Skip to main content

Predictive Simulation of Defects in Structural Composites

Periodic Reporting for period 1 - PSIDESC (Predictive Simulation of Defects in Structural Composites)

Reporting period: 2017-06-01 to 2018-11-30

Aviation has become an important part of our economy and contributes to the way the world interacts, communicates and grows. As the number of passengers flow increases regularly on an annual basis and given the need to replace aging fleets, innovation is at the heart of design and manufacturing of the aircraft of the future.

Two technical challenges stand out in aviation with respect to the ongoing work done in the PSIDESC project. The first is the environmental impact air travel could have if technologies don’t evolve to reduce emissions and fuel consumption of aviation over its complete life cycle. The second is related to following up with manufactures rates of aircrafts being ordered while keeping costs under control. Innovation also plays a key role here to ensure that the European aircraft industry will remain competitive with regards to emerging technologies.

Composite materials based on high strength and stiffness reinforcements such as carbon fibers, have the possibility to reduce the overall weight of the structure of an aircraft and improve its service life. Nevertheless, performance of these materials depend on the quality of manufacturing and this has been an ongoing topic of research since the beginning. As zero defects in a manufactured part will make manufacturing and inspection costs soar, managing the effect of defects is key for production to set manageable quality targets strictly in line with safety requirements.

The PSIDESC project contributes to this overall approach to manage the effects of defects with particular attention to composite materials manufacturing in and out-of-autoclave. Qualification through testing as well as scrapping a part on which a slight deviation from specifications is identified (wrinkles in fibers, small gaps in fiber alignment, resin rich regions) are both very costly. Ideally a part should be designed in a robust ways such that a small defect will not affect its structural capabilities. Also, if deviations are found in processing during manufacturing, a predictive tool that can simulate the effect of this deviation regardless of its location can help save time and money.

Cenaero, E-Xstream (MSC Software Company) and Sonaca, are joining efforts to implement and validate this numerical tool at the test specimen scale up to a large scale aeronautical structure defined with Airbus, the topic manager. The goal is to develop a tool capability to enlarge panel acceptance criteria to support composite part manufacturing for next generation aircraft.

To achieve this goal, the following work packages are defined:
- Manufacturing defect Selection (WP2): The goal is to select the most relevant defects to be treated in the project
- Development of probabilistic tool (WP3). The toolkit will allow the generation of random defects in the structure, the post processing of the mechanical response and computing the reliability of the structure
- Toolkit validation at coupon level (WP4) and at panel level (WP5)
Over the first 18 months of the project, the first four work packages have been started in order to define, develop and test the workflow at a specimen level for different manufacturing defects identified as critical and representative of processes under consideration.
A high level of cooperation was required throughout these work packages as knowledge and expertise on topics ranging from manufacturing, modeling, effect of defects, is in the hands of different partners and had to be discussed regularly between partners and with the topic manager.

Three main types of defects are under consideration after carefully analyzing the occurrence rate of defects, the amount of processes potentially concerned, the knock-down on properties these defects may have, the detectability and avoidability, the expected gain if the defect impact is reduced or suppressed, the repairability and finally the amount of data available that can be used in the frame of the developments. The defects that are selected are porosity, waviness and gaps/overlaps of tows.

Based on data available and description of the defect in terms of size, location and detection thresholds, the representation of the defect were started. For porosity, this was done at an elementary level as we can treat it as a change in local material properties, for gaps and waviness, the effect of the defect is intimately linked to its orientation and its geometry and must be treated as such when looking at the macro structural scale. A valid workflow to include these defects was thus thought up and will be tested with different variants on the demonstrators.

From a numerical standpoint, the representation of porosity and waviness are complete, while gaps will still undergo developments in the second period. For the design and statistical evaluation workflows, some tests were successfully performed on simple geometries and criteria and must now be confronted at the panel scale to check the validity and convergence of the evaluations.
First, from and design and manufacturing point of view, the project will be a key enabler to the expansion of the design space by effectively evaluation the impact of subsequent steps in the manufacturing process on the structural integrity in a non-deterministic approach. Currently, most stochastic methods look at a distribution of equivalent material properties (compression strength, interlaminar strength, Young’s modulus). This gives an idea of the sensitivity to material properties, but not how it can related to variations due to manufacturing. The link between defects arising through processing and structural integrity needs to be matured and validated and we are convinced that this project will address and bring innovation with respect to this aspect.

Technologies to answer to the challenge in material behaviour variability and distribution are already available in numerical tools such as Digimat or Genoa. Other tools such as PAM-Composites or Moldex3D allow identifying locations prone to defects with respect to a consolidation strategy. What will differentiate the developments of this project is the capacity to quantify the influence of different defects on structural integrity and therefore take the right design decisions or scrap a part or accept it after a deviation in the manufacturing.

In terms of numerical developments the project focus in on statistical material and defect representation.

Today three kind of methodologies have reached sufficient maturity in order to be used industrially:
- Advanced material engineering based on homogenization
- Multi-scale structural computation allowing accounting for the local effective properties
- Methodologies bridging the process to the structural FEA
The ambition of this project concerning numerical simulation are mainly:
- To improve material models and methodologies to make them applicable
- To perform variability analysis by setting-up a probabilistic toolkit integrating micromechanical material and structural models
- To set-up representative severity and spatial distributions of these defects
- To set-up a toolkit based on well-known industrial level software like Digimat, Abaqus and openTurns.
Achieving this and validating the toolkit at the level of an experimental demonstrator and full scale numerical demonstrator will ensure that these methodologies can help evolve the design and effect of defect paradigm in manufacturing of composite structures.
Example of a gap due to a twisted tow
Through thickness waviness due to improper consolidation near the stiffener
Representation of the effect of statistical distribution of a porosity defect on a panel
Porosities in a composite laminate