Futurities Magazine ha inserito tra le storie di successo il progetto MULCOM, con cui CETMA ha vinto la prima opencall di FF4EuroHPC assieme al partner Manta Group.
A partire da giugno 2021, i ricercatori del CETMA Pasquale Bene, Rosario Dotoli ed Antonio Gerardi hanno gestito uno dei 16 esperimenti applicativi previsti con l’obiettivo di mostrare l’enorme impatto ed i benefici legati all’uso di HPC (High Performance Computing) per le piccole e medie imprese del settore manifatturiero.
L’obiettivo di MULCOM è stato quello di sviluppare un modello numerico multi-scala e multi-fisico per ottimizzare il processo di autoclave per la produzione di parti in composito, massimizzando la qualità del componente finale e minimizzando il relativo sviluppo in termini di tempo e costi.
Di seguito l’articolo integrale pubblicato dalla rivista:
Multiphysics and multiscale modelling of aeronautical components
The challenge: developing a more effective manufacturing process for aeronautical components
The autoclave moulding process, in which composite layers are placed on a mould according to a lamination sequence and then cured inside an autoclave using vacuum, heat, and pressure, is the main method used in the aerospace industry for manufacturing composites.
This process involves both mechanical and chemical phenomena, and a correspondingly large number of variables influence the final result.
Working with innovative materials and geometries leads to an increase in the number of defects and voids in the finished components, which are then rejected. During the curing process, the mechanical stresses in the various materials increase, which can lead to unwanted consequences. For example, during the cooling phase, the different thermal expansion coefficients of the fibre, matrix, and mould materials generate high residual stresses that can lead to defects in the finished composite parts. Given the expense of the autoclave process, it is important to minimize defects in the finished parts.
Currently, a costly trial-and-error approach is used to find the optimal process parameters to produce components with complex shapes, minimizing the risk of voids or geometric distortions. This leads to long development times and high costs.
The MANTA GROUP’s goal is a more efficient production process for its products: finding the optimal process parameters by means of multiscale, multiphysics numerical simulations. This significantly reduces the development time and costs required compared to the trial-and-error approach currently used.
Building multiscale numerical models using HPC resources
To optimize the parameters of the autoclave process (e.g. lamination sequence, maximum temperature, curing times, heating times, maximum pressure), it is necessary to simulate the various phenomena that occur during the curing process in
order to predict the effects of the parameters on the quality of the components being manufactured.
To this end, two separate multiphysics and multiscale numerical models were set up using HPC resources. In detail, these are
1) a thermo-mechanical model (on the macro scale) to predict the dimensional variations of the laminates due to residual stresses generated during the autoclave process;
2) a fluid-structure model (on the micro scale) to simulate the resin flow during the application of pressure. Both numerical models were validated by comparison with experimental test results.
To reduce the computing time, the utilization of HPC resources and the scalability of the simulations were substantially analysed.
Thermo-mechanical model (macroscale)
The aim was to study the dimensional variations of flat and L-shaped profiles due to residual stresses generated during the curing process.
The technological process used to manufacture these profiles was vacuum bagging. To simulate the curing process, a numerical finite element (FE) model was implemented and subsequently validated through experimental tests.
To comprehensively simulate the manufacturing processes of composite materials, various phenomena that occur during the curing process such as thermal flows and curing kinetics were considered because they lead to macroscopic defects.
The curing process was modelled with a thermo-mechanical approach that covers the curing phase and analyses residual stresses. In this phase, heat transfer and curing influence the stresses and, consequently, the deformations of the part.
The material is treated as a composite and the temperature and pressure cycles (with their effects on shrinkage and spring-in) are simulated.
Fig. 1 shows the contour plot of the dimensional changes in the flat laminate during the autoclave process. As the temperature rises the composite material’s strength and mechanical properties increase as a result of the polymerization of the resin.
In addition, the chemical transformation of the resin generates tension stresses due to chemical shrinkage. During the cooling phase, residual stresses increase due to the CTE (coefficient of thermal expansion), and the misalignment of fibres, matrix, and mould materials.
Residual stresses cause unwanted dimensional changes during demoulding.
The geometric distortions after curing are approximately 0.5mm along the z-axis.
Fluid-structure model (microscale)
A microscale fluid-structure model was implemented to simulate the effects of pressure. The FE model generated at the microscale enabled simulations to be conducted concerning the effect of
resin flow pressure on the defined fibre layup (see Fig.2).
The optimum pressure value for the process is close to 0.6MPa (6 bar) and represents an acceptable compromise between tow integrity, void level, compaction of the fabric layers, and the mechanical
behaviour of the final composite material.
Analysis of the fluid-structure interaction at the microscale enabled the voids and resin pockets generated during the process and their positions in the RVE (representative volume element) geometry to be studied.
In addition, the analysis provided the tow displacement with respect to the initial tissue architecture. By
increasing the pressure, the rotation of the tow was acceptable if the load level did not exceed 6bar.
Furthermore, the fabric maintains its orthogonal architecture under greater compaction
minimizing the presence of voids.
Using the material properties, lamination sequence, part geometry, and the autoclave cure cycle specification as input parameters, the HPC simulations set up in the first phase proved able to
provide the required information on the resulting part distortion and the possible defects in the finished part in a very short time.
This HPC-supported simulation workflow now allows the MANTA GROUP to easily find the optimal parameters for the production process in a matter of minutes, thus reducing development time
and significantly reducing the number of physical tests required.
Business benefits: reduction of development time and costs and increased competitiveness
-MANTA (end-user) expects to reduce design costs by 50% (saving approximately €100,000 per year), material waste by 70% (saving approximately €60,000 per year), and raw material usage by 15% (saving approximately €150,000 per year;
– CETMA expects the success story to lead to new R&D projects and consultancy services with an increase in turnover of about €50,000 per year;
-CINECA aims to become MANTA’s HPC resource provider and estimates the related increase in turnover at €20,000 per year and will leverage the success story to attract new customers estimating a further increase in turnover of the same order.
HPC-based simulations were used to produce high-quality composite components, reducing development time and costs while increasing competitiveness.
As autoclave moulding is likely to remain the main production technology for aerospace structures
for at least the next ten years, this significantly strengthens the MANTA GROUP’s business position.
Furthermore, the improved know-how of the autoclave process offers MANTA the ability to profitably enter many other sectors besides aerospace (e.g. luxury boats, automotive, sport). This will help the company to attract new customers by offering a complete service from design to component production.
The FF4EuroHPC project has received funding from the European High-Performance Computing Joint Undertaking (JU) under grant agreement No. 951745. The JU receives support from the European Union’s Horizon 2020 research and innovation programme and from Germany, Italy, Slovenia, France, and Spain.