Skip to main content

Design against DISTortion of metallic aerospace parts based on combination of numeRical modelling ACTivities and topology optimisatION

Periodic Reporting for period 2 - DISTRACTION (Design against DISTortion of metallic aerospace parts based on combination of numeRical modelling ACTivities and topology optimisatION)

Reporting period: 2017-07-04 to 2019-01-03

Current design tools and methodologies do not account for part distortions in manufacturing. Distortions are critical during prototyping and ramp up stages of new fuselage components and they increase developing costs, times and generate wastes and scraps. Design Against Distortion topic is focused on the development of numerical modelling strategies which can anticipate distortions even from the design stage.
Within this context, the following main achievements must be highlighted as once the activities of DISTRACTION project are finished:
1. A methodology for prediction of machining induced distortions has developed. This methodology is based on FEM numerical models. The prediction models developed in DISTRACTION project present accuracy greater than 90%, validated by experimental trials carried out in aluminium and titanium machined components.
2. A simplified and rapid modelling approach has developed for Selective Laser Melting manufacturing process induced distortion prediction, based on FEM numerical simulations. The modelling methodology is based on linear transient inherent strain concept. Inherent strains calibration/validation methodology has been established: based on cantilever beam geometry experimental results and iterative fitting algorithm.
3. Prediction capability of SLM process modelling methodology has been guaranteed performing experimental trials. Several titanium mock ups have been manufactured by SLM and experimentally measured distortion field have been compared with the FEM predicted values. These comparisons have shown a very good correlation between experimental and FEM model predicted distortion fields.
4. A correlation methodology to compare experimentally measured distortion field with FEM simulations has been defined for complex geometry mock-ups. The methodology is based on the comparison of results (distortion field) by means of deviations (difference in the size and shape of a manufactured part versus its design specifications) instead of displacements (FEM).
5. A general optimization framework, considering topology, machining and SLM process parameters, has been developed.
6. The topology optimization code has been developed and the source code has been made available to the project partners.
7. The optimization methodology has been applied to an industrial use case. Based on numerical simulation, it has been demonstrated that gradient-based topology optimization procedures can be utilized to address fabrication issues due to AM. The flexibility of the formulation permits consideration of many AM process responses. Numerical analysis of the optimized designs showed that the final distortion and likelihood of jamming can be reduced.
The project is divided into 5 work packages (WP). WP1 is related with project management and WP5 is related to knowledge protection, exploitation and dissemination activities.
Regarding RTD activities, WP2 faces the development of rapid distortion prediction. WP3´s goal is the development of rapid distortion prediction methods for additive layer manufacturing. WP4 copes with the topology optimization accounting for distortion.
The work performed in the project will be summarized in the following lines:
Related to WP2:
1. A simplified mechanical approach was developed; the machining induced residual stresses has been mapped in the contour of the final geometry by means of different elements: shell elements, membrane elements and layer solid elements and boundary layers elements.
2. Machining experimental trials has been carried out for Al7075 and Ti6Al4V rolled plates.
3. Correlation of experimental trials and FEM models prediction has been carried out for aluminium and titanium mock ups. the mean absolute error between experimentally measured distortions and the FEM model predicted is below 10%.
As for WP3:
4. Different simplified and rapid modelling approaches have been investigated. Inherent strains calibration/validation methodology has been established: based on cantilever beam geometry experimental results and iterative fitting algorithm.
5. Validation coupons have been manufactured by SLM, following different scanning strategies. Additionally, a mock up with more complex geometry has been manufactured
6. Post-process distortion measurements of each coupon have been performed in order to validate the numerical model against these measurements.
7. After the preliminary validation of FEM model, a use case has been defined for further development. Pclip geometry has been defined as a use case. Based on the original design, different optimized designs have been manufactured by SLM. SLM process induced distortion field has been predicted for each optimized design.
8. Comparison methodology for numerical distortion prediction results with the experimental results has been determined in WP3.
Related to WP4:
9. General optimization framework with an embedded, single-step machining process simulation developed, implemented, and subjected to initial testing.
10. Adjoint sensitivity formulation of general distortion response function with respect to topology (density) and process parameter variables developed, implemented, and compared with common `virtual work' response function.
11. A linearized transient AM process simulation is formulated for inclusion in a topology optimization design loop.. Small-scale support structure topology optimization has been demonstrated. Structural topology optimization of a combined AM process and in-service distortion formulation is demonstrated. The novelty is found in the exploitation of the seemingly detrimental AM process distortions to compensate a priori for in-service distortion.
12. Use case redesign, pclip, for manufacture and support-structure optimization is carried out by means of topology optimizations with embedded AM process simulation in the optimization loop.
13. The feasibility of gradient-based topology optimization procedures has been demonstrated to address fabrication issues due to AM.
14. A user guide for the topology optimization code developed during the DISTRACTION project has been defined and summarized in a document.
The potential improvements focus on robust design for manufacturing using ALM (Additive Layer Manufaturing), considering robustness as the low distortion amount and its predictability.
As a general technical impact it should be considered that the consequences obtained about distortion prediction may be very useful for all those industries using machining. Results could recommend ALM or a hybrid process combining ALM and machining, as a much better process for high quality parts and so on. That is, machining industry could consider higher specialisation on specific parts/geometries, or even including ALM equipment for a more general solution. Anyway results may impact machining industries and those components that could be competitive by ALM (change of manufacturing process).
Furthermore, the integration of distortion simulation into topology optimisation procedure will also have a strong impact on the ALM technology development and its advancement toward higher TRLs, given that the study of the best manufacturing conditions, the manufacturing itself, and the final part validation will be performed.
The most relevant final result to realise the industrial potential of the project: the new topology optimisation code, capable of accounting for risk of part distortion.