www.pul-aero.eu

Contact details:
Dr. John Hartley - Director - Exel Composites UK - Fairoak Lane, Whitehouse, Runcorn Cheshire WA7 3DU - United Kingdom
E-mail address: john.hartley@exelcomposites.com

Related information

Reported by

FIBREFORCE COMPOSITES LTD
United Kingdom

Subjects

Transport
Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top