Periodic Reporting for period 1 - D-STANDART (DURABILITY MODELLING OF COMPOSITE STRUCTURES WITH ARBITRARY LAY-UP USING STANDARDIZED TESTING AND ARTIFICIAL INTELLIGENCE)
Okres sprawozdawczy: 2023-01-01 do 2024-06-30
Durability is the ability of a physical product to remain functional without requiring excessive maintenance or repair throughout its design lifetime. Although maximising durability seems to be an obvious goal, it typically requires more resources before operational use, it results in higher reserve factors for structural responses and it comes at a higher cost. Therefore, finding the optimal level of durability during the development process of a product requires models that can accurately predict the use of resources, the structural responses and the overall cost. For composites, the available models are slow, they do not have the required accuracy and lack integration into the design process. D-STANDART aims to make them fast and accurate, promoting their use in the design process thereby enabling a further acceleration of the net-zero transition.
The D-STANDART scope is limited to the improvement of two engineering activities of high-performance composite structures: the evaluation of their fatigue resistance and the characterisation of cradle-to-cradle environmental aspects.
In order to improve the evaluation of the fatigue resistance of composite structures, current material characterisation methods will be improved and new (high-frequency) fatigue characterisation standards will be developed. Also, simulation methodologies will be developed to correctly quantify parameters that determine the fatigue behaviour of composites. Data obtained from experiments will be used in high-fidelity models to simulate defect growth in various lay-ups and at various scales. To apply these computational expensive models in an industrial design environment, D-STANDART develops Artificial Intelligence (AI) surrogate models, trained using the aforementioned high-fidelity simulation data.
For the characterisation of cradle-to-cradle environmental aspects, life-cycle and techno-economic assessments are used. From these assessments, models are generated that allow the assessment of sustainability effects of design alternatives thereby giving product developers the opportunity to efficiently account for sustainability in their decision-making processes
The D-STANDART scope is limited to the improvement of two engineering activities of high-performance composite structures: the evaluation of their fatigue resistance and the characterisation of cradle-to-cradle environmental aspects.
In order to improve the evaluation of the fatigue resistance of composite structures, current material characterisation methods will be improved and new (high-frequency) fatigue characterisation standards will be developed. Also, simulation methodologies will be developed to correctly quantify parameters that determine the fatigue behaviour of composites. Data obtained from experiments will be used in high-fidelity models to simulate defect growth in various lay-ups and at various scales. To apply these computational expensive models in an industrial design environment, D-STANDART develops Artificial Intelligence (AI) surrogate models, trained using the aforementioned high-fidelity simulation data.
For the characterisation of cradle-to-cradle environmental aspects, life-cycle and techno-economic assessments are used. From these assessments, models are generated that allow the assessment of sustainability effects of design alternatives thereby giving product developers the opportunity to efficiently account for sustainability in their decision-making processes
In the area of material characterisation testing, selection of the materials of interest, manufacturing and conditioning of specimens for coupon testing were performed, and fatigue test data were generated and analysed. Detailed designs for both the aerospace and wind demonstrators have made.
Following a state-of-the-art review in micro-scale modelling of composites, the project generated the first AI surrogate model for static loading without defects. In addition, a framework for probabilistic meso-scale modelling was established and the workflow was defined for how to translate the AI surrogate fatigue model into a macro-scale element applicable to full-scale analysis.
Coupons with and without embedded defects were manufactured and test protocols for their fatigue characterisation were defined. The process modelling strategy to predict the formations of defects in composite laminates was developed. Also, the meso-scale modelling capability to predict the fatigue damage accumulation in arbitrary composite laminates was developed and validated.
The process flows to capture the financial and ecological costs of manufacturing for life cycle (LCA), life cycle costing (LCC) and techno-economic analysis (TEA) assessments were defined. An investigation into the relationship between LCA with durability and between LCC and durability were initiated. A new coefficient-based approach for simplified LCA in design was also defined.
For the digital thread, a thorough review of all the applicable procedural and test standards (from ASTM and ISO) related to composite materials testing was carried out and based on that the architecture framework for the digital thread was established and the Specimen Test Suite (STS) platform developed.
Following a state-of-the-art review in micro-scale modelling of composites, the project generated the first AI surrogate model for static loading without defects. In addition, a framework for probabilistic meso-scale modelling was established and the workflow was defined for how to translate the AI surrogate fatigue model into a macro-scale element applicable to full-scale analysis.
Coupons with and without embedded defects were manufactured and test protocols for their fatigue characterisation were defined. The process modelling strategy to predict the formations of defects in composite laminates was developed. Also, the meso-scale modelling capability to predict the fatigue damage accumulation in arbitrary composite laminates was developed and validated.
The process flows to capture the financial and ecological costs of manufacturing for life cycle (LCA), life cycle costing (LCC) and techno-economic analysis (TEA) assessments were defined. An investigation into the relationship between LCA with durability and between LCC and durability were initiated. A new coefficient-based approach for simplified LCA in design was also defined.
For the digital thread, a thorough review of all the applicable procedural and test standards (from ASTM and ISO) related to composite materials testing was carried out and based on that the architecture framework for the digital thread was established and the Specimen Test Suite (STS) platform developed.
Although the main D-STANDART project results will be generated during the second half of the reporting, a few highlights from the first half of the project are worth mentioning because they either are already beyond the state-of-the-art with respect to the beginning of the project or their intermediate results reinforce the confidence that D-STANDART is on the right track.
From the material characterisation activities, it is confirmed that there is a real need for standardising composite fatigue characterisation, because without standardisation, processing of raw fatigue test sensor output can lead to significant differences in (derived) material characteristics. Although discussions on how to standardize fatigue characterisation are in progress, the consortium is confident they can deliver a clear recommendation, ready to be adopted by standardisation bodies for both the low-frequency (current practice) as the high-frequency (novel) fatigue tests.
Another highlight is the development of a coefficient-based life cycle assessment (LCA) model. This allows for fast and accurate LCA assessments early in the design process of a product.
Also, the progress in the high-fidelity meso-scale modelling of fatigue damage initiation and propagation and the AI trained surrogate models obtained from the running these models is worth mentioning. Initial validation results of both the high-fidelity and the AI models show their excellent predictive capability. This is a key enabler for the further uptake of efficient fatigue resistance assessments of high-performance composite structures.
From the material characterisation activities, it is confirmed that there is a real need for standardising composite fatigue characterisation, because without standardisation, processing of raw fatigue test sensor output can lead to significant differences in (derived) material characteristics. Although discussions on how to standardize fatigue characterisation are in progress, the consortium is confident they can deliver a clear recommendation, ready to be adopted by standardisation bodies for both the low-frequency (current practice) as the high-frequency (novel) fatigue tests.
Another highlight is the development of a coefficient-based life cycle assessment (LCA) model. This allows for fast and accurate LCA assessments early in the design process of a product.
Also, the progress in the high-fidelity meso-scale modelling of fatigue damage initiation and propagation and the AI trained surrogate models obtained from the running these models is worth mentioning. Initial validation results of both the high-fidelity and the AI models show their excellent predictive capability. This is a key enabler for the further uptake of efficient fatigue resistance assessments of high-performance composite structures.